KRAS mutation testing of tumours in adults with metastatic colorectal cancer: a systematic review and cost-effectiveness analysis

Targeted mutation detection tests

All targeted tests are commercial kits and these look for different numbers of mutations within specific codons of the KRAS gene and have differing limits of detection. They may therefore differ in their ability to accurately differentiate patients who are likely to benefit from treatment with cetuximab in combination with standard chemotherapy from those who should receive standard chemotherapy alone.

The Therascreen KRAS RGQ PCR Kit is a CE-marked real-time PCR assay for the qualitative detection of seven mutations in codons 12 and 13 of the KRAS gene. It has been approved by the US FDA for the application covered by this assessment, that is, the selection of patients with mCRC for treatment with cetuximab. The Therascreen KRAS RGQ PCR Kit uses two technologies for the detection of mutations: ARMS (amplification-refractory mutation system) for mutation-specific DNA amplification and Scorpions for detection of amplified regions. Scorpions are bifunctional molecules containing a PCR primer covalently linked to a fluorescently labelled probe. A real-time PCR instrument [Rotor-Gene Q 5-Plex HRM (high-resolution melt) for consistency with CE marking] is used to perform the amplification and to measure fluorescence. 17 There is an earlier version of the Therascreen KRAS PCR Kit that also uses ARMS and Scorpions for the detection of KRAS mutations and is designed to detect the same KRAS mutations as the current, reformulated and revalidated version. Evidence for both versions will be included in this assessment.

The Therascreen KRAS Pyro Kit is a CE-marked test for the quantitative measurement of 12 mutations in codons 12, 13 and 61 of the KRAS gene. The kit is based on pyrosequencing technology and consists of two assays, one for detecting mutations in codons 12 and 13 and a second for detecting mutations in codon 61. The two regions are amplified separately by PCR and then amplified DNA is immobilised on streptavidin sepharose high-performance beads. Single-stranded DNA is prepared and sequencing primers added. The samples are then analysed using the PyroMark^® Q24 System (QIAGEN). The KRAS Plug-in Report is recommended by the manufacturer for the analysis of the results; however, the analysis tool within the pyrosequencer can also be used. 18

The cobas KRAS Mutation Test from Roche Molecular Systems is a CE-marked TaqMelt^™ real-time PCR assay intended for the detection of 19 mutations in codons 12, 13 and 61 of the KRAS gene. The assay uses DNA extracted from FFPE tissue and is validated for use with the cobas^® 4800 System.

The KRAS LightMix Kit from TIB MOLBIOL is a CE-marked test designed for the detection and identification of mutations in codons 12 and 13 of the KRAS gene. The first part of the test involves PCR amplification of the KRAS gene. To reduce amplification of the wild-type KRAS gene and therefore enrich the mutant KRAS gene, a wild type-specific competitor molecule is added to the reaction mix. This is called clamped mutation analysis. The second part of the test procedure involves melting curve analysis with hybridisation probes. The melting temperature is dependent on the number of mismatches between the amplification product and the probe and allows the detection and identification of a mutation within the sample. The test is run on the LightCycler^® instrument (Roche). 19

The KRAS StripAssay from ViennaLab is a CE-marked test for the detection of mutations in the KRAS gene. The test procedure involves three steps: the DNA is first isolated from the specimen; PCR amplification is then performed; and the amplification product is then hybridised to a test strip containing allele-specific probes immobilised as an array of parallel lines. Colour substrates are used to detect bound sequences, which can then be identified with the naked eye or by using a scanner and software. 20 There are two versions of the KRAS StripAssay: one is designed to detect 10 mutations in codons 12 and 13 of the KRAS gene and the other is designed to detect the same 10 mutations in codons 12 and 13 plus three mutations in codon 61.

Mutation screening tests

‘In-house’ laboratory-based tests are designed to detect all mutations within specific codons of the KRAS gene.

Pyrosequencing assays are the most commonly used method of KRAS mutation testing in UK laboratories (see Table 1). The process involves first extracting DNA from the sample and amplifying it using PCR. The PCR product is then cleaned up before the pyrosequencing reaction. The reaction involves the sequential addition of nucleotides to the mixture. A series of enzymes incorporate nucleotides into the complementary DNA strand, generate light proportional to the number of nucleotides added and degrade unincorporated nucleotides. The DNA sequence is determined from the resulting pyrogram trace. 21

Sanger sequencing is also a commonly used method (see Table 1); however, there is much variation in the detail of how the method is carried out. In general, after DNA is extracted from the sample it is amplified using PCR. The PCR product is then cleaned up and sequenced in both forward and reverse directions. The sequencing reaction uses dideoxynucleotides labelled with coloured dyes that randomly terminate DNA synthesis, creating DNA fragments of various lengths. The sequencing reaction product is then cleaned up and analysed using capillary electrophoresis. The raw data are analysed using analysis software to generate the DNA sequence. All steps are performed at least in duplicate to increase confidence that an identified mutation is real. It should be noted that sequencing works well only when viable tumour cells constitute at least 25% or more of the sample. 22

National Institute for Health and Care Excellence contact with laboratories (October/November 2012) suggested that several laboratories were planning to convert to next-generation sequencing in the coming year. As with Sanger sequencing, there is much variation in the methodology used to perform next-generation sequencing. The concept is similar to Sanger sequencing; however, the sample DNA is first fragmented into a library of small segments that can be sequenced in parallel reactions. 23

High-resolution melt analysis assays are also commonly used by laboratories participating in the UK NEQAS scheme (see Table 1). For this technique the DNA is first extracted from the sample and amplified using PCR. The HRM reaction is then performed. This involves a precise warming of the DNA during which the two strands of DNA ‘melt’ apart. Fluorescent dye that binds only to double-stranded DNA is used to monitor the process. A region of DNA with a mutation will ‘melt’ at a different temperature to the same region of DNA without a mutation. These changes are documented as melt curves and the presence or absence of a mutation can be reported. 24

Matrix-assisted laser desorption ionisation time-of-flight (MALDI-TOF) mass spectrometry is currently used by one laboratory participating in the UK NEQAS scheme. This technique involves extracting DNA and amplifying it using PCR. The PCR products are then cleaved and fragments separated based on mass by the MALDI-TOF mass spectrometer. This generates a ‘fingerprint’ of the DNA with each fragment represented as a peak with a certain mass. The ‘fingerprint’ of the test sample is compared with the ‘fingerprint’ of the wild-type DNA. A mutation would appear as a peak shift due to a change in the mass of a fragment caused by a base change. 25 MALDI-TOF mass spectrometry can be used to identify all mutations within selected codons in the KRAS oncogene and has a limit of detection of approximately 10% tumour DNA in a background of wild-type DNA. 26

Subgroup analyses of patients tested for KRAS mutation status from randomised controlled trials (RCTs) have shown that treatment with the EGFR-inhibiting monoclonal antibody cetuximab in combination with standard chemotherapy can increase progression-free survival (PFS) and tumour response in patients with KRAS wild-type tumours compared with standard chemotherapy alone. 27,28 In contrast, patients whose tumours were positive for KRAS mutations had reduced PFS and tumour response when treated with cetuximab in combination with standard chemotherapy compared with standard chemotherapy alone. 27,28 These two trials formed the basis of NICE technology appraisal 176,1 which recommends cetuximab in combination with standard chemotherapy for the first-line treatment of mCRC in patients whose tumours are KRAS wild type and whose metastases are confined to the liver and are unresectable. However, both of these trials used a pre-CE-marked version of the LightMix KRAS Kit (TIB MOLBIOL), which is not currently in use by any laboratory participating in the UK NEQAS scheme.

Diagnosis of colorectal cancer

This guideline states that patients referred to secondary care for suspected CRC should be assessed using colonoscopy, flexible sigmoidoscopy followed by barium enema or computed tomography (CT), dependent on comorbidities and local expertise and test availability. When a lesion suspicious of cancer is detected a biopsy should be performed to confirm the diagnosis. 29

All patients with histologically confirmed CRC should be offered contrast-enhanced CT of the chest, abdomen and pelvis to estimate the stage of the disease. Further imaging [e.g. contrast-enhanced magnetic resonance imaging (MRI) or positron emission tomography/computed tomography (PET/CT)] may be considered if the CT scan shows metastatic disease only in the liver. 29 The aim of further imaging is to identify those patients who have resectable metastases, or metastases that may become resectable following response to chemotherapy. For the second group of patients, European Society for Medical Oncology (ESMO) 2010 clinical practice guidelines for the treatment of advanced CRC30 recommend establishing KRAS mutation status to determine the best treatment regimen. These guidelines do not stipulate which specific mutations should be analysed or which test method should be used. The KRAS status of a patient’s tumour is identified through analysis of a biopsy sample or, more frequently, a section of resected tumour tissue. The tissue is fixed in formalin and embedded in a block of paraffin for storage by the pathologist, who also examines the histology and evaluates the tumour content of the sample. Macro dissection may be performed before DNA is extracted and mutation analysis is carried out to determine the KRAS status of the tumour.

To minimise turnaround time, guidance from the Royal College of Pathologists31 recommends that mutation testing should be ordered by the pathologist reporting on the cellular make-up of the tumour. However, this is not currently universal practice and often the decision to perform a KRAS mutation test is often taken at the multidisciplinary team (MDT) meeting. If a sample is stored as a FFPE specimen for a long time this can lead to DNA degradation, which can result in a higher chance of failure when testing for KRAS mutations. The timing of the KRAS test varies between patients, with some clinicians preferring to test at diagnosis, potentially before the disease becomes metastatic, and other clinicians waiting until the cancer has progressed to metastatic disease. If the KRAS status is tested early, then the result is then referred to if metastatic disease develops. It has been suggested that analysing multiple resection or biopsy samples from the same patient increases the chances of identifying a KRAS mutation because of potential heterogeneity between tumour sites. The evidence on this is conflicting, with studies reporting that testing a single site only will potentially misclassify between 2% and 10% of tumours as KRAS wild type. 32,33

Treatment of colorectal cancer

In patients with unresectable liver metastases whose primary tumour has been resected or is potentially operable and who are fit enough to undergo liver surgery, the aim of chemotherapy is to induce tumour response such that resection becomes possible. The KRAS mutation status of a patient’s tumour is used to determine the optimal chemotherapy regimen for this purpose. Evidence suggests that patients with KRAS wild-type tumours are more likely to benefit from treatment with an EGFR-inhibiting monoclonal antibody (cetuximab) in combination with standard chemotherapy. However, patients whose tumours are positive for KRAS mutations are more likely to benefit from standard chemotherapy alone. In addition, the overall health and the preferences of the patient should be taken into consideration when selecting treatment.

The choice of standard chemotherapy is covered by NICE clinical guideline 131,29 which recommends that one of the following sequences of chemotherapy is considered:

• oxaliplatin in combination with infusional fluorouracil plus folinic acid (FOLFOX) as first-line treatment and then single-agent irinotecan as second-line treatment

• FOLFOX as first-line treatment and then irinotecan in combination with infusional fluorouracil plus folinic acid (FOLFIRI) as second-line treatment

• oxaliplatin and capecitabine (XELOX) as first-line treatment and then FOLFIRI as second-line treatment.

The guideline further states that raltitrexed should be considered only for patients who are intolerant to fluorouracil and folinic acid or for patients for whom these drugs are not suitable. 29 NICE technology appraisal 6134 suggests that oral therapy with either capecitabine or tegafur with uracil (in combination with folinic acid) can also be considered as an option for the first-line treatment of mCRC.

With respect to the use of biological agents (EGFR inhibitors), NICE technology appraisal 1761 recommends cetuximab in combination with FOLFOX or FOLFIRI, within its licensed indication, for the first-line treatment of mCRC:

• in patients in whom the primary colorectal tumour has been resected or is potentially operable

• in patients in whom the metastatic disease is confined to the liver and is unresectable

• when the patient is fit enough to undergo surgery to resect the primary colorectal tumour and to undergo liver surgery if the metastases become resectable after treatment with cetuximab.

The European Medicines Agency marketing authorisation for cetuximab states that it is ‘indicated for the treatment of patients with EGFR-expressing, KRAS wild-type metastatic colorectal cancer’. 35 Therefore, KRAS mutation testing is an important component of the care pathway. Cetuximab (monotherapy or combination therapy) and bevacizumab (in combination with non-oxaliplatin chemotherapy) for the treatment of mCRC after first-line chemotherapy are not recommended in NICE technology appraisal 242. 36 However, these treatments may be given to some patients through the Cancer Drugs Fund. If cetuximab is considered in the third-line setting, KRAS status is often not retested but a decision will be made based on the result of the KRAS test performed earlier in the care pathway. No other biological agents are currently recommended by NICE for the first-line treatment of patients with unresectable liver metastases from CRC.

National Institute for Health and Care Excellence clinical guideline 13129 stipulates that all patients with primary CRC undergoing treatment with curative intent should be followed up at a clinic visit 4–6 weeks after the potentially curative treatment. They should then have regular surveillance including:

• a minimum of two CT scans of the chest, abdomen and pelvis in the first 3 years and

• regular serum carcinoembryonic antigen tests (at least every 6 months in the first 3 years).

They should also have a surveillance colonoscopy at 1 year after initial treatment and, if the result is normal, further colonoscopic follow-up after 5 years and thereafter as determined by cancer networks.

Measuring response to treatment

In 1979 the World Health Organization (WHO) and the International Union Against Cancer introduced criteria for the classification of the response of solid tumours to treatment. 37 These criteria were an early attempt to standardise reporting of response outcomes and were widely adopted; however, some problems with their use have subsequently developed. There has been variation in the methods used for incorporating into response assessments the change in size of measurable lesions, as defined by WHO; the minimum lesion size and number of lesions to be recorded have also varied; the definitions of progressive disease (PD) have sometimes been related to change in a single lesion and sometimes to change in overall tumour load (sum of the measurements of all lesions); and there has been confusion around how to use three-dimensional measures from new technologies such as CT and MRI in the context of the WHO criteria. 38 The Response Evaluation Criteria in Solid Tumours (RECIST) Group is a collaborative initiative that was instigated to review the WHO criteria. The RECIST criteria use the same categories as the WHO criteria [complete response (CR), partial response (PR), stable disease (SD) and PD]. 38 RECIST guidance recommends CT and MRI for measuring target lesions in response assessment and that imaging-based evaluation is generally preferable to clinical examination. It is suggested that follow-up assessments every 6–8 weeks is a ‘reasonable norm’. 38 Taking into account the longest diameter only for all target lesions, the RECIST criteria, as they are applicable to this assessment, can be summarised as follows:38

• CR – disappearance of all target lesions and no new lesions

• PR – at least a 30% decrease in the sum of the longest diameters of target lesions, taking the sum of the baseline diameters as the reference, and no new lesions

• PD – at least a 20% increase in the sum of the longest diameters of target lesions, taking the smallest sum of the longest diameters recorded since treatment started as the reference, or the appearance of one or more new lesions

• SD – neither sufficient shrinkage to be classified as PR or sufficient increase to be classified as PD, taking the smallest sum of the longest diameters recorded since treatment started as the reference, and no new lesions.

Best overall response is defined as the best response recorded from the start of treatment to disease progression. 38

First-line chemotherapy of unresectable colorectal liver metastases seeks to achieve a tumour response such that the tumour is judged to be resectable. For this reason, resection rate is considered the ideal reference standard for research question 2 and the optimal outcome measure for research question 3 (see Chapter 1). Objective response rate (ORR), defined as best overall response = CR + PR, is also of interest as there is some evidence that ORR correlates well with resection rate. 39 Tumour status following treatment/resection is defined by the residual tumour (R) classification, in which R0 = no residual tumour, R1 = microscopic residual tumour and R2 = macroscopic residual tumour.

This assessment compares the performance and cost-effectiveness of KRAS mutation testing options, currently available in the UK NHS, for the differentiation of adults with mCRC whose metastases are confined to the liver and are unresectable and who may benefit from first-line treatment with cetuximab in combination with standard chemotherapy from those who should receive standard chemotherapy alone.

Search strategy

Search strategies were based on target condition and intervention, as recommended in the CRD guidance for undertaking reviews in health care40 and the Cochrane Handbook for DTA Reviews. 42

Candidate search terms were identified from target references, browsing database thesauri [e.g. MEDLINE medical subject headings (MeSH) and EMBASE Emtree], existing reviews identified during the rapid appraisal process and initial scoping searches. These scoping searches were used to generate test sets of target references, which informed text mining analysis of high-frequency subject indexing terms using EndNote X5 reference management software (Thomson Reuters, CA, USA). Strategy development involved an iterative approach testing candidate text and indexing terms across a sample of bibliographic databases and aimed to reach a satisfactory balance of sensitivity and specificity.

The following databases were searched for relevant studies from 2000 to January 2013:

• MEDLINE (OvidSP) (2000 to Week 2 January 2013)

• MEDLINE In-Process and Other Non-Indexed Citations and Daily Update (OvidSP) (up to 21 January 2013)

• EMBASE (OvidSP) (2000 to Week 3 2013)

• Cochrane Database of Systematic Reviews (CDSR) (Wiley) (The Cochrane Library 2000 to Issue 12, 2012)

• Cochrane Central Register of Controlled Trials (CENTRAL) (Wiley) (The Cochrane Library 2000 to Issue 12, 2012)

• Database of Abstracts of Reviews of Effects (DARE) (Wiley) (The Cochrane Library 2000 to Issue 4, 2012)

• Health Technology Assessment (HTA) database (Wiley) (The Cochrane Library 2000 to Issue 4, 2012)

• Science Citation Index (SCI-EXPANDED) (Web of Knowledge) (2000 to 22 January 2013)

• Conference Proceedings Citation Index (CPCI-S) (Web of Knowledge) (2000 to 22 January 2013)

• Latin American and Caribbean Health Sciences Literature (LILACS) (Internet) (up to 24 January 2013), http://regional.bvsalud.org/php/index.php?lang=en

• BIOSIS Previews (Web of Knowledge) (2000 to 22 January 2013)

• National Institute for Health Research (NIHR) HTA programme (Internet) (up to 25 January 2013)

• International Prospective Register of Systematic Reviews (PROSPERO) (Internet) (up to 25 January 2013), www.crd.york.ac.uk/prospero/.

Completed and ongoing trials were identified by searches of the following resources:

• National Institutes of Health ClinicalTrials.gov (Internet) (2000 to 23 January 2013), www.clinicaltrials.gov/

• Current Controlled Trials (Internet) (2000 to 29 January 2013), www.controlled-trials.com/

• WHO International Clinical Trials Registry Platform (ICTRP) (Internet) (2000 to 25 January 2013), www.who.int/ictrp/en/.

Searches were undertaken to identify studies of KRAS testing for mCRC. The main EMBASE strategy for each set of searches was independently peer reviewed by a second information specialist using the Peer Review of Electronic Search Strategies Evidence-Based Checklist (PRESS-EBC). 43 Search strategies were developed specifically for each database and the keywords associated with CRC were adapted according to the configuration of each database. Searches took into account generic and other product names for the intervention. No restrictions on language or publication status were applied. Full search strategies are reported in Appendix 1.

Electronic searches were undertaken for the following conference abstracts:

• ASCO conference proceedings (Internet) (2007–13), www.asco.org/ASCOv2/Meetings/Abstracts

• ESMO conference proceedings (Internet) (2007–13), www.esmo.org/education-research/abstracts-virtual-meetings-and-meeting-reports.html

• American Association for Cancer Research conference proceedings (Internet) (2007–13), www.aacrmeetingabstracts.org/search.dtl

• Association for Molecular Pathology conference proceedings (Internet) (2007–13), www.amp.org/meetings/past_meetings.cfm.

Identified references were downloaded into EndNote X4 software for further assessment and handling.

References in retrieved articles were checked for additional studies. The final list of included papers was also checked on PubMed for retractions, errata and related citations. 44–46

Inclusion and exclusion criteria

Separate inclusion criteria were developed for each of the three clinical effectiveness questions; these are summarised in Table 2.

Inclusion screening and data extraction

Two reviewers (MW and PW) independently screened the titles and abstracts of all reports identified by searches and any discrepancies were discussed and resolved by consensus. Full copies of all studies deemed potentially relevant were obtained and the same two reviewers independently assessed these for inclusion; any disagreements were resolved by consensus. Details of studies excluded at the full paper screening stage are presented in Appendix 5.

Studies cited in materials provided by the manufacturers of the Therascreen KRAS RGQ PCR Kit and Therascreen KRAS Pyro Kit (QIAGEN), the cobas KRAS Mutation Test Kit (Roche Molecular Systems), the KRAS LightMix Kit (TIB MOLBIOL) and the KRAS StripAssay (ViennaLab) were first checked against the project reference database, in EndNote X4; any studies not already identified by our searches were screened for inclusion following the process described above.

Data were extracted on the following: study design/details, participant details (e.g. inclusion/exclusion criteria, age, liver metastases details, criteria for unresectability, performance status, previous treatments), KRAS mutation test(s) and mutations targeted, intervention details, clinical outcomes, test performance outcome measures (against treatment response as reference standard), details of specific mutations identified by outcome measure (when reported), test failure rates and limits of detection. Data were extracted by one reviewer, using a piloted standard data extraction form, and checked by a second (MW and PW); any disagreements were resolved by consensus. Full data extraction tables are provided in Appendix 2.

Quality assessment

The risk of bias in included RCTs was assessed using the Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. 47 Studies used to derive accuracy data, for the ability of KRAS mutation tests to predict treatment response, were assessed using the QUADAS-2 tool. 48 Risk of bias assessments were undertaken by one reviewer and checked by a second reviewer (MW and PW) and any disagreements were resolved by consensus.

The results of the risk of bias assessments were summarised and are presented in tables and graphs in the results section of this chapter and in full, by study, in Appendix 3.

Survey of laboratories providing KRAS mutation testing

We conducted a web-based survey to gather data on the technical performance characteristics of KRAS mutation tests. We sent an e-mail invitation via NEQAS to laboratories participating in the UK NEQAS pilot scheme for KRAS mutation testing. We used SurveyMonkey online software to run the survey. We structured the survey into sections on:

• laboratory details

• KRAS testing methods

• logistics

• technical methods

• costs.

When possible we used multiple choice options with tick boxes to make the survey quick and easy to complete. A copy of the survey is provided in Appendix 4.

Methods of analysis/synthesis

The results of studies included in this review were summarised by research question (see Chapter 1), that is, studies providing technical information on KRAS mutation testing in NHS laboratories in the UK (see What are the technical performance characteristics of the different KRAS mutation tests?), studies providing information on the accuracy of KRAS mutation tests for predicting response to treatment (see What is the accuracy of KRAS mutation testing for predicting response to treatment with cetuximab plus standard chemotherapy and subsequent resection rates?) and studies reporting information on how clinical outcomes may vary according to which test is used to select patients for treatment (see How do outcomes from treatment with cetuximab plus standard chemotherapy vary according to which test is used to select patients for treatment?). We planned to use a bivariate/hierarchical summary receiver operating characteristic (HSROC) random-effects model to generate summary estimates and a summary receiver operating characteristic (SROC) curve for test accuracy data49–51 and a DerSimonian and Laird random-effects model to generate summary estimates of treatment effects. However, because the review identified a small number of studies with between-study variation in participant characteristics, methods used to test for KRAS mutations and mutations targeted, we did not consider meta-analysis to be appropriate and have provided a structured narrative synthesis.

For all studies that provided data on accuracy for the prediction of response to treatment with cetuximab in combination with standard chemotherapy, the absolute numbers of true-positive (TP), false-negative (FN), false-positive (FP) and true-negative (TN) test results, as well as sensitivity and specificity values with 95% confidence intervals (CIs), are presented in results tables, for each reference standard response (e.g. ORR or resection rate) reported. When reported, data on the numbers of failed KRAS mutation tests and reasons for failure were also included in the results tables. The results of individual studies were plotted in the receiver operating characteristic (ROC) plane to illustrate the trade-off between sensitivity and specificity and for ease of comparison between test methods; separate plots were provided for each reference standard response. For RCTs providing information on how clinical outcomes may vary according to which test is used to select patients for treatment with cetuximab in combination with standard chemotherapy, hazard ratios (HRs) with 95% CIs were provided for PFS and odds ratios (ORs) with 95% CIs were reported for tumour response outcomes (ORR and resection rate). The results of individual studies were illustrated in forest plots. Between-study clinical heterogeneity was assessed qualitatively. There were insufficient studies to assess heterogeneity statistically using the chi-squared test or I² statistic.

What are the technical performance characteristics of the different KRAS mutation tests?

KRAS mutation test methods (Figure 2)

Fifteen laboratories stated that they used one method of KRAS mutation testing; one of these stated that they sometimes use a single KRAS mutation testing method and sometimes multiple methods (e.g. to confirm mutations).

FIGURE 2.

KRAS mutation tests used in NHS laboratories in the UK participating in the UK NEQAS pilot scheme for KRAS mutation testing.

Pyrosequencing, using in-house methods, was the most commonly used KRAS mutation test with nine laboratories using this approach, although one of the laboratories using pyrosequencing stated that it was in the process of switching to HRM analysis because of its quicker turnaround time. The cobas KRAS Mutation Test was used by three laboratories, Sanger sequencing was used by two laboratories and only a single laboratory used the Therascreen KRAS Pyro Kit. The final laboratory, which had initially reported using only a single method of KRAS mutation testing, stated that it used HRM analysis and direct sequencing. The laboratory that reported sometimes using multiple methods reported details for only one testing method (Sanger sequencing); however, this laboratory did state that it used Sanger sequencing in the event of an unusual pyrosequencing result. It also stated that its reason for using more than one testing method was the ability to fully characterise detected mutations. This suggests that its first choice method was in fact pyrosequencing not Sanger sequencing. This laboratory is therefore included in both methods in Figure 2 and the numbers above. All other laboratories stated that they used the reported KRAS mutation testing method for 100% of the samples.

Nine laboratories reported that samples were referred to their laboratory for testing on demand (one specified that this was MDT meetings, via pathologist, or oncologist), two reported mixed referral (some centres on demand, some all CRC samples), one laboratory reported that all resected primary CRC samples were sent for testing, one reported that samples were sent through clinical trials, one reported that all resected primary CRC plus metastatic samples were sent for testing and one laboratory did not answer this question.

The main reasons cited for choice of KRAS mutation testing method were mutation coverage (n = 13, 87%), ease of use (n = 12, 80%) and cost (n = 11, 73%). Nine laboratories (60%) also selected sensitivity (proportion of tumour cells required) and seven (47%) selected turnaround time. One laboratory did not answer this question. There was no apparent association between test method and reason for choice.

Of the eight laboratories that completed the questionnaire for pyrosequencing, all reported that they targeted mutations in codons 12, 13 and 61. Two of these laboratories reported that all mutations were targeted, one using commercial primers and one using self-designed primers. Two laboratories reported that they targeted specific mutations using self-designed primers. The others all used self-designed primers but did not state whether they targeted all or specific mutations; one laboratory stated that it also targeted mutations in codon 146. Two laboratories used Sanger sequencing. One stated that it targeted specific mutations but did not provide any further details. The other stated that it targeted mutations in codons 12, 13 and 61 using self-designed primers. One laboratory stated that it used a single testing method only, HRM analysis, and subsequently stated that it used HRM analysis and direct sequencing; mutations in codons 12, 13 and 61 and all mutations in exons 2 and 3 were targeted. Details on primers were not reported. The other four laboratories used commercial KRAS mutation testing kits.

KRAS mutation test logistics (Table 3 and Figure 3)

The number of samples screened for KRAS mutations in a typical week varied by laboratory, from less than five (three laboratories) to > 20 (four laboratories). The batch size ranged from less than one to two to 15–20 samples. Only laboratories with five or less samples screened per week ran batches of three or less. Only one laboratory had a batch size of > 15 (reported as 15–20 samples per week) and this laboratory screened > 20 samples per week; most other laboratories had batch sizes between five and 10. The two laboratories using Sanger sequencing both reported screening five or less samples per week and reported batch sizes of one or two. Of the four laboratories using commercial kits, one did not report on number of samples screened or batch size, two reported screening > 20 samples per week with batch sizes of 10 and 15, and one reported screening 6–10 samples per week with a similar batch size. Only one laboratory reported that it waited until it had a certain number of samples before running the KRAS mutation test; this laboratory waited until it had 10 samples before running the test.

TABLE 3 - Laboratory throughput by KRAS mutation test

NR, not reported.

KRAS mutation test	Samples per week	Batch size	Frequency of test	Wait for batch size?	Time from receiving sample to returning result to clinician
cobas KRAS Mutation Test Kit	6–10	6–10	Weekly	No	6–7 days
	> 20	10	Two to three times per week	Yes, 10	3–5 days
	NR	NR	NR	NR	NR
HRM analysis	6–10	4	Two to three times per week	No	3–5 days
Pyrosequencing	≤ 5	3	On demand	No	3–5 days
	6–10	6–10	Weekly	No	3–5 days
	> 20	15–20	Two to three times per week	No	3–5 days
	6–10	8	Weekly	No	6–7 days
	11–15	6–10	Two to three times per week	No	24–48 hours
	16–20	10	Two to three times per week	No	3–5 days
	6–10	5–10	Two to three times per week	No	3–5 days
	> 20	12	Daily	No	3–5 days
Sanger sequencing	≤ 5	≥ 1	Daily	No	3–5 days
	≤ 5	1–2 patients	Variable	No	3–5 days
Therascreen KRAS Pyro Kit	> 20	15	Two to three times per week	No	6–7 days

FIGURE 3.

Summary of logistic information.

Most laboratories reported an average waiting time from requesting the sample for KRAS testing to receiving the sample in the laboratory of 24–48 hours (three laboratories) or 3–5 days (six laboratories), although one laboratory reported a waiting time of < 24 hours and one reported a waiting time of 6–7 days. The range in waiting times was reported by four laboratories and was 1–10 days in two and 2–30 days in one, with the fourth stating that occasionally request dates are included in referral and so the range is 1–3 weeks. Four laboratories did not report data on time to receive samples at the laboratory once the sample had been requested. The majority of laboratories had a turnaround time from receiving the sample to reporting the result to the clinician of 3–5 or 6–7 days, with only one laboratory reporting a time of 24–48 hours. The laboratory with the shortest turnaround time was one that used pyrosequencing and tested 11–15 samples per week.

What is the accuracy of KRAS mutation testing for predicting response to treatment with cetuximab plus standard chemotherapy and subsequent resection rates?

One study, the CELIM trial,52 reported sufficient data to allow calculation of the accuracy of KRAS mutation testing for predicting response to treatment in patients with colorectal liver metastases who are treated with cetuximab plus FOLFOX6 or cetuximab plus FOLFIRI. This study is potentially useful in that it could provide full information on the extent to which KRAS mutation tests are able to discriminate between patients who will have benefit from the addition of cetuximab to standard chemotherapy regimens and those who will not. The utility of the study for this assessment is limited because reporting of outcome data by mutation status was limited to objective response. Thus, we defined TPs as those patients with KRAS wild-type tumours who have a positive response to treatment with cetuximab plus FOLFOX6 or cetuximab plus FOLFIRI (best observed response = CR or PR). FPs were defined as those patients with KRAS wild-type tumours who did not have a positive response to treatment with cetuximab plus FOLFOX6 or cetuximab plus FOLFIRI (SD or PD). FNs were defined as those with KRAS mutant tumours who had a positive response to treatment with cetuximab plus FOLFOX6 or cetuximab plus FOLFIRI. TNs were defined as those with KRAS mutant tumours who did not have a positive response to treatment with cetuximab plus FOLFOX6 or cetuximab plus FOLFIRI. Full definitions of CR, PR, SD and PD are provided in Chapter 2 (see Measuring response to treatment). The publication of the results of the CELIM trial included resection rates for liver metastases; however, these data were reported for all participants and by treatment group (cetuximab plus FOLFOX6 or cetuximab plus FOLFIRI) only and not by tumour KRAS mutation status. 52 For all study participants, the R0 resection rate was 36/106 (34%, 95% CI 25% to 44%) and the R0/R1 resection and/or radiofrequency ablation rate was 49/106 (46%, 95% CI 36% to 56%). 52 (Commercial-in-confidence information has been removed.)59 All participants in the CELIM trial received treatment with cetuximab in addition to standard chemotherapy; therefore, this trial could not contribute data to question 3 on how outcomes from treatment with cetuximab plus standard chemotherapy vary according to which test is used to select patients for treatment.

Additional data, supplied by the COIN trial investigators,54 allowed calculation of the accuracy of KRAS mutation testing, using a combination of pyrosequencing and MALDI-TOF and targeting mutations in codons 12, 13 and 61, for predicting response to treatment in patients with colorectal liver metastases who are treated with cetuximab plus FOLFOX or XELOX. These data could be viewed as being of limited applicability to this assessment because the standard chemotherapy regimen used in the COIN trial does not exactly match that in our inclusion criteria (some participants received XELOX rather than FOLFOX or FOLFIRI). However, the additional data supplied did allow the calculation of accuracy with respect to prediction of the more clinically relevant outcome of potentially curative resection. In this case we defined TPs as those patients with KRAS wild-type tumours who had a potentially curative resection following treatment with cetuximab plus FOLFOX or cetuximab plus XELOX. FPs were defined as those patients with KRAS wild-type tumours who did not have a potentially curative resection following treatment with cetuximab plus FOLFOX or cetuximab plus XELOX. FNs were defined as those with KRAS mutant tumours who had a potentially curative resection following treatment with cetuximab plus FOLFOX or cetuximab plus XELOX. TNs were defined as those with KRAS mutant tumours who did not have a potentially curative resection following treatment with cetuximab plus FOLFOX or cetuximab plus XELOX.

KRAS mutation test accuracy

Data from the CELIM trial52 provided estimates for the accuracy of the Therascreen KRAS RGQ PCR Kit for discriminating between patients who are likely to benefit from addition of cetuximab to standard chemotherapy regimens and those who are not. Sensitivity for the prediction of objective response was moderate (74.6%, 95% CI 62.1% to 84.7%) and specificity was poor (35.5%, 95% CI 19.2% to 54.6%) (Table 8 and Figure 4). Additional data supplied by the COIN trial investigators allowed the calculation of estimates for the accuracy of pyrosequencing and MALDI-TOF, targeting mutations in codons 12, 13 and 61, for predicting potentially curative resection following treatment with cetuximab plus FOLFOX or XELOX (D Fisher, personal communication). Sensitivity and specificity were both poor at 52.0% (95% CI 31.3% to 72.2%) and 45.6% (95% CI 37.0% to 54.3%) respectively (see Table 8 and Figure 4).

TABLE 8 - Accuracy of KRAS mutation testing for the prediction of response to treatment with cetuximab in addition to standard chemotherapy

NPV, negative predictive value; PPV, positive predictive value.

Additional data supplied by the COIN trial investigators (D Fisher, personal communication).

Study	KRAS mutation test method and mutations targeted	Non-evaluable samples	Reference standard	TP	FP	FN	TN	Sensitivity (95% CI) (%)	Specificity (95% CI) (%)	Prevalence (%)	PPV (95% CI) (%)	NPV (95% CI) (%)
Folprecht 201052 (CELIM)	Therascreen KRAS RGQ PCR Kit	12/111 unknown mutation status (KRAS mutation testing not carried out successfully, no reasons reported)	Objective response	47	20	16	11	74.6 (62.1 to 84.7)	35.5 (19.2 to 54.6)	67	70.1 (58.3 to 79.8)	40.7 (24.5 to 59.3)
COINa	Pyrosequencing and MALDI-TOF (codons 12, 13 and 61)		Resection rate	13	74	12	62	52.0 (31.3 to 72.2)	45.6 (37.0 to 54.3)	16	14.9 (8.9 to 23.9)	83.9 (73.8 to 90.5)

FIGURE 4.

Summary receiver operating characteristic plot showing estimates of sensitivity and specificity together with 95% CIs from the CELIM trial52 (black lines) and the COIN trial (green lines) (D FIsher, personal communication).

How do outcomes from treatment with cetuximab plus standard chemotherapy vary according to which test is used to select patients for treatment?

Four RCTs (six publications)27,28,53–56 provided data on the clinical effectiveness of cetuximab plus standard chemotherapy compared with standard chemotherapy alone in patients with colorectal liver metastases and no extrahepatic metastases whose tumours were KRAS wild type. The trials compared cetuximab in combination with standard chemotherapy (FOLFOX or FOLFIRI) with standard chemotherapy alone (see Table 11); in one trial54 standard chemotherapy could be either FOLFOX or XELOX. One trial55 included only participants with unresectable colorectal liver metastases and no extrahepatic metastases whose tumours were KRAS wild type. This trial was reported as a conference abstract only and some additional information on this trial was derived from the trial registry entry. 60 The remaining three trials, CRYSTAL,27,53 OPUS28,53,56 and COIN54 (five publications), included participants with mCRC, conducted tumour KRAS mutation testing in a subgroup of these participants and reported data for a smaller subgroup of participants whose metastases were confined to the liver; in all cases outcome data on participants whose metastases were confined to the liver were reported only for those with KRAS wild-type tumours.

Clinical outcomes in patients with colorectal metastases limited to the liver and KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy compared with clinical outcomes in those treated with standard chemotherapy

All studies in this section reported that the addition of cetuximab to standard chemotherapy was associated with an increase in the rate of R0 resections (Table 11); however, this increase reached statistical significance only in the trial by Xu et al. 55 (OR 4.57, 95% CI 1.56 to 13.34). All three studies that assessed ORR reported a statistically significant higher response rate for participants treated with cetuximab plus standard chemotherapy than for those treated with standard chemotherapy alone; ORs ranged from 3.00 (95% CI 1.49 to 6.03)53 to 4.93 (95% CI 1.42 to 17.06). 28 Only the COIN trial54 reported an improvement in PFS associated with the addition of cetuximab to standard chemotherapy (HR 0.68, 95% CI 0.48 to 0.97). The study by Xu et al. 55 reported a significant improvement in 3-year survival rates for participants treated with cetuximab plus standard chemotherapy compared with those treated with standard chemotherapy alone (OR 2.76, 95% CI 1.12 to 6.26). There were no clear differences in treatment effect regardless of which KRAS mutation test was used to identify participants whose tumours were KRAS wild type (Figures 5–7). The median PFS for participants with KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy was 11.8 months in the CRYSTAL trial and 11.9 months in the OPUS trial; the corresponding PFS values in the standard chemotherapy groups were 9.2 months and 7.9 months. 53 The median OS for participants with KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy was 27.8 months in the CRYSTAL trial and 26.3 months in the OPUS trial; the corresponding OS values in the standard chemotherapy groups were 27.7 months and 23.9 months. 53

TABLE 11 - Effectiveness of cetuximab plus standard chemotherapy compared with standard chemotherapy alone in patients with KRAS wild-type tumours and liver-limited metastases

OSR, 3-year survival rate; R0R, R0 resection rate.

Calculated value.

Study	KRAS test (mutations targeted)	Participant details	Intervention	Comparator	Outcome	Effect estimate (95% CI)
Bokemeyer 201128,53,56 (OPUS)	LightMix k-ras Gly12 assay (KRAS codons 12 and 13 missense mutations)	n = 337 KRAS wild type 179; KRAS wild type with liver-limited metastases 48; KRAS mutation status unknown 22/337 (no reason reported)	Cetuximab + FOLFOX-4 (n = 25)	FOLFOX-4 (n = 23)	PFS	HR 0.64 (0.23 to 1.79)
					ORR	OR 4.93 (1.42 to 17.06)a
					R0R	OR 4.19 (0.43 to 40.62)a
Van Cutsem 200927,53 (CRYSTAL)	LightMix k-ras Gly12 assay (KRAS codons 12 and 13 missense mutations)	n = 1198 KRAS wild type 348; KRAS wild type with liver-limited metastases 140; KRAS mutation status could not be evaluated 658/1198 (no reason reported)	Cetuximab + FOLFIRI (n = 68)	FOLFIRI (n = 72)	ORR	OR 3.00 (1.49 to 6.03)a
					R0R	OR 2.59 (0.76 to 8.86)a
Maughan 201154 (COIN)	Pyrosequencing and MALDI-TOF mass array with Sanger sequencing for discordant samples (< 1%) (KRAS mutations in codons 12, 13 and 61)	n = 1630 KRAS wild type 729; KRAS wild type with liver-limited metastases 178; KRAS mutation status unknown 336/1630 (141 tumour blocks not available, 173 blocks contained insufficent tumour material for processing, 22 not successfully genotyped)	Cetuximab + FOLFOX or XELOX (n = 87)	FOLFOX or XELOX (n = 91)	PFS	HR 0.68 (0.48 to 0.97)
					R0R	OR 1.16 (0.50 to 2.70)a
Xu 201255	Pyrosequencing (KRAS mutations in codons 12 and 13)	n = 116 KRAS wild type 116; KRAS wild type with liver-limited metastases 116	Cetuximab + FOLFIRI or FOLFOX6 (n = 59)	FOLFIRI or FOLFOX6 (n = 57)	OSR	OR 2.76 (1.12 to 6.26)a
					ORR	OR 3.90 (1.80 to 8.43)a
					R0R	OR 4.57 (1.56 to 13.34)a

FIGURE 5.

Progression-free survival in patients with colorectal metastases limited to the liver and KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy compared with those treated with standard chemotherapy.

FIGURE 6.

Objective response in patients with colorectal metastases limited to the liver and KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy compared with those treated with standard chemotherapy.

FIGURE 7.

R0 resection rate in patients with colorectal metastases limited to the liver and KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy compared with those treated with standard chemotherapy.

Search strategy

Searches were undertaken to identify cost-effectiveness studies of KRAS mutation testing in mCRC. As with the clinical effectiveness searching, the main EMBASE strategy for each set of searches was independently peer reviewed by a second information specialist using the PRESS-EBC. 43 Search strategies were developed specifically for each database and searches took into account generic and other product names for the intervention. All search strategies are reported in Appendix 1.

The following databases were searched for relevant studies from 2000 to January 2013:

• MEDLINE (OvidSP) (2000 to Week 3 January 2013)

• MEDLINE In-Process and Other Non-Indexed Citations and Daily Update (OvidSP) (up to 28 January 2013)

• EMBASE (OvidSP) (2000 to Week 4 2013)

• NHS Economic Evaluation Database (NHS EED) (Wiley) (The Cochrane Library 2000 to Issue 4, 2012)

• Health Economic Evaluations Database (HEED) (Wiley) (up to 30 January 2013), http://onlinelibrary.wiley.com/book/10.1002/9780470510933

• EconLit (EBSCO) (2000 to 30 January 2013)

• Science Citation Index (SCI-EXPANDED) (Web of Science) (2000 to 25 January 2013).

Inclusion criteria

Studies reporting a full economic analysis that related explicitly to the test–treat combination of KRAS mutation testing and treatment with cetuximab were eligible for inclusion. Specifically, one of the comparators included KRAS mutation testing and for this comparator the treatment decision was guided by the test result; patients whose tumour was KRAS mutant were also included in the treatment pathway.

Results

The search retrieved 445 references. Following title and abstract screening, 416 references were excluded. Of the remaining 29 titles, the full papers were screened, which led us to exclude another 24, leaving five references: one HTA report (from Ontario)61 and four papers. 62–65 A summary of the included studies is provided in Table 13 with a quality checklist based on Drummond et al. 66 provided in Table 14.

The Ontario HTA report61 aimed to determine the cost-effectiveness of KRAS mutation testing for the third-line treatment of (stage IV) mCRC in Ontario. For this purpose, seven strategies were compared:

• 0 best supportive care (BSC)

• 1(a) cetuximab with KRAS mutation testing

• 1(b) cetuximab without KRAS mutation testing

• 2(a) panitumumab with KRAS mutation testing

• 2(b) panitumumab without KRAS mutation testing

• 3(a) cetuximab plus irinotecan with KRAS mutation testing

• 3(b) cetuximab plus irinotecan without KRAS mutation testing.

In the strategies with KRAS mutation testing, only patients with wild-type KRAS tumours receive the therapy in question. In the strategies without KRAS mutation testing, all patients receive the therapy. A cost–utility analysis was performed by means of a Markov model with a lifetime time horizon. Inputs for PFS, OS, utility weights and adverse events were obtained from various clinical studies. Although a probabilistic sensitivity analysis (PSA) was performed for the BSC strategy and all KRAS mutation testing strategies, the information on uncertainty around the incremental cost-effectiveness ratio (ICER) presented in the report was limited to some percentages taken from the cost-effectiveness acceptability curve (CEAC) (which was not shown). Also, it appears that the parameters that were varied within the PSA were highly aggregated (PFS, OS, utility scores and total costs) and the distribution used was not mentioned (possibly uniform, as only a range was given). The deterministic results showed that, compared with BSC, all monotherapy strategies are either dominated or extendedly dominated. The ICER for cetuximab plus irinotecan with KRAS mutation testing compared with BSC is C$42,710. The ICER for cetuximab plus irinotecan without KRAS mutation testing compared with cetuximab plus irinotecan with KRAS mutation testing is C$163,396. The authors concluded that, although KRAS mutation testing is cost-effective for all strategies considered, it is not equally cost-effective for all treatment options.

Vijayaraghavan et al. 62 developed a Markov model to compare six hypothetical strategies for the second-line treatment of patients with mCRC who have failed previous chemotherapy:

1. combination therapy (cetuximab plus irinotecan/FOLFIRI) with KRAS mutation testing

2. combination therapy (cetuximab plus irinotecan/FOLFIRI) without KRAS mutation testing

3. cetuximab alone with KRAS mutation testing

4. cetuximab alone without KRAS mutation testing

5. panitumumab alone with KRAS mutation testing

6. panitumumab alone without KRAS mutation testing.

In treatment strategies without KRAS mutation testing, all patients received EGFR inhibitor-based chemotherapy, as did the patients with KRAS wild-type tumours in the treatment strategies with KRAS mutation testing. Patients with KRAS mutant tumours received chemotherapy without EGFR inhibitors or BSC. The model results were calculated for a situation in the USA as well as in Germany, with country-specific chemotherapy regimens and associated costs. Clinical effects were assumed to be the same for the USA and Germany and were based on published studies. In the results section the treatment strategies were compared as KRAS mutation testing compared with no KRAS mutation testing within a certain treatment regimen. For all of these comparisons KRAS mutation testing saved costs at equivalent clinical outcomes.

The cost-effectiveness of both KRAS and v-raf murine sarcoma viral oncogene homolog B (BRAF) mutation screening in patients with mCRC who are chemorefractory was the subject of a paper by Behl et al. 63 A decision-analytic model was developed comparing the following strategies:

1. KRAS plus BRAF mutation screening (before providing anti-EGFR therapy)

2. KRAS mutation screening (before providing anti-EGFR therapy)

3. anti-EGFR therapy (no screening)

4. no anti-EGRF therapy (no screening).

Inputs for the model were estimated using observations from RCTs. The model followed each patient for a maximum of 10 years. The no cetuximab strategy was least costly (US$34,291) but also least effective [0.6686 (LYs)], followed by the KRAS plus BRAF screening therapy, which offered more LYs (0.7025) but at a higher cost (US$56.324). Screening for KRAS mutations again added more LYs (0.7029) at a higher cost (US$57,348) but at an unfavourable ratio compared with KRAS plus BRAF mutation screening (ICER > US$2M). Finally, the anti-EGFR strategy – providing cetuximab to all patients without screening – was the most effective (0.7055 LYs) and the most expensive (US$64,841) strategy, which only proved cost-effective at a willingness to pay of > US$3M. Therefore, KRAS plus BRAF mutation screening appeared to be the most cost-effective option compared with no anti-EGFR therapy, but still had an ICER of US$648,396 per life-year gained.

Shiroiwa et al. 64 performed a cost-effectiveness analysis of KRAS testing in a Japanese population of patients with mCRC in whom previous chemotherapy had failed. The included strategies were KRAS mutation testing, no KRAS mutation testing (all patients receive cetuximab) and no cetuximab (all patients receive BSC). In the KRAS mutation testing strategy, patients with KRAS wild-type tumours received cetuximab whereas patients with KRAS mutant tumours received BSC. The analysis involved a three-state Markov model to estimate and extrapolate survival curves and treatment costs. Transition probabilities for disease progression and survival after disease progression were derived from the NCIC CO.17 trial. 67 The time horizon of the analysis was 2.5 years. As expected, the no cetuximab strategy was least costly (US$6800) and least effective [0.36 quality-adjusted life-years (QALYs)]. The no KRAS mutation testing strategy (all patients receive cetuximab) cost US$35,000 and resulted in 0.48 QALYs whereas the KRAS mutation testing strategy cost US$29,000 and resulted in 0.49 QALYs. Therefore, KRAS mutation testing is considered dominant compared with no KRAS mutation testing. The ICER for KRAS mutation testing compared with no cetuximab was US$180,000 per QALY. The authors concluded that, although KRAS mutation testing compared with no KRAS mutation testing could be considered dominant, the ICER for cetuximab treatment is too high, even if treatment is limited to patients with KRAS wild-type tumours.

Blank et al. 65 constructed a model comparing the cost-effectiveness of four strategies for chemorefractory patients with mCRC: KRAS mutation testing, KRAS mutation testing with subsequent BRAF mutation testing of KRAS wild-type tumours, cetuximab treatment without testing and the reference strategy of no cetuximab. In the testing strategies, cetuximab treatment was initiated if no mutations were detected. BSC was given to all patients. Survival times and utilities were derived from published RCTs. Costs were assessed from the perspective of the Swiss health system. Adding cetuximab to BSC increased costs considerably, but the increase in costs in the testing strategies was distinctly lower than that in the no testing strategy. The costs of mutation testing were overcompensated for by savings associated with the restriction of cetuximab administration. The least costly and least effective strategy was the reference strategy (no cetuximab). Testing for KRAS and BRAF mutations led to an ICER of €62,653 per QALY compared with the reference strategy. Testing for KRAS mutations only compared with testing for KRAS and BRAF mutations, as well as the no testing strategy compared with KRAS mutation testing, both had ICERs well above €300,000 per QALY. The authors concluded that testing for KRAS and BRAF mutations before cetuximab treatment of chemorefractory mCRC patients is clinically appropriate and economically favourable, despite high costs for predictive testing [i.e. given that cetuximab should be the next step in the treatment pathway, it is worthwhile to test for mutations first; it appears that the reference strategy (no cetuximab) was not included in this recommendation].

Based on all of these publications it can be said that, in general, although KRAS testing is obviously a more cost-effective option than administering cetuximab to all patients, the ICER of KRAS testing and treating with cetuximab only those patients with KRAS wild-type tumour status compared with treating all patients with standard chemotherapy alone seems rather high.

KRAS mutation tests considered in the model

The health economic analysis will determine the cost-effectiveness of different methods for KRAS mutation testing to decide between standard chemotherapy or standard chemotherapy plus cetuximab in adults with mCRC and a resected or resectable primary tumour, whose metastases are confined to the liver and are unresectable but may become resectable after response to chemotherapy. Standard chemotherapy regimens considered include FOLFOX and FOLFIRI. A range of methods for KRAS mutation testing are currently used in NHS laboratories in England and Wales. Ideally, the performance of these tests would be assessed against an objective measure of the true presence/absence of a clinically relevant KRAS mutation (the ‘reference standard’). The comparative effectiveness of treatment (cetuximab plus chemotherapy vs. chemotherapy alone) conditional on the true or false presence/absence of the KRAS mutation could then be determined. However, each testing method targets a different range of mutations and has different limits of detection (the lowest proportion of mutation detectable in tumour cells) and the exact combination of mutation type and level that will provide optimal treatment selection remains unclear. For this reason, assessment of test performance based on comparison with a conventional ‘reference standard’ is currently not possible. In this situation, an alternative way to determine the relative value of diagnostic methods for KRAS mutation testing is to use studies that report on the comparative treatment effect in patients with different KRAS mutation status (positive, negative or unknown), as defined using different KRAS mutation tests. As outlined in the previous chapter, information on the accuracy of tests (either based on ORR or tumour resection rate) for distinguishing between patients with KRAS wild-type tumours and patients with KRAS mutant tumours with metastases confined to the liver was available only for the Therascreen KRAS RGQ PCR Kit52 and pyrosequencing and MALDI-TOF mass spectometry. 54 A major assumption underlying the use of these accuracy data in the health economic modelling is that differences in response rates and resection rates between the two trials are solely due to the use of different KRAS mutation tests.

In the COIN trial54 patients were tested using both pyrosequencing and the MALDI-TOF mass array, with a reported concordance of > 99%. It was therefore assumed that, for the economic evaluation, MALDI-TOF and pyrosequencing are equal, that is, all results reported for pyrosequencing also apply to MALDI-TOF. However, survey data were available only for pyrosequencing and therefore pyrosequencing data are reported in the results tables.

For all other KRAS mutation tests listed in the scope, no accuracy data were available. As a result, for the remaining tests it was only possible to make a comparison based on differences in technical performance and test costs retrieved from the online survey of NHS laboratories in England and Wales (see Chapter 3, What are the technical performance characteristics of the different KRAS mutation tests?), whilst assuming a prognostic value equal to pyrosequencing across all tests. The latter assumption was not based on evidence of equality but rather on the absence of any reliable evidence to model a difference in prognostic value for these tests.

Based on the information available, two analyses were performed:

1. ‘Linked evidence’ analysis – for all tests for which information on accuracy was available. In this analysis the Therascreen KRAS RGQ PCR Kit was compared with pyrosequencing. For the Therascreen KRAS RGQ PCR Kit, accuracy based on ORRs was taken from the CELIM trial. 52 Resection rates for patients with a KRAS wild-type test result treated with cetuximab plus chemotherapy were also based on CELIM trial data (data not reported in the published study but provided as part of the submissions that informed NICE guidance TA1761) whereas resection rates for patients with KRAS mutant and unknown test results who were treated with chemotherapy alone were taken from the Groupe Coopérateur Multidisciplinaire en Oncologie (GERCOR) study68 as the CELIM trial did not contain a chemotherapy-only strategy. For pyrosequencing, both accuracy data (based on resection rates) and resection rates for cetuximab plus chemotherapy as well as chemotherapy alone were taken from the COIN trial54 and from additional data supplied by the COIN triallists. PFS and OS after successful resection were assumed to be conditional on resection and treatment independent.

2. ‘Assumption of equal prognostic value’ analysis – for all tests for which information on technical performance was available from the online survey. In this analysis we assessed whether the tests were likely to be cost-effective given an assumption of equal prognostic value and test-specific information on failure rate only. The equal prognostic value assigned was based on data for the pyrosequencing test (as this was the only test for which accuracy data were available on resection rates following treatment with chemotherapy, with and without cetuximab, for patients with initially inoperable liver metastases and both KRAS mutant and KRAS wild-type tumours). The following tests were included in this analysis:

   1. cobas KRAS Mutation Test Kit (Roche Molecular Systems)

   2. Therascreen KRAS RGQ PCR Kit (QIAGEN)

   3. Therascreen KRAS Pyro Kit (QIAGEN)

   4. KRAS LightMix Kit (TIB MOLBIOL)

   5. KRAS StripAssay (ViennaLab)

   6. HRM analysis

   7. pyrosequencing

   8. MALDI-TOF mass spectrometry

   9. next-generation sequencing

   10. Sanger sequencing.

Consistency with related assessments

This assessment does not update the appraisal of cetuximab for the first-line treatment of mCRC. 1 To ensure consistency between the modelling approach used in technology appraisal TA1761 and the assessment of the cost-effectiveness of different methods for KRAS mutation testing in this report, the assessment group received the electronic health economic model submitted by Merck Serono for technology appraisal TA176. This model calculates the expected cost-effectiveness of chemotherapy plus cetuximab compared with chemotherapy for the first-line treatment of mCRC patients whose metastases are confined to the liver and are unresectable and whose tumours are KRAS wild type as tested with a pre-CE-marked version of the KRAS LightMix Kit.

This model, together with the amendments suggested and made by the Evidence Review Group and NICE, was used to inform the development of a de novo model in which the long-term consequences of using different KRAS mutation tests were assessed not only in patients with KRAS wild-type tumours but also in patients with KRAS mutant tumours or an unknown test result.

Model structure

In the health economic model the mean expected costs and QALYs were calculated for each alternative. As specified in the protocol, this economic evaluation takes a ‘no comparator’ approach, which implies that the cost-effectiveness of each strategy will be presented compared with the next most cost-effective strategy.

The health economic analysis considers the long-term consequences of the technical performance and accuracy of the different tests followed by treatment with cetuximab plus standard chemotherapy or standard chemotherapy alone in patients (average starting age 60 years) with mCRC whose metastases are confined to the liver and are unresectable. For this purpose a decision tree and a Markov model were developed. The decision tree was used to model the test result (KRAS wild type, KRAS mutant or unknown) and the accompanying treatment decision. In the model, patients with a KRAS wild-type tumour receive cetuximab plus standard chemotherapy. It is assumed that patients with a KRAS mutant tumour will receive standard chemotherapy (i.e. FOLFOX). Patients with an unknown KRAS status are also assumed to receive standard chemotherapy as cetuximab is indicated only for patients with KRAS wild-type tumour status. 1 The decision tree is shown in Figure 8.

FIGURE 8.

Decision tree structure. Ctx, chemotherapy.

The long-term consequences in terms of costs and QALYs were estimated using a Markov model with a cycle time of 1 week and a lifetime time horizon (23 years were modelled using 1200 cycles). Health states in the Markov model are numbered according to NICE technology appraisal TA176:1

1. progression-free first line – never operated

2. PD second line – never operated

3. PD second line – unsuccessful resection

4. survival after curative resection

5. progression-free first line – unsuccessful resection

6. PD third line – never operated

7. PD third line – unsuccessful resection

8. dead.

The Markov model structure is shown in Figure 9. The model is described in more detail in NICE technology appraisal TA176. 1

FIGURE 9.

Markov model structure.

Model parameters

Estimates for model input parameters were retrieved from NICE technology appraisal TA1761 and the manufacturer’s submission for TA176,59,69 the assessment of the clinical effectiveness of different KRAS mutation tests (see Chapter 3, What is the accuracy of KRAS mutation testing for predicting response to treatment with cetuximab plus standard chemotherapy and subsequent resection rates?) and an online survey of NHS laboratories in England and Wales (see Chapter 3, What are the technical performance characteristics of the different KRAS mutation tests?).

Overview of the main model assumptions

The main assumptions in the health economic analysis were:

1. The differences between ORRs and resection rates for cetuximab plus chemotherapy compared with chemotherapy alone reported in the CELIM trial52 combined with the GERCOR trial68 and those reported in the COIN trial54 are solely the result of the different tests used (Therascreen KRAS RGQ PCR Kit and pyrosequencing respectively) to distinguish between patients whose tumours are KRAS wild type (and who receive cetuximab) and patients whose tumours are KRAS mutant (and who receive chemotherapy) (linked evidence analysis).

2. To calculate the sensitivity and specificity of the tests, required to calculate the proportion of KRAS wild-type and KRAS mutant test results (see Table 8), patients tested as KRAS wild type were categorised as FP if no objective response was observed (for the Therascreen KRAS RGQ PCR Kit) or no liver resection was performed (for pyrosequencing) after treatment with cetuximab whereas patients were categorised as TP if an objective response was observed (Therascreen KRAS RGQ PCR Kit) or a liver resection was performed (pyrosequencing). Similarly, patients tested as tumour KRAS mutant were categorised as FN if an objective response was observed (for the Therascreen KRAS RGQ PCR Kit) or a liver resection was performed after treatment with cetuximab (for pyrosequencing) whereas patients were categorised as TN if no objective response was observed (Therascreen KRAS RGQ PCR Kit) or no liver resection was performed (pyrosequencing).

3. Test accuracy based on objective response can be compared with accuracy based on resection rates. 39

4. The number of patients with unknown mutation status relative to the number of patients for whom a tissue sample was available in the trials52,54 provides a realistic approximation of the proportion of patients with an unknown test result in clinical practice (both analyses).

5. As the COIN trial54 tests for KRAS mutations using both pyrosequencing and MALDI-TOF, with a reported concordance of > 99%, it was assumed that the accuracy derived from this trial and also the resection rates reported apply to both pyrosequencing and MALDI-TOF, that is, all pyrosequencing results in this report also apply to MALDI-TOF.

6. The standard chemotherapy used in the COIN trial54 (FOLFOX or XELOX) is comparable to FOLFOX6 as used in the CELIM trial. 52

Sensitivity analyses

For both the linked evidence and the assumption of equal prognostic value analysis, the following sensitivity analyses were performed:

• mortality in the second line was based on the average of the first- and third-line mortality instead of background mortality as in NICE technology appraisal TA176. 1

• the proportion of patients with unknown mutation status was based on the results of the online survey instead of the literature (see Table 5).

Linked evidence analysis

The linked evidence analysis includes two tests, that is, only those tests for which evidence on test accuracy based on either resection rate or ORR was available. Table 22 shows the probabilistic results of this analysis. It should be noted that this analysis was based on a number of substantial assumptions, which are outlined in the previous section. In short, we have only the COIN54 and CELIM52 trials to rely on, of which the COIN trial used pyrosequencing to test for KRAS mutations and the CELIM trial used the Therascreen KRAS RGQ PCR Kit. We assumed that the differences between the outcomes of these trials are exclusively caused by the different tests used (assumption 1). Table 23 provides a summary of the comparability of the study populations across the COIN,54 CELIM52 and GERCOR68 trials used in the linked evidence analysis. In addition, we assumed that all KRAS wild-type patients would respond perfectly to cetuximab – or would all have a liver resection after cetuximab – and that all KRAS mutant patients would not (assumption 2), and that test accuracy based on objective response can be compared with accuracy based on resection rates (assumption 3).

As is apparent from Table 22, pyrosequencing has the lowest total cost. The Therascreen KRAS RGQ PCR Kit is the more expensive but also more effective strategy, with an ICER of £17,019 per QALY gained. The CEAC in Figure 14 shows that, for lower values of the threshold, pyrosequencing is the preferred strategy and that at thresholds of ≥ £17,000 the Therascreen KRAS RGQ PCR Kit is the most cost-effective option. The results of the sensitivity analyses (see Table 22) do not differ substantially from the base-case results in the sense that the Therascreen KRAS RGQ PCR Kit is consistently more expensive and more effective than pyrosequencing, with ICERs ranging from £14,860 to £20,528 per QALY gained. CEACs for the sensitivity analyses are presented in Appendix 6.

FIGURE 14.

Cost-effectiveness acceptability curve for the linked evidence analysis: base case.

Assumption of equal prognostic value analysis

The assumption of equal prognostic value analysis includes all tests for which information on technical performance was available from the online survey of NHS laboratories in England and Wales. This includes the tests for which accuracy data, based on either response rates or resection rates, were not available. Therefore, this analysis assessed whether the tests were likely to be cost-effective given an assumption of equal prognostic value based on the prognostic value of testing with pyrosequencing (as this was the only test for which full data were available on resection rates following treatment with chemotherapy, with and without cetuximab, for patients with initially inoperable liver metastases and both KRAS mutant and KRAS wild-type tumours) and test-specific information on technical failures within the laboratory only (Table 24). In the base case and the first sensitivity analysis, the total technical failure rate (pre-laboratory plus within-laboratory technical failures) is assumed to be equal for all tests. As a result, the strategies in these analyses differ only with respect to costs (because of differences in within-laboratory technical failures). In the base case the average QALYs for all comparators were 1.48 (95% CI 1.33 to 1.64). The total costs associated with the various testing strategies (see Table 24) are highly similar. The same applies to the first sensitivity analysis (Table 25): costs are similar across strategies and average QALYs are equal by assumption at 1.28 (95% CI 1.12 to 1.44).

In the second sensitivity analysis the total technical failure rate is also test specific, which impacts on the proportion of patients with unknown (and therefore also wild-type and mutant) tumour KRAS status. Therefore, in this sensitivity analysis the strategies differ with respect to both effects and costs. All other input parameters, such as test costs and test accuracy, are still considered equal. The probabilistic results in Table 26 show that the cobas KRAS Mutation Test is the least costly and least effective strategy. HRM analysis and Sanger sequencing have equal costs and effects and their ICER compared with the cobas KRAS Mutation Test is £69,815 per QALY gained. Pyrosequencing and the Therascreen KRAS RGQ PCR Kit are ruled out by extended dominance in this analysis. From the CEAC (Figure 15) it is apparent that the cobas KRAS Mutation Test is the preferred strategy for all threshold values of < £60,000.

FIGURE 15.

Cost-effectiveness acceptability curve for assumption of equal prognostic value sensitivity analysis: proportion of patients with unknown mutation status based on the survey.

Clinical effectiveness

There was no clear evidence to suggest any differences between the KRAS mutation testing techniques for any of the measures assessed (technical performance, accuracy for predicting response to treatment with cetuximab in combination with standard chemotherapy, or variation in clinical outcomes following treatment with cetuximab in combination with standard chemotherapy depending on which method is used to classify patients as having KRAS wild-type tumours).

The survey of laboratories providing KRAS mutation testing indicated that in-house pyrosequencing methods, targeting KRAS mutations in codons 12, 13 and 61 and using self-designed primers, were the most commonly used approaches (9/15 respondents); reasons cited by respondents for their choice of this technique were the proportion of tumour cells required, ease of use, cost, mutations covered, turnaround time and experience of pyrosequencing techniques available in the laboratory. There was no apparent association between test method and reason for choice. Commercial kits used were the cobas KRAS Mutation Test (three laboratories) and the Therascreen KRAS Pyro Kit (one laboratory). More than half of the responding laboratories reported that KRAS mutation testing was carried out on request (e.g. from a pathologist or oncologist); only one laboratory reported routine testing of all CRC samples. In general, there was no clear indication that choice of test method was related to volume of throughput, although both of the laboratories that reported using Sanger sequencing had a low throughput (up to five samples per week). Most respondents reported turnaround times, from receipt of sample to reporting to the clinician, of between 3 and 5 days. The only laboratory to report a turnaround time of < 3 days (24–48 hours) used an in-house pyrosequencing method. Frequency of running the test did not appear to relate to laboratory throughput and only one laboratory reported waiting for a minimum batch size (10 samples) before running the test; this laboratory had a high throughput (> 20 samples per week). The minimum percentage of tumour cells required for testing varied widely across laboratories (< 1% to > 30%), even when the same test method was being used. When reported, the minimum requirement for the cobas KRAS Mutation Test was ≤ 10%. With the exception of those using Sanger sequencing, all laboratories reported a limit of detection for percentage mutation of ≤ 10%. The laboratory that used the Therascreen KRAS Pyro Kit did not provide any data on technical performance. The proportion of samples rejected before analysis was < 2% for all responding laboratories. The rate of failure for analysed samples did not appear to be dependent on test method (3–6% for the cobas KRAS Mutation Test and 0.2–10% for in-house pyrosequencing methods). The majority of responding laboratories reported using microdissection techniques before DNA extraction; however, there was no clear indication that non-use of this technique was associated with a higher rate of sample rejection or test failure. The laboratory that used the Therascreen KRAS Pyro Kit did not provide any data on failure rates. Although most respondents included cost in their reasons for choosing a particular test, it is worth noting that a relatively narrow range of costs was reported across all tests (£100–150), with one laboratory reporting a higher cost (£273) for running a single sample. Prices charged, to both Merck Serono and the NHS, ranged from £99 to £150.

When contacted by NICE in relation to a previous diagnostic assessment of EGFR mutation testing in non-small-cell lung cancer, UK NEQAS stated that ‘Error rates are not always method related and it is not always possible to obtain data from all the labs committing critical genotyping errors. Therefore, any data which could be provided would be skewed with processing and reporting issues rather than being method related.’ Only one KRAS mutation testing method is currently approved by the US FDA; this is the Therascreen KRAS RGQ PCR Kit when used with the QIAamp DSP DNA FFPE Tissue Kit and the QIAGEN Rotor-Gene Q MDx (software version 2.1.0) and KRAS Assay Package. 16 The clinical trial67 used to support FDA approval was not included in this assessment as it did not match our inclusion criteria; it compared treatment with cetuximab and BSC to treatment with BSC alone in patients with mCRC who had previously failed all available chemotherapy. It should be noted that none of the laboratories participating in the UK NEQAS scheme who responded to our survey reported using the Therascreen KRAS RGQ PCR Kit.

Evidence to allow comparison of the accuracy of different KRAS mutation tests was very limited. Only one publication,52 from the CELIM trial, provided sufficient data to allow estimation of the accuracy of a KRAS mutation test (version 1 of the Therascreen KRAS PCR Kit) for predicting response to treatment with cetuximab plus standard chemotherapy. This study reported objective response data and thus did not provide direct information on the value of the KRAS mutation test for predicting resection rate. Because the aim of KRAS mutation testing is to predict likely response to the addition of cetuximab to standard chemotherapy, test positive was defined as a KRAS wild-type tumour. The positive predictive value (70.1%, 95% CI 58.3% to 79.8%), reported in Chapter 3 (see What is the accuracy of KRAS mutation testing for predicting response to treatment with cetuximab plus standard chemotherapy and subsequent resection rates?), indicated that KRAS wild type, as determined using the Therascreen KRAS PCR Kit, may be moderately predictive of tumour response. If the published strong correlation between ORRs and resection rates in patients with isolated liver metastases39 treated with various chemotherapy regimens were assumed to extrapolate to patients with KRAS wild-type tumours treated with standard chemotherapy plus cetuximab, then the expected R0 and R1 resection rates for these patients would be approximately 67%, based on data from the CELIM trial. 52 By contrast, the negative predictive value (40.7%, 95% CI 24.5% to 59.3%) could be interpreted as indicating that the presence of a KRAS mutation, as determined using the Therascreen KRAS PCR Kit, is a relatively poor predictor of non-response. Additional data supplied by the COIN trial investigators54 allowed the calculation of estimates for the accuracy of pyrosequencing and MALDI-TOF (in which both tests were performed on all samples), targeting mutations in codons 12, 13 and 61, for predicting potentially curative resection following treatment with cetuximab plus FOLFOX or XELOX. The positive and negative predictive values derived from these data were 14.9% (95% CI 8.9% to 23.9%) and 83.9% (95% CI 73.8% to 90.5%) respectively; this could be interpreted as indicating that a tumour which is defined as KRAS wild type by this method is a poor predictor of resectability following treatment with cetuximab plus standard chemotherapy, whereas the presence of a KRAS mutation is a good predictor of non-response (tumour remaining unresectable after treatment). The COIN trial reported > 99% concordance on KRAS genotyping between pyrosequencing and MALDI-TOF; it may therefore be assumed that accuracy data from the COIN trial are also representative of the accuracy of both pyrosequencing and MALDI-TOF when used as single tests. It should be noted that any apparent differences in the ability of KRAS mutation tests to predict response to treatment between the CELIM trial and the COIN trial may be caused by other differences between studies (e.g. participant characteristics, in particular the definition of baseline unresectability, and treatment regimens).

Four further studies (six publications27,28,53–56) were included in the review; all were RCTs comparing cetuximab plus standard chemotherapy with standard chemotherapy alone in patients whose tumours were KRAS wild type and all reported data on patients with CRC metastases that were confined to the liver. The standard chemotherapy regimen was different in each of the four trials (FOLFOX4,28,53,56 FOLFIRI,27,53 FOLFIRI or FOLFOX655 and FOLFOX or XELOX54). There was no substantial evidence to indicate a significant difference in treatment effect depending on which of three KRAS mutation tests (LightMix k-ras Gly12, pyrosequencing and MALDI-TOF mass array for mutations in codons 12, 13 and 61, or pyrosequencing for KRAS mutations in codons 12 and 13) was used to identify patients with KRAS wild-type tumours. All three studies that assessed ORR reported a statistically significant higher response rate for participants treated with cetuximab plus standard chemotherapy than for those treated with standard chemotherapy alone; ORs ranged from 3.00 (95% CI 1.49 to 6.03)53 to 4.93 (95% CI 1.42 to 17.06). 28 All four studies reported that the addition of cetuximab to standard chemotherapy was associated with an increase in the rate of R0 resections following treatment. However, it should be noted that the only trial to report a statistically significant treatment effect for R0 resection rate used pyrosequencing to identify KRAS mutations in codons 12 and 13 only. 55 This was also the only trial in which all participants had CRC metastases that were limited to the liver.

Effectiveness data from the CRYSTAL27 and OPUS28 trials were used to inform the technology appraisal underpinning NICE guidance TA1761 on cetuximab for the first-line treatment of mCRC. Data from an interim analysis of the CELIM trial were used as a source of UK data for resection rates following treatment with cetuximab plus standard chemotherapy. 1 Data from the COIN trial54 and the trial by Xu et al. 55 were published subsequently to TA176.

Cost-effectiveness

The review of economic analyses of different methods for KRAS mutation testing to decide between standard chemotherapy and cetuximab in combination with standard chemotherapy in adults with mCRC found four full papers62–65 and one HTA report. 61 Based on all of these publications it can be said that, in general, although KRAS testing is obviously a more cost-effective option than administering cetuximab to all patients, the ICER of KRAS testing and treating only patients with KRAS wild-type tumours with cetuximab compared with treating all patients with standard chemotherapy alone seems rather high.

In the health economic analysis, the cost-effectiveness of different methods of KRAS mutation testing to decide between standard chemotherapy and cetuximab in combination with standard chemotherapy in adults with mCRC was assessed. In light of the limited amount of evidence available, two analyses were performed: ‘linked evidence’ and ‘assumption of equal prognostic value’. All analyses took a ‘no comparator’ approach.

In the linked evidence analysis, the Therascreen KRAS RGQ PCR Kit was compared with pyrosequencing, using the available objective response and resection rate, respectively, to estimate lifetime costs and QALYs. The results of this analysis suggested that the Therascreen KRAS RGQ PCR Kit was more costly and more effective than pyrosequencing, with an ICER of £17,019 per QALY gained. Sensitivity analyses did not show substantial differences compared with the base case. The key driver behind the outcome was the difference in resection rate between treatment with and treatment without cetuximab and the proportions of patients with KRAS wild-type, KRAS mutant and unknown tumours. This was determined by test accuracy and therefore, for the most part, was dependent on ORR (for the Therascreen KRAS RGQ PCR Kit) or resection rate (for pyrosequencing).

It should be noted that this analysis was based on a number of substantial assumptions, which are outlined in Chapter 4 (see Overview of main model assumptions). The following assumptions used were particularly problematic as they are open to doubt and probably have a considerable impact on the model results:

• The differences between ORRs and resection rates for cetuximab plus chemotherapy compared with chemotherapy alone as reported in the CELIM trial52 combined with the GERCOR trial68 and the COIN trial54 are solely the result of the different tests used (Therascreen KRAS RGQ PCR Kit and pyrosequencing respectively) to distinguish between patients whose tumours are KRAS wild type (and who receive cetuximab) and patients whose tumours are KRAS mutant (and who receive chemotherapy).

• To calculate the sensitivity and specificity of the tests, required to calculate the proportion of KRAS wild-type and KRAS mutant test results (see Table 8), patients with a tumour tested as KRAS wild type were categorised as FP if no objective response was observed (for the Therascreen KRAS RGQ PCR Kit) or no liver resection was performed (for pyrosequencing) after treatment with cetuximab, whereas patients were categorised as TP if an objective response was observed (Therascreen KRAS RGQ PCR Kit) or a liver resection was performed (pyrosequencing). Similarly, patients with a tumour tested as KRAS mutant were categorised as FN if an objective response was observed (for the Therascreen KRAS RGQ PCR Kit) or a liver resection was performed (for pyrosequencing) after treatment with cetuximab, whereas patients were categorised as TN if no objective response was observed (Therascreen KRAS RGQ PCR Kit) or no liver resection was performed (pyrosequencing).

The results of the linked evidence analysis should therefore be interpreted on the condition that these assumptions hold. Moreover, the uncertainty presented surrounding the results is an underestimation of the true uncertainty as the uncertainty associated with the assumptions was not parameterised in the model and is therefore not reflected in the PSA.

The assumption of equal prognostic value analysis included all tests for which information on technical performance was available from the online survey of NHS laboratories in England and Wales. This includes the tests for which accuracy data based on either response rates or resection rates were not available. Therefore, this analysis assessed whether the tests were likely to be cost-effective given an assumption of equal prognostic value based on the prognostic value of testing with pyrosequencing (as this was the only test for which full data were available on resection rates following treatment with chemotherapy, with and without cetuximab, for patients with initially inoperable liver metastases and both KRAS mutant and KRAS wild-type tumours) and test-specific information on technical failures within the laboratory only, which implies that strategies can differ only with respect to costs. The results of the assumption of equal prognostic value analysis indicated that the strategies were almost equal. The first sensitivity analysis confirmed this. The second sensitivity analysis, in which the proportion of patients with unknown mutation status was taken from the survey instead of the literature, was slightly different in the sense that for this analysis the effectiveness was not assumed equal among all tests and therefore ICERs were available. The results showed that the cobas KRAS Mutation Test was the least expensive and least effective strategy and that Sanger sequencing and HRM analysis were equally the most costly and most effective, with an ICER of £69,815 per QALY gained compared with the cobas KRAS Mutation Test. The other two strategies included in this analysis, that is, the Therascreen KRAS RGQ PCR Kit and pyrosequencing, are ruled out by extended dominance.

Clinical effectiveness

Extensive literature searches were conducted in an attempt to maximise retrieval of relevant studies. These included electronic searches of a variety of bibliographic databases as well as screening of clinical trial registers and conference proceedings to identify unpublished studies. Because of the known difficulties in identifying test accuracy studies using study design-related search terms,76 and the potential need to include non-RCTs, search strategies were developed to maximise sensitivity at the expense of specificity. Thus, large numbers of citations were identified and screened, very few of which met the inclusion criteria of the review. The specificity of searches was further reduced as it was not possible to target publications focusing on patients whose metastases were limited to the liver only; these patients were a subgroup in the majority of included studies.

The possibility of publication bias remains a potential problem for all systematic reviews. Considerations may differ for systematic reviews of test accuracy studies. It is relatively simple to define a positive result for studies of treatment, for example a significant difference between the treatment group and the control group that favours treatment. This is not the case for test accuracy studies, which measure agreement between the index test and the reference standard. It would seem likely that studies finding greater agreement (high estimates of sensitivity and specificity) will be published more often. This distinction may be less applicable to studies in this review that provided accuracy data, as these studies aimed to assess the effectiveness of treatment with cetuximab plus standard chemotherapy in different patient groups (CELIM trial52) or compare the effectiveness of cetuximab plus standard chemotherapy and standard chemotherapy alone (COIN trial54); neither study was primarily focused on test performance. Our review included a very small number of clinically heterogeneous studies, both for the accuracy of KRAS mutation testing to predict response to treatment with cetuximab plus standard chemotherapy and for the relative effectiveness of cetuximab plus standard chemotherapy and standard chemotherapy alone in populations with KRAS wild-type tumours, selected using different KRAS mutation test methods. We were therefore unable to undertake any meta-analyses or formal assessment of publication bias. However, our search strategy included a variety of routes to identify unpublished studies and resulted in the inclusion of a number of conference abstracts.

Clear inclusion criteria were specified in the protocol for this review and only one protocol modification occurred during the assessment. The eligibility of studies for inclusion is therefore transparent. In addition, we have provided specific reasons for excluding any studies considered potentially relevant at initial citation screening (see Appendix 5). The review process followed recommended methods to minimise the potential for error and/or bias. 40 Studies were independently screened for inclusion by two reviewers and data extraction and quality assessment were carried out by one reviewer and checked by a second (MW and PW). Any disagreements were resolved by consensus.

Studies included in this review were assessed for risk of bias using published tools appropriate to the study design and/or the type of data extracted. Studies that provided data on the accuracy of KRAS mutation testing for predicting response to treatment with cetuximab plus standard chemotherapy were assessed using a modification of the QUADAS-2 tool. 48 The QUADAS-2 tool is structured into four key domains covering participant selection, index test, reference standard and the flow of patients through the study (including timing of tests). Each domain is rated for risk of bias (low, high or unclear); the participant selection, index test and reference standard domains are also separately rated for concerns regarding the applicability of the study to the review question (low, high or unclear). Studies that provided data on the effectiveness of treatment with cetuximab plus standard chemotherapy compared with standard chemotherapy alone in patients with KRAS wild-type tumours were all RCTs or subgroup analyses from RCTs. These studies were therefore assessed using The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. 42,47 The results of the risk of bias assessment are reported in full for all included studies in Appendix 3 and in summary form in Chapter 3 (see What is the accuracy of KRAS mutation testing for predicting response to treatment with cetuximab plus standard chemotherapy and subsequent resection rates? and How do outcomes from treatment with cetuximab plus standard chemotherapy vary according to which test is used to select patients for treatment?). The main potential sources of bias identified were exclusion of withdrawals from the analyses (for studies providing data on the accuracy of KRAS mutation tests for predicting response to treatment with cetuximab plus standard chemotherapy) and blinding of participants and personnel in treatment trials. Both of the studies that provided data on the accuracy of KRAS mutation testing for predicting response to treatment had some limitations in their applicability to the target population for this assessment. In the case of the CELIM trial52 data were available to calculate accuracy for the prediction of objective response only rather than for the preferred direct measure resection of liver, and in the case of the COIN trial54 the standard chemotherapy regimen did not fully match that in the inclusion criteria for this assessment. In addition, participants in the CELIM trial52 were described as having technically non-resectable or five or more liver metastases from CRC and it was therefore unclear whether some participants may have had potentially resectable metastases at baseline.

All of the studies included in this review have some limitations with regard to their ability to address the overall aim of comparing the clinical effectiveness of different KRAS mutation tests to determine which patients may benefit from the addition of cetuximab to standard chemotherapy and which should receive standard chemotherapy alone. The COIN trial54 is likely to represent the closest approximation to the ideal study in that, when additional data supplied by the trial investigators are also considered, it provides full information on the comparative treatment effect (cetuximab plus standard chemotherapy vs. standard chemotherapy alone) for patients with both KRAS wild-type tumours and KRAS mutant tumours. In addition, the trial was conducted in the UK and hence provides data that are likely to be directly applicable to practice in the NHS in England and Wales. However, data included in this assessment were derived from subgroup analyses of patients included in the original trial; not all patients included in the original trial had samples available for KRAS mutation testing and, in addition, a much smaller subgroup of patients had metastases that were limited to the liver. Further, the standard chemotherapy regimen used in the COIN trial54 allowed a choice between FOLFOX or XELOX (depending on local hospital practice and patient preference); the use of XELOX as standard chemotherapy does not match the standard chemotherapy described in the inclusion criteria for this assessment, as determined by the recommendations of NICE technology appraisal TA176. 1 The COIN trial used a combination of pyrosequencing and MALDI-TOF, targeting mutations in codons 12, 13 and 61, to determine KRAS mutation status; in common with all other studies included in this assessment, the study was not designed to assess KRAS mutation testing and did not provide any comparative data for other testing methods.

Because methods of testing for KRAS mutation status can differ in terms of both the mutations targeted and the limit of detection (the lowest proportion of tumour cells with a mutation that can be detected), the definitions of KRAS wild type and KRAS mutant vary according to which test is used. All testing methods are essentially reference standard methods for classifying mutation status, as defined by the specific test characteristics. The essential clinical question is to determine which testing method is best at classifying patients such that the maximum treatment effect is achieved both for patients whose tumours are classified as KRAS wild type, who receive cetuximab in addition to standard chemotherapy, and for those whose tumours are classified as KRAS mutant, who receive standard chemotherapy alone. To fully address this question, data of the type supplied by the COIN trial investigators (i.e. treatment effectiveness data for the addition of cetuximab to standard chemotherapy in both patients whose tumours are classified as KRAS wild type and those whose tumours are classified as KRAS mutant) would be required for each proposed KRAS mutation testing method. Ideally, data for all tests would be derived from the same study population, to allow meaningful comparison of the performance of tests for predicting treatment response without confounding by between-study variations in key participant characteristics. Following the recommendations made in NICE technology appraisal TA176,1 obtaining these data may be problematic as it could be argued that a trial in which patients are randomised to receive cetuximab in addition to standard chemotherapy or standard chemotherapy alone, regardless of tumour KRAS mutation status, would be unethical. Although the COIN trial54 was published after TA176,1 more recent UK trials such as New EPOC77 have tended to focus on determining the effectiveness of adding cetuximab to standard chemotherapy in patients with KRAS wild-type tumours. The recently complete, but as yet unpublished, New EPOC trial was a randomised open-label comparison between oxaliplatin/irinotecan plus fluorouracil plus cetuximab and oxaliplatin/irinotecan plus fluorouracil. The trial aimed to assess the effect on PFS of adding cetuximab to standard chemotherapy in patients with KRAS-wild type resectable CRC liver metastases who require chemotherapy. Trials of this type are not primarily concerned with the method used to establish mutation status. An alternative approach to this problem is provided by studies that report sufficient data to calculate the accuracy of different KRAS mutation tests for predicting response to treatment with cetuximab plus standard chemotherapy. These studies can potentially provide information on the extent to which KRAS mutation tests are able to predict resectability of liver metastases following treatment with cetuximab plus standard chemotherapy; outcome data (resection rates or ORRs) are reported for both patients with KRAS wild-type tumours and patients with KRAS mutant tumours. However, we were able to identify only one study of this type, the CELIM trial,52 which used an older version of the Therascreen RGQ PCR Kit. Neither the CELIM trial52 or the COIN trial54 were intended to assess KRAS mutation testing and neither reported comparative data for more than one KRAS mutation test; hence, any apparent differences in test performance observed between the two studies may have arisen as a result of differences in study populations. Of particular note is the way in which unresectable liver metastases were defined in the two studies: participants in the CELIM trial52 were described as having technically non-resectable or five or more liver metastases from CRC and it was therefore unclear whether some participants may have had potentially resectable metastases at baseline, whereas the COIN trial54 explicitly excluded patients receiving combination chemotherapy before resection of operable liver metastases. This difference may partially account for the marked difference in resection rates observable between the two studies for patients with KRAS wild-type tumours who were treated with cetuximab plus standard chemotherapy; (Commercial-in-confidence information has been removed),59 compared with a resection rate of 13/87 (15%) from the CELIM trial. 52 (Commercial-in-confidence information has been removed),59 whereas the COIN trial focused on ‘potentially curative liver resections’,54 and the standard chemotherapy regimens were different in the two trials.

Trials that compared the effectiveness of cetuximab plus standard chemotherapy with that of standard chemotherapy alone in patients with unresectable liver-limited metastases from CRC and whose tumours were KRAS wild type were also included in this review. These trials were included with the aim of providing some indication of how the favourable effects from the addition of cetuximab in these patients may vary according to how patients are selected for treatment (which KRAS mutation test is used). However, it should be noted that differences between these studies, other than the way in which KRAS wild-type mutation status is defined, particularly in relation to the baseline participant characteristics, are likely to contribute to any differences in treatment effects observed. In addition, these trials can provide no information about the relative effectiveness of cetuximab and standard chemotherapy compared with standard chemotherapy alone in patients whose tumours are classified as KRAS mutant.

The effectiveness data available to inform this assessment were very limited. In anticipation of this problem, our assessment included a survey of UK laboratories participating in the NEQAS scheme. This survey aimed to provide additional data on the technical performance of KRAS mutation tests, as seen in routine practice in the UK. We consider that data of this type are potentially more informative than data on the technical performance characteristics of tests obtained under research conditions, using non-clinical samples.

Cost-effectiveness

The review of economic analyses of different methods for KRAS mutation testing included only full economic analyses. Hence, economic studies that had information on test costs may be excluded from the review.

A de novo probabilistic model was developed to assess the cost-effectiveness of different methods for KRAS mutation testing to decide between standard chemotherapy and cetuximab in combination with standard chemotherapy in adults with mCRC in whom metastases are confined to the liver and are initially unresectable. To be consistent with related assessments/appraisals, it was first ensured that the model structure, model assumptions and input parameters in the de novo model were consistent with the manufacturer’s model used in NICE technology appraisal 176. 1,59,69 Model results were also consistent for patients with KRAS wild-type tumours in the sense that the use of cetuximab would still be considered cost-effective according to the de novo model.

In the assessment of the economic value of different tests, a link has to be established between test accuracy, clinical value (e.g. ORR, resection rate) and relative cost-effectiveness. Ideally, the performance of KRAS mutation tests would be assessed against an objective measure of the true presence/absence of a clinically relevant KRAS mutation (the ‘reference standard’), and the comparative effectiveness of treatment (chemotherapy plus cetuximab vs. chemotherapy alone) conditional on the true presence/absence of the KRAS mutation would be determined. However, different testing methods target different ranges of mutations and have different limits of detection (the lowest proportion of mutation detectable in tumour cells) and the optimal combination of mutation location and level for treatment selection remains unclear. For this reason, assessment of test performance based on comparison with a conventional ‘reference standard’ is currently not possible. An alternative way to determine the relative value of diagnostic methods for KRAS mutation testing is to use studies that report on the comparative treatment effect (or a substitute) in patients with both wild-type and mutant KRAS tumours. Thus, ORR or liver resection rate after treatment with cetuximab was assumed to correlate perfectly with the ‘true’ absence/presence of the KRAS mutation. The use of alternative outcome measures to determine test accuracy for the assessment of cost-effectiveness might impact the proportion of KRAS wild types to KRAS mutations and thus might substantially impact the assessment of cost-effectiveness (in either direction) as division of patients over the tumour mutation status categories is a major driver of cost-effectiveness. In the absence of an objective measure of the ‘true’ presence/absence of a clinically significant KRAS mutation, the current cost-effectiveness assessment is, at best, an approximation of the ‘true’ cost-effectiveness of test–treat combinations.

Evidence on test accuracy was available for only two tests (Therascreen KRAS RGQ PCR Kit and pyrosequencing); this was derived from ORR for the Therascreen KRAS RGQ PCR Kit and from resection rate for pyrosequencing. A major assumption underpinning our analyses was that the differences in liver resection rates observed in the two included studies from which these data were derived,52,54 and therefore also differences in the subsequent PFS and OS, can be attributed exclusively to the specific test used. In practice, this assumption would seem unlikely to hold true. These differences could also be caused by, for instance, differences in the characteristics of the respective study populations (i.e. with respect to the type of metastases) or differences in the standard chemotherapy regimen used. In addition, if the assumption of comparability of accuracy rates based on different measures (i.e. ORR and resection rate) holds true,39 this would reduce the likelihood that the main assumption holds.

Clinical effectiveness

As noted in Strengths and limitations, one important consideration when selecting a KRAS mutation testing method is the variation between tests in the limit of detection (i.e. the minimum percentage of mutation in tumour cells required to produce a positive result). A lower limit of detection can enhance the ability of laboratories to produce results from poor-quality samples. However, it should not be assumed that a lower limit of detection will necessarily result in a more clinically effective test, as it is possible that the addition of cetuximab to standard chemotherapy may still be effective in patients with KRAS mutant tumours in which mutations are present at a very low level (a low proportion of tumour cells harbouring a mutation). None of the studies that met the inclusion criteria for this review reported any data on variation in treatment effect according to the limit of detection used to define a KRAS mutant tumour.

A further area of uncertainty concerns the clinical value of detecting rarer KRAS mutations. The majority of the evidence on the effectiveness of first-line treatment with cetuximab plus standard chemotherapy in patients with liver-limited colorectal metastases whose tumours are KRAS wild type was derived from patients selected using tests that target mutations in codons 12 and 13; only the COIN trial54 used a test method that also targeted mutations in codon 61. Indeed, although no testing method was specified, the ASCO PCO published in 200915 recommended universal KRAS mutation testing in patients with mCRC in whom treatment with EGFR inhibitors is being considered. The recommendation also stated that testing should be carried out in an accredited laboratory and that patients whose tumours have KRAS mutations in codons 12 or 13 should not be treated with EGFR receptor inhibitors. The PCO also highlighted the uncertainty around the clinical relevance of detecting rare mutations in codons 61 and 146. 15 The COIN trial54 reported detection of the following mutations in codon 61, for all samples successfully analysed: Q61H (13/1059, 1.2%), Q61L (5/1289, 0.4%) and Q61R (6/1289, 0.5%); it was not clear whether any of these mutations were detected in patients with liver-limited metastases. The additional clinical value of using tests that target a wider range of mutations remains uncertain, as the low frequency of most KRAS mutations makes it very difficult to adequately assess treatment effects or resistance to EGFR inhibitors in patients with these mutations. A large multicentre observational study conducted in Italy by the KRAS aKtive network78 (a programme promoted by the Italian Association of Medical Oncology and the Italian Society of Surgical Pathology and Cytopathology to support the activity of oncologists and pathologists involved in the management of mCRC patients who require KRAS mutation testing) has collected data on a total of 7432 KRAS mutation analyses. The majority (77%) of testing was conducted using Sanger sequencing, and mutations other than those in codons 12 and 13 represented approximately 5% of the total detected. In addition to the issue of rare mutations, questions have been raised whether all codon 13 mutations predict lack of benefit from treatment with EGFR inhibitors; De Roock et al. 79 suggested that the KRAS G13D mutation may not predict lack of benefit. The COIN trial54 identified 110 participants with this mutation and reported no difference in outcome with the addition of cetuximab to standard chemotherapy; the HR for PFS was 1.11 (95% CI 0.76 to 1.63) in patients with the KRAS G13D mutation and 1.05 (95% CI 0.87 to 1.27) for all other mutations (data for the whole trial population, not the liver-limited metastases subgroup).

As discussed in Strengths and limitations, when assessing the performance of different KRAS mutation tests for the prediction of response to treatment, it is important to have information on the relative effectiveness of different treatment options in patients whose tumours are KRAS mutant as well as in those whose tumours are KRAS wild type. This is because, even when the benefits of adding cetuximab to standard chemotherapy in patients with KRAS wild-type tumours have been established, it is important to determine whether there are any negative effects associated with adding cetuximab to the treatment of patients with KRAS mutant tumours. If there are no negative effects associated with ‘overtreatment’ of patients with KRAS mutant tumours with cetuximab, then a conservative classification of patients with rare or low-level mutations as ‘wild type’ for treatment purposes may be considered clinically appropriate. Similarly, the ability of a test to detect rare mutations and/or a low limit of detection may be considered less important. None of the studies included in this assessment reported any difference in OS between patients with KRAS mutant tumours treated with cetuximab plus standard chemotherapy and those treated with standard chemotherapy alone. The CRYSTAL trial27 also reported no difference in ORRs or PFS, whereas the OPUS trial28 reported a lower ORR (OR 0.46, 95% CI 0.23 to 0.92) and shorter PFS (HR 1.72, 95% CI 1.10 to 2.68) for patients with KRAS mutant tumours who were treated with cetuximab plus standard chemotherapy than for those treated with standard chemotherapy alone. 27 These data were for all patients in the trials with KRAS mutant tumours; for both the CYSTAL and OPUS trials, data on treatment effectiveness in patients with KRAS mutant tumours were not available for the subgroup of patients with liver-limited metastases. Additional data supplied by the COIN trial investigators that were specific to patients with inoperable liver-limited metastases showed no significant difference in PFS or potentially curative resection rates between patients with KRAS mutant tumours who were treated with cetuximab plus standard chemotherapy and patients with KRAS mutant tumours who were treated with standard chemotherapy alone (D Fisher, personal communication).

The timing of KRAS mutation testing can vary, with some clinicians/hospitals undertaking routine testing of all CRC patients at diagnosis, potentially before the disease becomes metastatic, and others waiting until metastases have been detected. It should be noted that only one of the UK laboratories responding to our survey reported routine KRAS testing in all CRC patients. It could be argued that routine testing avoids potential delays in the start of treatment; however, clinical opinion suggested that any such delays would be unlikely to have measurable effects on clinical outcomes. Also, because cetuximab is added to standard chemotherapy in patients with KRAS wild-type tumours, standard chemotherapy can be commenced whilst awaiting the results of KRAS testing so that only the potential additional benefit of cetuximab is subject to delay. A related question is whether a stored biopsy sample from the primary tumour is adequate for KRAS mutation testing once metastases have been detected, or whether potential heterogeneity between tumour sites means that a sample from the metastatic site is preferable. Use of the primary tumour sample is likely to be considered preferable because all patients should have already undergone a biopsy at diagnosis for histological typing; thus, the risks and discomfort of a further invasive procedure (liver biopsy) could potentially be avoided. None of the studies included in this assessment considered the potential impact of sample site on the results of KRAS mutation testing. A systematic review80 identified by our searches, which did not meet the inclusion criteria for this assessment, assessed the concordance of KRAS mutations between primary CRC tissue and mCRC tissue. This review included 19 publications reporting data on a total of 986 paired samples from primary tumours and distant metastases (including, but not limited to, the liver) and reported a pooled concordancy rate of 94.1% (95% CI 88.3% to 95.0%). One of the primary studies included in this review specifically assessed KRAS mutation concordancy between primary colorectal tumours and liver metastases in 305 paired samples;32 KRAS mutation status was determined based on pyrosequencing of codons 12 and 13. This study reported a concordancy rate of 96.4% (95% CI 93.6% to 98.2%), with clinically relevant discordance in six participants (2.0% of the study population); five primary tumours had a KRAS mutation with a wild-type metastasis and one primary tumour was wild type with a KRAS mutation in the metastasis. Although outside the scope of this assessment, these studies could be interpreted as supporting the view that KRAS mutation testing using stored samples from the primary tumour is a valid approach and that testing using liver biopsy samples is unlikely to produce significant clinical benefit.

A variety of KRAS mutation testing methods is currently used by accredited NHS laboratories in England and Wales. None of the methods reported in our survey exactly matched the methods used in any of the studies identified in our systematic review. However, because the COIN trial54 reported > 99% concordance on KRAS genotyping between pyrosequencing and MALDI-TOF, it may be assumed that accuracy data from the COIN trial are also representative of the accuracy of pyrosequencing (used as a single test) for KRAS mutations in codons 12, 13 and 61, the method used by the majority of UK laboratories who responded to our survey. It should be noted that the performance of pyrosequencing methods may vary when different primers are used and that the potential clinical effects of using different KRAS mutation test methods to make decisions on first-line treatment in patients with unresectable liver-limited CRC metastases remain uncertain. The Therascreen KRAS RGQ PCR Kit is the only product currently approved by the FDA;16 however, the clinical study used to support its approval was not conducted in the population specified for this assessment and none of the respondents to our survey of UK laboratories reported using this product. The Therascreen KRAS RGQ PCR Kit, Therascreen KRAS Pyro Kit, cobas KRAS Mutation Test, KRAS LightMix Kit and KRAS StripAssay are all CE marked. No direct data, either from our systematic review or from our survey of UK laboratories, are currently available for the following KRAS mutation testing methods listed in the scope: next-generation sequencing of codons 12, 13 and 61; KRAS StripAssay; MALDI-TOF mass spectrometry of codons 12, 13 and 61 used alone; and HRM analysis of codons 12, 13 and 61 used alone. As was the case for pyrosequencing, concordance between the two KRAS mutation testing methods used in the COIN trial54 means that accuracy data derived from the trial may also be assumed to be representative of the performance of MALDI-TOF when used as a single test for the detection of KRAS mutations in codons 12, 13 and 61.

Cost-effectiveness

Major assumptions were made to be able to model the relative cost-effectiveness of different KRAS mutation tests. It was assumed that the differences in resection rates between the CELIM trial52 and the COIN trial54 and associated subsequent PFS and OS were exclusively attributable to the different mutation tests used (the Therascreen KRAS RGQ PCR Kit and pyrosequencing respectively) to distinguish between patients whose tumours are KRAS wild type and those whose tumours are KRAS mutant. As discussed previously, it is questionable whether this assumption would hold true. Furthermore, to calculate the proportions of patients with KRAS wild-type and KRAS mutant test results, patients with a KRAS wild-type test result were categorised as FP if no objective response was observed on cetuximab (for the Therascreen KRAS RGQ PCR Kit) or when no liver resection was performed (for pyrosequencing) or as TP if a objective response was observed or a resection was performed. Likewise, patients with a KRAS mutant test result were classified as TN when no objective response was observed on cetuximab (for the the Therascreen KRAS RGQ PCR Kit) or no resection was performed (for pyrosequencing) and as FN when an objective response was observed or a liver resection was performed. Ideally, the categorisation of TPs/FP and TNs/FNs should be based on an objective measure of the true presence/absence of a clinically relevant KRAS mutation. However, as previously described, the uncertainty around the exact definition of a clinically relevant mutation is such that, at present, there is no such thing as an objective measure or gold standard.

Moreover, as this model was partially based on the evidence and model structure used in the technology appraisal of cetuximab for the first-line treatment of mCRC,1,59,69 the assumptions underlying that appraisal also apply to this assessment. One example, which applies only to the linked evidence analysis, is the implicit assumption in the manufacturer’s model that, in the absence of a chemotherapy-only arm in the CELIM trial,52 resection rates from the GERCOR trial68 can be applied to patients with KRAS mutant and KRAS unknown status tumours treated with standard chemotherapy, whereas resection rates for patients with KRAS wild-type tumours treated with cetuximab were taken from the CELIM-trial. 52

Finally, it should be emphasised that the uncertainty resulting from the above-mentioned assumptions was not parameterised in the model and is therefore not reflected in the PSA or in the CEACs.

EMBASE (OvidSP), 2000 to Week 3 2013

Searched 22 January 2013

1. exp colon cancer/ or exp rectum cancer/ or colorectal tumor/ (169,199)

2. ((colorect$ or rectal$ or rectum$ or colon$ or sigma$ or sigmo$ or rectosigm$ or bowel$ or anal or anus) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot,hw. (245,923)

3. (m-CRC or CRC).ti,ab,ot. (14,043)

4. ((cecum or cecal or caecum or caecal or il?eoc?ecal or il?eoc?ecum) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (2124)

5. (large intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (1871)

6. (lower intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (26)

7. or/1-6 (249,697)

8. k ras oncogene/ (4953)

9. (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS).af. (17,025)

10. (Kirsten adj3 (murine or rat) adj3 sarcoma$).ti,ab,ot. (396)

11. (thera?screen$ or therascreen$).af. (67)

12. (Cobas adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (8)

13. (sanger sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (15)

14. (pyrosequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (25)

15. ((HRM or HRMA or dHPLC) adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (13)

16. (high resolution adj3 melt$ adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (8)

17. (SNapShot adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (5)

18. (Next generation sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

19. high resolution melting analysis/ (691)

20. 19 and (8 or 9 or 10) (62)

21. or/8-18,20 (17,279)

22. 7 and 21 (5716)

23. limit 22 to yr=“2000 -Current” (5036)

24. limit 23 to embase (4540)

MEDLINE (OvidSP), 2000 to January Week 2 2013

Searched 22 January 2013

1. exp Colorectal Neoplasms/ (134,723)

2. ((colorect$ or rectal$ or rectum$ or colon$ or sigma$ or sigmo$ or rectosigm$ or bowel$ or anal or anus) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot,hw. (165,769)

3. (m-CRC or CRC).ti,ab,ot. (8215)

4. ((cecum or cecal or caecum or caecal or il?eoc?ecal or il?eoc?ecum) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (1570)

5. (large intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (1541)

6. (lower intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (23)

7. or/1-6 (170,682)

8. Genes, ras/ (11,077)

9. (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS).af. (9538)

10. (Kirsten adj3 (murine or rat) adj3 sarcoma$).ti,ab,ot. (346)

11. (thera?screen$ or therascreen$).af. (16)

12. (Cobas adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (2)

13. (sanger sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (4)

14. (pyrosequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (12)

15. ((HRM or HRMA) adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (5)

16. (high resolution adj3 melt$ adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (4)

17. (SNapShot adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (3)

18. (Next generation sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

19. or/8-18 (16,696)

20. 7 and 19 (3083)

21. limit 20 to yr=“2000 -Current” (2293)

22. remove duplicates from 21 (2278)

MEDLINE In-Process & Other Non-Indexed Citations (OvidSP) and Daily Update (OvidSP), up to 21 January 2013

Searched 22 January 2013

1. exp Colorectal Neoplasms/ (194)

2. ((colorect$ or rectal$ or rectum$ or colon$ or sigma$ or sigmo$ or rectosigm$ or bowel$ or anal or anus) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot,hw. (7747)

3. (m-CRC or CRC).ti,ab,ot. (1006)

4. ((cecum or cecal or caecum or caecal or il?eoc?ecal or il?eoc?ecum) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (108)

5. (large intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (28)

6. (lower intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (0)

7. or/1-6 (7930)

8. Genes, ras/ (8)

9. (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS).af. (659)

10. (Kirsten adj3 (murine or rat) adj3 sarcoma$).ti,ab,ot. (17)

11. (thera?screen$ or therascreen$).af. (5)

12. (Cobas adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

13. (sanger sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (2)

14. (pyrosequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

15. ((HRM or HRMA) adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

16. (high resolution adj3 melt$ adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

17. (SNapShot adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

18. (Next generation sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

19. or/8-18 (667)

20. 7 and 19 (269)

Cochrane Database of Systematic Reviews (CDSR) (Wiley), 2000 to Issue 12, 2012; Cochrane Central Register of Controlled Trials (CENTRAL) (Wiley), 2000 to Issue 12, 2012; Database of Abstracts of Reviews of Effects (DARE) (Wiley), 2000 to Issue 4, 2012; and Health Technology Assessment (HTA) database (Wiley), 2000 to Issue 4, 2012

Searched 22 January 2013

#1 MeSH descriptor: [Colorectal Neoplasms] explode all trees (4380)

#2 ((colorect* or rectal* or rectum* or colon* or sigma* or sigmo* or rectosigm* or bowel* or anal or anus) near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (7773)

#3 (m-CRC or CRC) (715)

#4 ((cecum or cecal or caecum or caecal or ileocecal or ileocaecal or ileocaecum or ileocecum) near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (24)

#5 (large intestin* near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor*or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (86)

#6 (lower intestin* near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (114)

#7 #1 or #2 or #3 or #4 or #5 or #6 (8053)

#8 MeSH descriptor: [Genes, ras] this term only (46)

#9 (k ras or kras or K-ras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS) (386)

#10 (Kirsten near/3 (murine or rat) near/3 sarcoma*) (7)

#11 (thera screen* or thera-screen* or therascreen*) (13)

#12 (Cobas) (115)

#13 (sanger sequencing) (7)

#14 (pyrosequencing) (18)

#15 (HRM or HRMA) (11)

#16 (high resolution near/3 melt*) (1)

#17 (SNapShot) (50)

#18 (“Next generation sequencing”) (2)

#19 #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 (605)

#20 #7 and #19 from 2000 (98)

The CDSR search retrieved nine references, the CENTRAL search retrieved 65 references, the DARE search retrieved 11 references and the HTA database search retrieved nine references.

National Institute for Health Research (NIHR) Health Technology Assessment (HTA) programme (Internet), up to 25 January 2013

www.hta.ac.uk/

Searched 25 January 2013

Browsed by relevant terms; retrieved two references.

Science Citation Index (SCI-EXPANDED) (Web of Knowledge) and Conference Proceedings Citation Index (CPCI-S) (Web of Knowledge), 2000 to 22 January 2013

Searched 23 January 2013

Databases=SCI-EXPANDED, CPCI-S Timespan=2000-01-01 - 2013-01-23

#18 3597 #17 AND #6

#17 10,477 #16 OR #15 OR #14 OR #13 OR #12 OR #11 OR #10 OR #9 OR #8 OR #7

#16 26 TS=((Next SAME generation SAME sequencing) SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#15 23 TS=(SNapShot SAME (k ras or kras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#14 75 TS=((high SAME resolution SAME melt*) SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#13 59 TS=((HRM or HRMA or dHPLC) SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#12 94 TS=(pyrosequencing SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#11 49 TS=((sanger SAME sequencing) SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#10 2 TS=(Cobas SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#9 17 TS=(thera$screen* or therascreen*)

#8 96 TS=(Kirsten NEAR/3 (murine or rat) NEAR/3 sarcoma*)

#7 10,467 TS=(k ras or kras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)

#6 134,422 #1 or #2 or #3 or #4 or #5

#5 1328 TS=(lower SAME intestin* NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#4 1113 TS=(large SAME intestin* NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#3 484 TS=((cecum or cecal or caecum or caecal or il$eoc$ecal or il$eoc$ecum) NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#2 9622 TS=(m-CRC or CRC)

#1 130,942 TS=((colorect* or rectal* or rectum* or colon* or sigma* or sigmo* or rectosigm* or bowel* or anal or anus) NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

BIOSIS Previews (Web of Knowledge), 2000 to 22 January 2013

Searched 23 January 2013

Databases=BIOSIS Previews Timespan=2000-2013

#18 2641 #17 AND #6

#17 8621 #16 OR #15 OR #14 OR #13 OR #12 OR #11 OR #10 OR #9 OR #8 OR #7

#16 39 TS=((Next SAME generation SAME sequencing) SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#15 36 TS=(SNapShot SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#14 73 TS=((high SAME resolution SAME melt*) SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#13 55 TS=((HRM or HRMA or dHPLC) SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#12 133 TS=(pyrosequencing SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#11 95 TS=((sanger SAME sequencing) SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#10 4 TS=(Cobas SAME (k-ras or k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#9 14 TS=(thera$screen* or therascreen*)

#8 153 TS=(Kirsten NEAR/3 (murine or rat) NEAR/3 sarcoma*)

#7 8611 TS=(k-ras or k ras or kras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)

#6 97,980 #1 or #2 or #3 or #4 or #5

#5 1020 TS=(lower SAME intestin* NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#4 805 TS=(large SAME intestin* NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#3 424 TS=((cecum or cecal or caecum or caecal or il$eoc$ecal or il$eoc$ecum) NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#2 7235 TS=(m-CRC or CRC)

#1 95,876 TS=((colorect* or rectal* or rectum* or colon* or sigma* or sigmo* or rectosigm* or bowel* or anal or anus) NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

Latin American and Caribbean Health Sciences (LILACS) (Internet), up to 24 January 2013

Searched 23 January 2013

International Prospective Register of Systematic Reviews (PROSPERO) (Internet), up to 25 January 2013

www.crd.york.ac.uk/prospero/

Searched 25 January 2013

Searched for terms in ‘All Fields’

Clinicaltrials.gov (Internet), 2000 to 23 January 2013

http://clinicaltrials.gov/ct2/search/advanced

Searched 23 January 2013

Advanced search option – search terms box

Limited to results received from 1 January 2000 to 23 January 2013

World Health Organization (WHO) International Clinical Trials Registry Platform (ICTRP) (Internet), 2000 to 25 January 2013

www.who.int/ictrp/en/

Searched 25 January 2013

Advanced search option

Current Controlled Trials (metaRegister of Controlled Trials) (Internet), up to 29 January 2013

www.controlled-trials.com/

Searched 29 January 2013

Conference searches

EMBASE (OvidSP), 2000 to Week 4 2013

Searched 29 January 2013

1. exp colon cancer/ or exp rectum cancer/ or colorectal tumor/ (169,460)

2. ((colorect$ or rectal$ or rectum$ or colon$ or sigma$ or sigmo$ or rectosigm$ or bowel$ or anal or anus) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot,hw. (246,253)

3. (m-CRC or CRC).ti,ab,ot. (14,068)

4. ((cecum or cecal or caecum or caecal or il?eoc?ecal or il?eoc?ecum) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (2125)

5. (large intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (1871)

6. (lower intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (26)

7. or/1-6 (250,031)

8. k ras oncogene/ (4967)

9. (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS).af. (17,128)

10. (Kirsten adj3 (murine or rat) adj3 sarcoma$).ti,ab,ot. (399)

11. (thera?screen$ or therascreen$).af. (67)

12. (Cobas adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (8)

13. (sanger sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (15)

14. (pyrosequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (25)

15. ((HRM or HRMA or dHPLC) adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (13)

16. (high resolution adj3 melt$ adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (8)

17. (SNapShot adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (5)

18. (Next generation sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

19. high resolution melting analysis/ (701)

20. 19 and (8 or 9 or 10) (62)

21. or/8-18,20 (17,382)

22. 7 and 21 (5731)

23. health-economics/ (32,282)

24. exp economic-evaluation/ (194,421)

25. exp health-care-cost/ (186,276)

26. exp pharmacoeconomics/ (160,882)

27. or/23-26 (445,930)

28. (econom$ or cost or costs or costly or costing or price or prices or pricing or pharmacoeconomic$).ti,ab. (542,934)

29. (expenditure$ not energy).ti,ab. (21,678)

30. (value adj2 money).ti,ab. (1191)

31. budget$.ti,ab. (22,178)

32. or/28-31 (565,244)

33. 27 or 32 (823,889)

34. letter.pt. (811,274)

35. editorial.pt. (424,059)

36. note.pt. (543,769)

37. or/34-36 (1,779,102)

38. 33 not 37 (742,302)

39. (metabolic adj cost).ti,ab. (800)

40. ((energy or oxygen) adj cost).ti,ab. (3005)

41. ((energy or oxygen) adj expenditure).ti,ab. (18,652)

42. or/39-41 (21,682)

43. 38 not 42 (737,540)

44. 22 and 43 (310)

45. limit 44 to yr=“2000 -Current” (303)

46. limit 45 to embase (285)

Costs filter: Centre for Reviews and Dissemination. NHS EED Economics Filter: EMBASE (Ovid) Weekly Search. York: CRD; 2010. URL: www.york.ac.uk/inst/crd/intertasc/nhs_eed_strategies.html (accessed 17 March 2011).

MEDLINE (OvidSP), 2000 to January Week 3 2013

Searched 29 January 2013

1. exp Colorectal Neoplasms/ (134,899)

2. ((colorect$ or rectal$ or rectum$ or colon$ or sigma$ or sigmo$ or rectosigm$ or bowel$ or anal or anus) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot,hw. (165,996)

3. (m-CRC or CRC).ti,ab,ot. (8243)

4. ((cecum or cecal or caecum or caecal or il?eoc?ecal or il?eoc?ecum) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (1571)

5. (large intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (1542)

6. (lower intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (23)

7. or/1-6 (170,912)

8. Genes, ras/ (11,084)

9. (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS).af. (9558)

10. (Kirsten adj3 (murine or rat) adj3 sarcoma$).ti,ab,ot. (346)

11. (thera?screen$ or therascreen$).af. (16)

12. (Cobas adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (2)

13. (sanger sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (5)

14. (pyrosequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (12)

15. ((HRM or HRMA) adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (5)

16. (high resolution adj3 melt$ adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (4)

17. (SNapShot adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (3)

18. (Next generation sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

19. or/8-18 (16,720)

20. 7 and 19 (3092)

21. economics/ (26,342)

22. exp “costs and cost analysis”/ (168,037)

23. economics, dental/ (1847)

24. exp “economics, hospital”/ (18,317)

25. economics, medical/ (8474)

26. economics, nursing/ (3868)

27. economics, pharmaceutical/ (2383)

28. (economic$ or cost or costs or costly or costing or price or prices or pricing or pharmacoeconomic$).ti,ab. (375,308)

29. (expenditure$ not energy).ti,ab. (15,563)

30. (value adj1 money).ti,ab. (18)

31. budget$.ti,ab. (15,762)

32. or/21-31 (493,254)

33. ((energy or oxygen) adj cost).ti,ab. (2454)

34. (metabolic adj cost).ti,ab. (667)

35. ((energy or oxygen) adj expenditure).ti,ab. (14,408)

36. or/33-35 (16,880)

37. 32 not 36 (489,444)

38. letter.pt. (758,034)

39. editorial.pt. (307,072)

40. historical article.pt. (288,506)

41. or/38-40 (1,339,895)

42. 37 not 41 (463,260)

43. 20 and 42 (74)

44. limit 43 to yr=“2000 -Current” (69)

Costs filter: Centre for Reviews and Dissemination. NHS EED Economics Filter: MEDLINE (Ovid) Monthly Search. York: CRD; 2010. URL: www.york.ac.uk/inst/crd/intertasc/nhs_eed_strategies.html (accessed 28 September 2010).

MEDLINE In-Process & Other Non-Indexed Citations (OvidSP) and MEDLINE Daily Update (OvidSP), up to 28 January 2013

Searched 29 January 2013

1. exp Colorectal Neoplasms/ (132)

2. ((colorect$ or rectal$ or rectum$ or colon$ or sigma$ or sigmo$ or rectosigm$ or bowel$ or anal or anus) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot,hw. (7699)

3. (m-CRC or CRC).ti,ab,ot. (1022)

4. ((cecum or cecal or caecum or caecal or il?eoc?ecal or il?eoc?ecum) adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (111)

5. (large intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (28)

6. (lower intestin$ adj3 (cancer$ or neoplas$ or oncolog$ or malignan$ or tumo?r$ or carcinoma$ or adenocarcinoma$ or metasta$ or meta-sta$ or sarcoma$ or adenom$ or lesion$)).ti,ab,ot. (0)

7. or/1-6 (7889)

8. Genes, ras/ (6)

9. (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS).af. (666)

10. (Kirsten adj3 (murine or rat) adj3 sarcoma$).ti,ab,ot. (17)

11. (thera?screen$ or therascreen$).af. (5)

12. (Cobas adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

13. (sanger sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

14. (pyrosequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

15. ((HRM or HRMA) adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

16. (high resolution adj3 melt$ adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

17. (SNapShot adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (1)

18. (Next generation sequencing adj3 (k ras or kras or k-ras or V-Ki-ras$ or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)).af. (0)

19. or/8-18 (671)

20. 7 and 19 (270)

21. economics/ (1)

22. exp “costs and cost analysis”/ (143)

23. economics, dental/ (0)

24. exp “economics, hospital”/ (8)

25. economics, medical/ (0)

26. economics, nursing/ (0)

27. economics, pharmaceutical/ (1)

28. (economic$ or cost or costs or costly or costing or price or prices or pricing or pharmacoeconomic$).ti,ab. (33,295)

29. (expenditure$ not energy).ti,ab. (992)

30. (value adj1 money).ti,ab. (3)

31. budget$.ti,ab. (1659)

32. or/21-31 (35,017)

33. ((energy or oxygen) adj cost).ti,ab. (186)

34. (metabolic adj cost).ti,ab. (46)

35. ((energy or oxygen) adj expenditure).ti,ab. (681)

36. or/33-35 (890)

37. 32 not 36 (34,752)

38. letter.pt. (19,083)

39. editorial.pt. (12,353)

40. historical article.pt. (144)

41. or/38-40 (31,562)

42. 37 not 41 (34,345)

43. 20 and 42 (17)

Costs filter:

Costs filter: Centre for Reviews and Dissemination. NHS EED Economics Filter: MEDLINE (Ovid) Monthly Search. York: CRD; 2010. URL: www.york.ac.uk/inst/crd/intertasc/nhs_eed_strategies.html (accessed 28 September 2010).

NHS Economic Evaluation Database (NHS EED) (Wiley), 2000 to Issue 4, 2012

Searched 22 January 2013

#1 MeSH descriptor: [Colorectal Neoplasms] explode all trees (4380)

#2 ((colorect* or rectal* or rectum* or colon* or sigma* or sigmo* or rectosigm* or bowel* or anal or anus) near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (7773)

#3 (m-CRC or CRC) (715)

#4 ((cecum or cecal or caecum or caecal or ileocecal or ileocaecal or ileocaecum or ileocecum) near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (24)

#5 (large intestin* near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor*or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (86)

#6 (lower intestin* near/3 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (114)

#7 #1 or #2 or #3 or #4 or #5 or #6 (8053)

#8 MeSH descriptor: [Genes, ras] this term only (46)

#9 (k ras or kras or K-ras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS) (386)

#10 (Kirsten near/3 (murine or rat) near/3 sarcoma*) (7)

#11 (thera screen* or thera-screen* or therascreen*) (13)

#12 (Cobas) (115)

#13 (sanger sequencing) (7)

#14 (pyrosequencing) (18)

#15 (HRM or HRMA) (11)

#16 (high resolution near/3 melt*) (1)

#17 (SNapShot) (50)

#18 (“Next generation sequencing”) (2)

#19 #8 or #9 or #10 or #11 or #12 or #13 or #14 or #15 or #16 or #17 or #18 (605)

#20 #7 and #19 from 2000 (98)

The NHS EED search retrieved three references.

Science Citation Index (SCI-EXPANDED) (Web of Knowledge), 2000 to 25 January 2013

Searched 30 January 2013

Databases=SCI-EXPANDED Timespan=2000-01-01 - 2013-01-30.

#27 117 #6 and #17 and #26

#26 532,023 #21 not #25

#25 33,411 #22 or #23 or #24

#24 15,970 TS=((energy or oxygen) NEAR expenditure)

#23 2095 TS=(metabolic NEAR cost)

#22 17,130 TS=((energy or oxygen) NEAR cost)

#21 551,568 #18 or #19 or #20

#20 909 TS=(value NEAR money)

#19 10,944 TS=(expenditure* not energy)

#18 546,907 TS=(economic* or cost or costs or costly or costing or price or prices or pricing or pharmacoeconomic* or budget*)

#17 10,434 #16 OR #15 OR #14 OR #13 OR #12 OR #11 OR #10 OR #9 OR #8 OR #7

#16 27 TS=((Next SAME generation SAME sequencing) SAME (k ras or kras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or k-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#15 23 TS=(SNapShot SAME (k ras or kras or k-rasor V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#14 76 TS=((high SAME resolution SAME melt*) SAME (k ras or kras or k-ras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#13 59 TS=((HRM or HRMA or dHPLC) SAME (k ras or kras or k-ras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#12 95 TS=(pyrosequencing SAME (k ras or kras or k-ras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#11 50 TS=((sanger SAME sequencing) SAME (k ras or kras or k-ras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#10 3 TS=(Cobas SAME (k ras or kras or k-ras or V-Ki-ras* or V-K-ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS))

#9 17 TS=(thera$screen* or therascreen*)

#8 97 TS=(Kirsten NEAR/3 (murine or rat) NEAR/3 sarcoma*)

#7 10,424 TS=(k ras or kras or K-ras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS)

#6 132,645 #1 or #2 or #3 or #4 or #5

#5 1321 TS=(lower SAME intestin* NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#4 1097 TS=(large SAME intestin* NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#3 481 TS=((cecum or cecal or caecum or caecal or il$eoc$ecal or il$eoc$ecum) NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo$r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

#2 8999 TS=(m-CRC or CRC)

#1 129,783 TS=((colorect* or rectal* or rectum* or colon* or sigma* or sigmo* or rectosigm* or bowel* or anal or anus) NEAR/3 (cancer* or neoplas* or oncolog* or malignan* or tumo?r* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*))

EconLit (EBSCO), 2000 to 30 January 2013

Searched 30 January 2013

Search modes – Boolean/phrase

1. S6 s1 or s2 or s3 or s4 Limiters - Published Date from: 20000101- (104)

2. S5 s1 or s2 or s3 or s4 (135)

3. S4 TX ((colorect* or rectal* or rectum* or colon* or sigma* or sigmo* or rectosigm* or bowel* or anal or anus) N4 (cancer* or neoplas* or oncolog* or malignan* or tumour* or tumor* or carcinoma* or adenocarcinoma* or metasta* or meta-sta* or sarcoma* or adenom* or lesion*)) (90)

4. S3 TX(thera screen* or thera-screen* or therascreen*) (0)

5. S2 TX(Kirsten murine sarcoma* or Kirsten rat sarcoma*) (0)

6. S1 TX(k ras or kras or k-ras or V-Ki-ras* or V-K-ras or V-Ki-ras or v ki ras or c-ki-ras or c-k-ras or ki-ras or ki ras or Kras1 or Kras2 or KRAS1P or RASK or RASK1 or RASK2 or Kirsten RAS) (45)

Health Economic Evaluations Database (HEED) (Internet), up to 30 January 2013

http://onlinelibrary.wiley.com/book/10.1002/9780470510933

Searched 30 January 2013

Compound search (all data), unable to limit by date

The HEED search retrieved 13 records.

Folprecht et al.52 (CELIM)

Maughan et al.54 (COIN)

Bokemeyer et al.28^,53^,56 (OPUS)

Van Cutsem et al.27^,53 (CRYSTAL)

Maughan et al.54 (COIN)

Xu et al.55