Molecular testing strategies for Lynch syndrome in people with colorectal cancer

The committee reviewed the available evidence on the clinical effectiveness of using immunohistochemistry (IHC) testing of tumour tissue for mismatch repair (MMR) proteins and microsatellite instability (MSI) to identify tumours with deficient DNA mismatch repair in people with colorectal cancer to assess their risk of having Lynch syndrome.

The committee discussed the generalisability of data from the studies, which were identified in the clinical review, to the decision problem. It noted that estimates for sensitivity values were taken from all the studies in the review, which included colorectal cancer patients who were identified as being at high risk of Lynch syndrome and age-limited patient populations that would be expected to have a higher prevalence of Lynch syndrome. The committee heard from the external assessment group (EAG) that the incidence of MSI in sporadic colorectal cancer increases with age and that this may alter test accuracy values in different age groups. The committee concluded that although there were differences in the trial populations in identified studies and the population of people with colorectal cancer in the UK, the effect of this on test accuracy was likely to be minimal.

The committee considered the evidence available on the diagnostic accuracy of MSI and IHC testing for MMR proteins. It noted that no identified studies directly compared MSI and IHC testing. It heard from the clinical experts that, in their experience, these tests are comparable in diagnostic accuracy. The committee noted that the tests appeared to be accurate enough for detecting MSI or abnormal expression of MMR proteins, but noted that these findings alone are not enough to diagnose Lynch syndrome without second-line tumour-based testing and subsequent genetic testing. Further, it heard that external quality assurance programmes are used to ensure the accuracy and consistency of testing between laboratories. The committee also heard that both tests are used in current practice in the NHS, and that the choice of test used is often determined by locally available services and expertise. The committee concluded that these tests are broadly comparable in accuracy.

The committee discussed the issue of unclassified variants, that is, when genetic testing identifies variations in the sequence of MMR genes that are of unknown clinical significance. This can affect whether results from the reference standard test are classified as positive or negative. The committee noted that relatively few studies identified in the clinical review had enough data to allow alternative analyses when unclassified variants were considered as positive reference standard results for Lynch syndrome. The clinical experts commented that in practice, unclassified variants are investigated further by asking for additional clinical information and testing before a diagnosis is given. The committee also heard that there are ongoing efforts to classify sequence variants in MMR genes and that the number of unclassified variants is therefore decreasing. The committee concluded that unclassified variants are unlikely to have a large effect on diagnosing Lynch syndrome in clinical practice.

The committee considered the cost effectiveness of the different testing strategies to identify Lynch syndrome in people with colorectal cancer. It noted that 10 strategies had been modelled, each using different combinations of tumour-based tests and genetic testing (see table 3).

The committee discussed the assumptions about the effectiveness of aspirin as a risk-reducing strategy for people with Lynch syndrome that were made in the economic model. It heard from the EAG that in the model, the effect of aspirin in reducing the risk of colorectal and endometrial cancer was assumed to occur instantaneously and last for 10 years, after which time the effect was assumed to stop instantaneously. However, the committee heard from the clinical experts that the Colorectal adenoma/carcinoma prevention programme 2 (CaPP2) trial of aspirin prophylaxis in Lynch syndrome reported that there is a lag time in the protective effect after starting therapy and that its effects can continue after people stop taking aspirin. The committee noted that the scenario analysis without the costs and effects of aspirin prophylaxis showed no substantial effect on overall results. The committee concluded that the effect of the assumptions about aspirin prophylaxis was likely to be small.

The committee discussed the effect estimates of colonoscopic surveillance used in the model. It noted that a study used to estimate the effectiveness of colonoscopic surveillance in people with Lynch syndrome in the model was about 15 years old and questioned whether this represents current practice in the NHS. The committee heard from the clinical experts that recent technological developments in this area and the introduction of standards by the Joint Advisory Group on Gastrointestinal Endoscopy have improved the effectiveness of colonoscopic surveillance. Data from cancer screening programmes have also shown that colonoscopic surveillance can lead to the detection of colorectal cancer at an earlier stage, which could improve patient outcomes. The committee also noted that the effectiveness of colonoscopic surveillance is likely to be influenced not only by the effectiveness of the test, but also by patient uptake and were reassured by the clinical experts that uptake of surveillance was high among people with Lynch syndrome. The committee concluded that colonoscopic surveillance is likely to reduce the risk of cancer developing in people with Lynch syndrome, and that consequently the effect estimate used in the base-case analysis was appropriate.

The committee considered the results of the base-case analysis, which suggested that strategies that began with IHC for MMR proteins were more cost effective than those that began with MSI testing, and that overall, strategy 5 appeared to be the most cost effective. The committee discussed the extent to which the results of the model were driven by the sensitivity and specificity parameter values used in the model. It noted that in the base-case analysis, MSI testing was assumed to be both less sensitive and less specific than IHC testing. The committee considered that, given the perceived equivalence of MSI and IHC testing and the absence of direct comparative data, there was not enough evidence to conclude that testing strategies that begin with MSI testing are not cost effective compared with IHC testing for MMR proteins.

The committee discussed the scenario analysis in which MSI‑Low (MSI‑L) was classified as a positive result for Lynch syndrome and noted that only MSH‑High (MSI‑H) was considered a positive result in the base case. It heard from the clinical experts that in current practice, both MSI‑L and MSI‑H results are generally considered indicative of Lynch syndrome. The results of the scenario analysis suggested that including MSI‑L as a positive result did not affect the cost effectiveness of the MSI‑based testing strategies. The committee concluded that both MSI‑L and MSI‑H should be considered as positive results.

The committee discussed the role of BRAF V600E and MLH1 promoter hypermethylation testing in the modelled strategies. It heard from the clinical experts that some tumours that test positive for MSI, or that have abnormal MLH1 protein expression, are sporadic colorectal cancers. Further, it heard that BRAF V600E and MLH1 promoter hypermethylation testing, particularly in combination, can be used to identify sporadic colorectal cancers and so reduce the number of people who are referred for genetic testing for Lynch syndrome. In addition, the clinical experts advised that testing strategies that aim to decrease the number of false-positive diagnoses for Lynch syndrome reduce the number of people having unnecessary colonoscopic surveillance. The committee concluded that strategies 5 and 9, which include tests to identify sporadic colorectal cancers, after first having MSI or IHC testing, are likely to be the most cost-effective options.

The committee discussed the cost effectiveness of the testing strategies in different age groups. It noted that the age-restricted subgroup analysis had little effect on the overall conclusions, but that referral straight to genetic testing was unlikely to be cost effective in older age groups. The committee heard from the clinical experts that although the prevalence of Lynch syndrome is much higher in younger people with colorectal cancer, it can still cause colorectal cancer in older people. It also heard that despite the lower prevalence of Lynch syndrome in older people, the greater number of colorectal cancer diagnoses in these age groups could mean that the absolute number of people who could benefit from a Lynch syndrome diagnosis may be similar to that in younger age groups. Therefore, the committee considered that there is no clinical reason to treat age groups differently. The committee concluded that all people, regardless of their age, with colorectal cancer should have tumour-based testing to assess the risk of Lynch syndrome.

The committee considered the joint effect of parameter uncertainty used in the model, and noted that this had not been explored in a probabilistic sensitivity analysis. It heard from the EAG that the univariate deterministic sensitivity analyses did not result in large changes to the incremental cost-effective ratios (ICERs) or the net health benefit values, and that it was unlikely that negative net health benefit values would be seen in a probabilistic sensitivity analysis. The committee considered that parameter uncertainty had been explored sufficiently, and that further analyses were unlikely to substantially change the overall results of the economic modelling. Therefore, the committee concluded that testing all people with colorectal cancer using strategies 5 (IHC plus BRAF V600E and MLH1 promoter methylation) and 9 (MSI plus BRAF V600E and MLH1 promoter methylation) would be a cost-effective use of NHS resource.

The committee discussed the timing of testing for Lynch syndrome in people who have been diagnosed with colorectal cancer. It heard from the clinical experts that a person's MMR tumour or gene status may be used to determine treatment options for colorectal cancer, for example, to direct surgical decisions or chemotherapy, although the clinical utility of using tumour testing to guide the selection of chemotherapy is not fully understood at present. However, it noted that it is very unlikely that definitive genetic testing will be completed before treatment for colorectal cancer begins. The committee therefore concluded that testing for Lynch syndrome should be started as soon as colorectal cancer is diagnosed, but should not delay the start of treatment.

The committee discussed which tissue samples should be used for testing. It heard from the clinical experts that there is good correlation between results for tissue from biopsies and tissue from resections. The committee concluded that clinical judgement should be used to determine the tumour material to be tested, and that tissue from a biopsy, resected colorectal tumour or polyp can be used. The committee also noted that people with Lynch syndrome may develop more than 1 colorectal cancer at the same time. It heard from the clinical experts that some of these cancers may differ in DNA mismatch repair functionality because people with Lynch syndrome can get sporadic colorectal cancers. The committee concluded that testing for Lynch syndrome should be considered for each individual cancer.

The committee noted that Lynch syndrome is not the only inherited condition that increases the risk of colorectal cancer. It heard from the clinical experts that other inherited causes of colorectal cancer include familial adenomatous polyposis. The clinical experts also emphasised that it is important that these additional inherited conditions are considered if someone is found not to have Lynch syndrome but the clinician suspects that the person's family history suggests that a genetic cause is likely. The committee concluded that clinical judgement should be used to determine whether a referral to clinical genetics is appropriate when Lynch syndrome has been ruled out by tumour-based testing, but other genetic causes are suspected.

The committee considered ongoing developments in genetic testing technologies. It noted that in the future, broad-range genetic sequencing or specific cancer panels using next-generation sequencing technology may be considered for diagnosing Lynch and other inherited colorectal cancer syndromes. The committee concluded that these advances may identify alternative and more rapid methods for diagnosing Lynch syndrome.

The committee discussed the value of developing research recommendations for tumour testing for Lynch syndrome. It considered that further research was unlikely to change its recommendations on molecular testing strategies for Lynch syndrome in people diagnosed with colorectal cancer.

The committee heard that good communication between colorectal cancer multidisciplinary teams and genetics or pathology laboratories is important for implementing tumour-based testing for Lynch syndrome to ensure that testing and reporting of results is coordinated. The committee noted that similar systems are embedded in breast cancer care pathways, in which reflex testing for human epidermal growth factor receptor 2 (HER2) and BRCA are done as part of the first assessment. The committee therefore wished to encourage centres adopting Lynch syndrome testing strategies to audit and publish their clinical and diagnostic outcomes to ensure that assessment of Lynch syndrome is timely and appropriate.

The committee heard from the clinical experts that centres already offering tumour-based testing for Lynch syndrome often carry out both MSI and IHC testing on samples. The committee encouraged these centres to publish their previously generated comparative results.