Article Text

Download PDFPDF

Screening for Lynch syndrome and referral to clinical genetics by selective mismatch repair protein immunohistochemistry testing: an audit and cost analysis
  1. Richard Colling1,
  2. David N Church2,
  3. Juliet Carmichael2,
  4. Lucinda Murphy3,
  5. James East4,
  6. Peter Risby5,
  7. Rachel Kerr2,
  8. Runjan Chetty6,
  9. Lai Mun Wang3
  1. 1Nuffield Division of Clinical Laboratory Sciences, University of Oxford, Oxford, UK
  2. 2Oxford Cancer Centre, University of Oxford, Oxford, UK
  3. 3Department of Cellular Pathology, Oxford University Hospitals NHS Trust, Oxford, UK
  4. 4Department of Gastroenterology, Oxford University Hospitals NHS Trust, Oxford, UK
  5. 5Department of Genetics, Oxford University Hospitals NHS Trust, Oxford, UK
  6. 6Department of Pathology, University Health Network, Toronto, Ontario, Canada
  1. Correspondence to Dr Richard Colling, Nuffield Division of Clinical Laboratory Sciences, University of Oxford, Cellular Pathology, Level 1, John Radcliffe Hospital, Headington, Oxford OX3 9DU, UK; rtcolling{at}gmail.com

Abstract

Lynch syndrome (LS) accounts for around 3% of colorectal cancers (CRCs) and is caused by germline mutations in mismatch repair (MMR) genes. Recently, screening strategies to identify patients with LS have become popular. We audited CRCs screened with MMR immunohistochemistry (IHC) in 2013. 209 tumours had MMR IHC performed at a cost of £12 540. 47/209 (21%) cases showed IHC loss of expression in at least one MMR protein. 28/44 cases with loss of MLH1 had additional BRAF V600E testing, at a cost of £5040. MMR IHC reduced the number of potential clinical genetics referrals from 209 to 47. BRAF mutation testing, performed in a subset of cases with MLH1 loss, further reduced this to 21. At a cost of £1340 per referral, this model of LS screening for clinical genetics referral had significant potential savings (£234 340) and can be easily implemented in parallel with MMR IHC done for prognostication in CRCs.

  • COLORECTAL CANCER
  • AUDIT
  • QUALITY ASSURANCE

Statistics from Altmetric.com

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Introduction

Colorectal cancers (CRC) account for 12.5% of new (non-skin) malignancies diagnosed each year in the UK and almost 10% of cancer-related deaths.1 Most CRCs are sporadic but 25% are associated with hereditary factors. Lynch syndrome (LS) is the most common genetically defined inherited cause of CRC and represents 3% of all newly diagnosed tumours.2 The lifetime risk of CRC in individuals with LS is between 30% and 70%, in contrast to 5.5% in the general population.2 There is also an increased risk of developing other common malignancies such as gastric, endometrial and ovarian carcinoma. Individuals diagnosed with LS are offered at least regular colonoscopy surveillance, with or without additional more intensive surveillance programmes.3

LS typically presents as an autosomal-dominant trait and is caused by germline mutations in genes encoding DNA mismatch repair (MMR) proteins—principally MLH1, PMS2, MSH2 and MSH6.3 These genes encode proteins that repair errors introduced during DNA replication and recombination, the loss of which results in a phenotype of hypermutation and variation at DNA microsatellites—a phenomenon known as microsatellite instability (MSI). In addition to LS cases, around 15% of sporadic CRCs display MSI as a result of MLH1 loss secondary to promoter hypermethylation—important clinically as it predicts favourable prognosis and lack of benefit from adjuvant chemotherapy in stage II disease.4 Thus, while testing of CRCs for MSI by PCR of DNA microsatellites or by immunohistochemistry (IHC) for MMR protein expression identifies patients who may have LS, most cases will be sporadic in aetiology.2 ,3 Informed by this consideration, one common strategy for LS screening is to test tumour MMR protein expression by IHC in patients meeting the Revised Bethesda criteria (box 1).2 ,3 Cancers with MLH1 loss (the majority of cases) are then tested for the BRAFV600E mutation, which is present in almost all sporadic MSI CRCs with MLH1 promoter hypermethylation, but very rare in LS-associated tumours with germline MLH1 mutation.5 ,6 Loss of MSH2 or MSH6, either alone or (more commonly) in combination, suggests germline LS mutation, as do rare cases of PMS2 loss in the absence of MLH1 loss.3 ,7 Patients in whom germline MMR LS mutation is suspected following MMR IHC±BRAF testing are referred to clinical genetics for counselling and MMR gene and MLH1 promoter methylation testing as required. This screening protocol (see figure 1) has been recommended internationally and recently endorsed by the Royal College of Pathologists in the updated 2014 dataset for CRC histopathology reporting.8

Box 1

The latest Revised Bethesda guidelines for immunohistochemical screening for microsatellite instability (MSI)

  1. Colorectal carcinoma (CRC) diagnosed at younger than 50 years.

  2. Presence of synchronous or metachronous CRC or other Lynch syndrome (LS)-associated tumours.

  3. CRC with MSI-high pathologic-associated features (Crohn-like lymphocytic reaction, mucinous/signet cell differentiation, or medullary growth pattern) diagnosed in an individual younger than 60 years old.

  4. Patient with CRC and CRC or LS-associated tumour diagnosed in at least one first-degree relative younger than 50 years old.

  5. Patient with CRC and CRC or LS-associated tumour at any age in two first-degree or second-degree relatives.

  6. Patients meeting one or more criterion should receive tissue screening. LS-associated tumours include colon, rectum, stomach, ovary, endometrium, pancreas, uterus, kidney, biliary tract, brain, small bowel and some skin tumours.

Figure 1

Flow chart of the current recommended screening pathway for Lynch syndrome (LS) using microsatellite instability (MSI) and BRAF testing.2 1, regardless of PMS2 expression; 2, regardless of MLH1 or PMS2 expression. IHC, immunohistochemistry.

However, despite these recommendations, LS screening within the UK National Health Service (NHS) is patchy and inconsistent, with many differing screening protocols in existence (see box 2). Many centres lack the funding and facilities to perform the additional IHC or BRAF mutation testing and the uncertainty around this is a barrier to implementation. Furthermore, a recent UK systematic review of the cost-effectiveness of the various MSI and LS screening approaches concluded that while there was an overall financial benefit to testing, data on the financial impact of individual protocols were lacking, and called for more research into the issue.9

Box 2

Protocols currently in existence for screening and diagnosing Lynch syndrome (LS) in the UK9

▸ Clinical screening (Amsterdam or Bethesda), then confirmatory mismatch repair (MMR) sequencing.

▸ Immunohistochemical (IHC) screening then, confirmatory MMR sequencing.

▸ IHC with BRAF mutation PCR screening, (if appropriate), then confirmatory MMR sequencing.

▸ Microsatellite instability (MSI) PCR screening, then confirmatory MMR sequencing.

▸ Combined MSI and BRAF PCR, then confirmatory MMR sequencing.

▸ Combined IHC, BRAF and MSI PCR screening, then confirmatory MMR sequencing.

▸ Universal MMR sequencing.

Until recently, our institution only offered clinical genetics referral to patients with CRC predicted to have high-risk of LS according to the Revised Bethesda criteria. As we routinely performed MMR IHC on stage II CRCs to guide clinical decisions regarding adjuvant chemotherapy, we expanded this to include all cases that fulfil the Revised Bethesda criteria and worked with the Clinical Genetics Department to streamline LS screening informed by these results, using existing resources. In this audit, we assessed the potential financial impact of this model of LS screening, taking into account the costs of its implementation compared with the potential savings from reduced referrals to clinical genetics.

Methods

We carried out a retrospective audit of all CRCs at our institution in the calendar year 2013. Inclusion criteria were colorectal adenocarcinoma resections that had undergone IHC screening for MMR protein expression. Referred cases were excluded. Costing information was acquired from the departmental budget leads of the Department of Cellular Pathology, the Department of Clinical Genetics at Oxford University Hospitals NHS Trust and from the Oxford Molecular Diagnostics Centre.

Results

We received 284 resections in 2013. In total, 209/284 cases (74%) were either stage II CRC (pT3N0 and pT4N0) or above or had met the Revised Bethesda criteria. All 209 cases had IHC for MLH1, PMS2, MSH2 and MSH6. A summary of the results can be found in table 1.

Table 1

A breakdown of the results from the 2013 Lynch syndrome (LS) screening audit

In total, 47/209 cases (22%) showed loss of expression of at least one of the four IHC markers. Also, 44/47 cases (94%) had loss of MLH1 expression with all but one showing concurrent loss of PMS2. One case showed loss of PMS2 alone. Two cases showed loss of both MSH2 and MSH6 concurrently.

BRAF V600E mutation testing by PCR using the Cobas 4800 system (Roche Molecular Diagnostics) was funded in 28 MLH1-negative cases. Of these, two tested negative and were therefore eligible for clinical genetics review. The 16 cases not tested for BRAF V600E were regarded eligible for direct referral to clinical genetics. The one case with isolated PMS2 loss and the two cases showing concurrent loss of MSH2 and MSH6 brought the total eligible for clinical assessment to 21.

A summary of the expenditure can be seen in table 2. At our institution, referral to a genetics clinic (£540) and germline testing (£800) had a cost of £1340 per patient. Had the clinical screening alone option (see box 2) been followed, genetics referral and germline testing of all 209 cases would have cost £280 060.

Table 2

A breakdown of the costs of Lynch syndrome (LS) screening during the audit period

The unit price of cutting and staining a slide for a single antibody is £15. Therefore, the cost we incurred for four IHC stains on each of the 209 cases was £12 540. This reduced the number of patients needing clinical genetics follow-up from 209 to 47, reducing the referral cost down to £69 980. With IHC costs included, the total expenditure is £75 520, and therefore, IHC screening alone had a potential annual saving of £204 540.

The cost of cutting multiple sections for PCR (£20) and referring the case for BRAF V600E testing (£160) had a unit price of £180 per case. Of the 44 cases with MLH1 loss, 28 cases had BRAF V600E tested and this incurred a cost of £5040. By additional BRAF V600E testing, it had reduced the number of 28 potential patients for clinical genetics assessment down to two, therefore, reducing the referral cost further by £34 840. The clinical follow-up, combined with the cost for IHC and PCR, gives a total expenditure for LS screening in 2013 of £45 720. Therefore, the total potential annual saving (compared with clinical assessment alone) for our institution was £234 340.

Discussion

Inherited cancer syndromes are increasingly recognised by clinicians and the public alike.2 ,3 ,8–10 There is a now a national and international push to screen for LS, but current protocols are inconsistent and their implementation patchy. For many centres in the UK, a lack of funding and perceived high costs are the main obstacles to implementing LS screening.2 ,3 ,5–9 However, a recent NHS Health Technology Assessment recognised that screening for LS (by any protocol) and preventing further LS-related tumours is cost-effective in the long term. However, the report stressed the need for further evidence comparing the financial impact of individual screening protocols.9

IHC for MMR proteins can be used to screening for LS and is also useful in guiding decisions on adjuvant therapy after surgery for stage II CRC.2–4 The updated 2014 dataset for CRC histopathology reporting by the Royal College of Pathologists8 noted that while there is a strong case for performing MMR IHC on all newly diagnosed early CRCs, MMR status is currently not considered a core data item due to the resource implications of reflex testing. In our institution, due to the lack of funding for reflex testing all CRCs, we have implemented a four-antibody IHC panel for MMR testing in all stage II tumours for prognostication, together with cases where patients met the Revised Bethesda criteria for LS screening—representing 74% of CRC resections in our centre during the study period. As noted previously, additional BRAF V600E PCR testing in cases with MLH1 loss was only available for a limited period of our audit. The direct cost of additional IHC to the department during the study period was £12 540, and the cost of BRAF PCR was £5040. Had funding been in place to test all 44 tumours with MLH1 loss for BRAF mutations (£7920), the overall departmental cost would have only increased by £2880. While few centres in the UK or elsewhere offer full genetic assessment as first-line screening for LS, in the absence of alternatives this is considered a valid approach.9 Modelling the cost implications of this strategy in our cohort, the 209 cases potentially eligible for genetics review at £1340 per patient would have resulted in a total cost of £280 060. Triage by targeted IHC screening alone using our approach reduced the number of potential referrals to 47, corresponding to a theoretical total annual saving of £234 340 for our institution. Had funding been available for BRAF testing in the 16 cases with MLH1 loss in which it was indicated but not performed, the cost savings are likely to have been greater, as only 7% of MLH1-negative CRCs tested for BRAF mutation needed genetics referral. However, it must be noted that this is likely to overstate the real-world savings of our approach, given that most centres do not refer all patients with CRC for genetic testing for LS.

Our cost modelling has limitations. For instance, not all patients suspected of carrying LS will typically agree to clinical genetics referral, and of these, not all will consent to germline mutation testing following counselling. While we were unable to examine this specifically in our audit, our experience is that roughly half of patients offered will take up the opportunity in our trust. However, these caveats notwithstanding, our strategy is still likely to result in substantial cost savings compared with referring all patients with CRC for genetics assessment. Similarly, while it was beyond the scope of our study to fully explore the downstream financial impacts of our model of LS screening, the increased costs associated with intensive surveillance in patients in whom LS is confirmed are likely to be outweighed by the savings as a result of preventing future cancers and the associated costs of surgery and chemotherapy.

Our strategy represents a pragmatic approach developed within the funding constraints of the UK NHS practice. As such, we would argue that the potential cost savings are immediately realisable, in contrast to other promising strategies for LS screening such as the use of cancer gene panels, which, while potentially cost saving,11 are likely to be further from clinical implementation. While we note that smaller numbers of cases in many smaller units will have greater costs per IHC stain and limited access to molecular testing, this can be easily circumvented by a tissue referral service for MMR IHC and BRAF mutation testing at cost price—a service we provide to district general hospitals in our region.

Conclusions

In this study, we presented our experience of the introduction of LS screening to our CRC histopathology service using MMR IHC and BRAF mutation testing to triage referrals to clinical genetics. As CRCs are one of the commonest oncological surgical specimens, the pathologist has an pivotal role in initiating MSI testing and, consequently, LS screening. Our data suggest that, when resources are limited, and MMR IHC is principally done for prognostication in stage II CRC, our model may both facilitate LS screening and to limit unnecessary clinical genetics referrals.

Take home messages

  • Lynch syndrome (LS) screening is currently fragmented and inconsistent in the UK.

  • Reflex mismatch repair immunohistochemistry (MMR IHC) on all colorectal cancer (CRC) is now recommended by the Royal College of Pathologists, but MMR IHC status remains a non-core item for all CRCs due to resource implications.

  • LS screening through selective MMR IHC model performed for prognostication in stage II CRC can be implemented when funding is limited, especially when expenditure has to be justified by impact on oncological decisions.

References

Footnotes

  • Handling editor Cheok Soon Lee

  • Contributors RC and LMW conceived the project. RC collected the data with help from LM. DNC assisted with the costing information. RC drafted the manuscript. All authors contributed to editing the manuscript, and RCh and LMW oversaw the project.

  • Competing interests None declared.

  • Provenance and peer review Not commissioned; externally peer reviewed.