Tumor mismatch repair immunohistochemistry and DNA MLH1 methylation testing of patients with endometrial cancer diagnosed at age younger than 60 years optimizes triage for population-level germline mismatch repair gene mutation testing

Abstract

Purpose: Clinicopathologic data from a population-based endometrial cancer cohort, unselected for age or family history, were analyzed to determine the optimal scheme for identification of patients with germline mismatch repair (MMR) gene mutations.

Patients and methods: Endometrial cancers from 702 patients recruited into the Australian National Endometrial Cancer Study (ANECS) were tested for MMR protein expression using immunohistochemistry (IHC) and for MLH1 gene promoter methylation in MLH1-deficient cases. MMR mutation testing was performed on germline DNA of patients with MMR-protein deficient tumors. Prediction of germline mutation status was compared for combinations of tumor characteristics, age at diagnosis, and various clinical criteria (Amsterdam, Bethesda, Society of Gynecologic Oncology, ANECS).

Results: Tumor MMR-protein deficiency was detected in 170 (24%) of 702 cases. Germline testing of 158 MMR-deficient cases identified 22 truncating mutations (3% of all cases) and four unclassified variants. Tumor MLH1 methylation was detected in 99 (89%) of 111 cases demonstrating MLH1/PMS2 IHC loss; all were germline MLH1 mutation negative. A combination of MMR IHC plus MLH1 methylation testing in women younger than 60 years of age at diagnosis provided the highest positive predictive value for the identification of mutation carriers at 46% versus ≤ 41% for any other criteria considered.

Conclusion: Population-level identification of patients with MMR mutation-positive endometrial cancer is optimized by stepwise testing for tumor MMR IHC loss in patients younger than 60 years, tumor MLH1 methylation in individuals with MLH1 IHC loss, and germline mutations in patients exhibiting loss of MSH6, MSH2, or PMS2 or loss of MLH1/PMS2 with absence of MLH1 methylation.

Fig 1.

Selection, recruitment, mismatch repair (MMR) tumor, and MMR gene germline testing of Australian National Endometrial Cancer Study (ANECS) population-based endometrial cancer cases. Details of nonpolymorphic MMR gene variants identified are shown in Appendix Table A1 (online only). The prior probabilities of pathogenicity for MSH6 variants of uncertain clinical significance from bioinformatics analysis as per Thompson et al are as follows: MSH6 c.3694_3696del p.(Val1232del), 0.97; MSH6 c.2314C>T p.(Arg772Trp), 0.89; MSH6 c.3725_3727del p.(Arg1242_Thr1243delinsPro), 0.99; MSH6 c.3469G>A p.(Gly1157Ser), 0.41. (For deletion variants, the prior probability of pathogenicity is that for the highest prior possible for a missense substitution at that residue.) No unclassified variants in the MLH1 or MSH2 genes were detected in this study. Methylation analysis was performed for all MMR-deficient patients with sufficient tumor DNA available for bisulfite conversion and methylation analysis and for a subset of 77 randomly selected patients with immuno-normal MMR protein tumor expression. The two immuno-normal cases (2.6%; 95% CI, 0.7% to 9.0%) demonstrating methylation showed 5.2% and 9.3% methylation compared with reference. In cases with loss of MLH1/PMS2 expression only, MLH1 methylation was detected in 99 of 111 (89.2%; 95% CI, 82.1% to 93.7%) mutation-negative cases, and in eight of 10 (80.0%; 95% CI, 49.0% to 94.3%) mutation-unknown cases. The individual with MSH6 loss and MLH1 methylation (reference value 11.9%) also demonstrated immunohistochemistry (IHC) loss of MLH1 and PMS2. Overall, a truncating gene mutation was identified in 22 of 158 MMR-deficient cases with germline DNA available for testing (13.9%, 95% CI, 9.4% to 20.2%), equating to an overall carrier frequency of 3.1% (95% CI, 2.1% to 4.7%). NEG, negative for MMR gene deleterious mutation or unclassified variant of uncertain clinical significance; POS, positive for MMR gene deleterious mutation; UNK, mutation status unknown (no germline DNA available for testing); UV, positive for MMR gene unclassified variant.

Fig 2.

Proposed scheme for identification of germline mismatch repair (MMR) mutation carriers at diagnosis of endometrial cancer. (*) Class 5 from qualitative or quantitative analysis according to criteria accepted by InSiGHT (http://www.insight-group.org/classifications; Thomson et al). (†) Consider microsatellite instability (MSI) testing for patients with high clinical suspicion. (‡) Unclassified variants can be submitted for further analyses at www.insight-group.org. IHC, immunohistochemistry; MSI-H, microsatellite-high instability; MSI-L, microsatellite-low instability; MSS, microsatellite stable.