2018 ASCO Annual Meeting!
Session: Tumor Testing in Precision Oncology: From Heredity to Counseling to Implementation
Type: Clinical Science Symposium
Time: Monday June 4, 8:00 AM to 9:30 AM
Pan-cancer microsatellite instability to predict for presence of Lynch syndrome.
Cancer Prevention, Hereditary Genetics, and Epidemiology
2018 ASCO Annual Meeting
J Clin Oncol 36, 2018 (suppl; abstr LBA1509)
Author(s): Alicia Latham Schwark, Preethi Srinivasan, Yelena Kemel, Jinru Shia, Chaitanya Bandlamudi, Diana Mandelker, Marianne Dubard-Gault, Christina Tran, Sumit Middha, Jaclyn Frances Hechtman, Alexander Penson, Anna M. Varghese, Liying Zhang, Mark E. Robson, David B. Solit, Luis A. Diaz, Barry S. Taylor, Kenneth Offit, Michael F. Berger, Zsofia Kinga Stadler; Memorial Sloan Kettering Cancer Center, New York, NY
Background: The success of immunotherapy in microsatellite unstable (MSI-H) and/or DNA mismatch repair deficient (MMR-D) tumors has resulted in routine MSI-H/MMR-D testing in advanced solid tumors. Unlike colorectal (CRC) and endometrial cancer (EC), where this has long been undertaken, the characterization of Lynch syndrome (LS) across heterogeneous MSI-H/MMR-D tumors is unknown. Methods: Through a targeted NGS panel, MSI status was determined via MSIsensor. Scores of < 3, ≥3 to < 10, or ≥10 designated Microsatellite stable (MSS), MSI-Indeterminate (MSI-I) or MSI-H status, respectively. Germline mutations were assessed in MLH1, MSH2, MSH6, PMS2, EPCAM. Immunohistochemical staining (IHC) for MMR-D and tumor signatures in LS patients were assessed. Clinical variables were correlated with MSI and compared via Chi square or T-test. Results: Of 15,045 tumors spanning > 50 cancers , 93.2% were MSS, 4.6% MSI-I, and 2.2% MSI-H. Germline mutations were identified in 0.3% (37/14,020), 1.9% (13/699), and 16.3% (53/326) in the MSS, MSI-I, and MSI-H groups, respectively (p-value < 0.001). 25% of 1,025 MSI-H/MSI-I tumors were CRC/EC, but 50% (33/66) of LS patients had MSI-H/MSI-I tumors less commonly or not previously associated with LS (mesothelioma, sarcoma, adrenocortical, melanoma, ovarian germ cell). LS pts with MSI-H/MSI-I non-CRC/EC tumors only met testing criteria in 63.6% of cases, had lower MSIsensor scores, and were more likely to be MSI-I (MSI-I: non-CRC/EC, 30.3% (10/33) vs CRC/EC 9.1% (3/33); p-value = 0.03). IHC was completed in 86.4% (57/66) of LS MSI-H/MSI-I tumors, with 98.3% MMR-D-concordance. Of LS pts with MSS tumors, 78% had MSH6/PMS2 mutations, but 71% of LS pts with MSI-H/MSI-I tumors had MLH1/MSH2/EPCAM mutations(p-value < 0.001). 89% (33/37) of MSS tumors of LS pts had non-MMR-D signatures. Conclusions: MSI-H/MMR-D is predictive of LS across tumor types and suggests a more heterogeneous spectrum of LS-associated cancers than previously appreciated. Nearly 40% of LS pts with MSI-H/MMR-D non-CRC/EC tumors did not meet clinical criteria for genetic testing, suggesting that MSI-H/MMR-D tumors, regardless of cancer type or family history, should prompt germline testing for the evaluation of LS.