Benchmarking mismatch repair testing for patients with cancer receiving immunotherapy.

Abstract

179 Background: Mismatch repair deficiency (MMR-D) serves as a theranostic marker for directing immune checkpoint inhibitor (ICI) treatment across various cancer stages and tumor types. The College of American Pathologists endorses immunohistochemistry (IHC) as the primary approach for assessing mismatch repair (MMR) testing in colorectal (CRC) and endometrial cancer (EC) patients. An alternative method involves genetic evaluation of microsatellite instability (MSI). Previous studies have shown that resistance to immune checkpoint inhibitors (ICIs) may occur due to inaccurate MSI or MMR IHC status determination. Accurate assessment of MMR status is critical for treatment decision-making. Methods: We hypothesized that tumor-only next generation sequencing (NGS) for MMR mutation signature assessment for identification of MMR-D tumors is more sensitive than MMR IHC testing. Using a standardized, clinically validated, tumor-only NGS mutation signature pipeline in place in the Oncopanel assay at DFCI/BWH, we compared the concordance in MMR determination between NGS and IHC. We then evaluated a subset of ICI-treated patients to assess the clinical implications of these findings. Results: 1655 colorectal cancer (CRC) and endometrial carcinoma (EC) patients who had both NGS MSI (NGSd = deficient or NGSp = proficient) and MMR IHC (IHC- = absent or IHC+ = intact) assays performed were assessed. Discordant NGSd/IHC+ results were identified in 1% of CRC and 5% of EC cases that were IHC+. Importantly, patients with discordant NGSd/IHC+ (median OS: not reached [NR], 95% CI, 9.9 months-NR; median TTF: 44.4 months, 95% CI, 4.1-NR) and concordant NGSd/IHC- results, showed a similar overall survival and time to treatment failure with ICI treatment (p=non-significant) which contrasted with results for patients with concordant NGSp/IHC+ results (median OS: 14.9 months, 95% CI, 7.3-24.6, p-value 0.0025 and < 0.0001; median TTF: 5.0 months, 95% CI, 3.2-7.3; p-value 0.0014 and < 0.0001). Patients with EC or CRC and discordant NGSd/IHC+ results had better survival when treated with ICI-based regimens, compared to those who received non-ICI treatments (p= 0.054 and 0.017). Conclusions: In 2022, 151,030 new CRC and 65,950 new EC cases were diagnosed in the USA. Extrapolating our findings to this larger population, use of NGS mutational signature analysis instead of IHC would amount to about 1510 additional MMR-D CRC cases and 3297 additional MMR-D endometrial cases identified each year for whom first-line ICI therapy is preferred. Our findings support revisiting guideline recommendations for diagnostic testing of MMR to incorporate both IHC and targeted NGS tumor panel testing using sensitive and specific mutation signature calling algorithms. Larger prospective studies are warranted to confirm these findings in patients with mCRC and mEC treated with ICIs. Evaluation of this approach in other cancer types is also warranted.