Rates of genotyping for KRAS, NRAS, BRAF, microsatellite instability (MSI), and mismatch repair (MMR) in metastatic colon cancer patients: Gaps and implications.

Abstract

e15123 Background: Molecular genotyping is essential to optimal treatment selection in metastatic colon cancer (mCC) as mutations in exons 2,3,4 of KRAS and NRAS (expanded RAS) and BRAF V600E predict lack of response to anti-HER2 therapy, and microsatellite instability high (MSI-H) predicts positive response to immune checkpoint inhibitor (ICPI)s. The National Comprehensive Cancer Network (NCCN) first recommended molecular genotyping for mCC in 2009 and by 2016 recommended all pts with mCC have expanded RAS, BRAF, and mismatch repair deficiency (MMRd) testing. This study updates genotyping rates in metastatic colon cancer (mCC) in a real-world practice setting. Methods: We performed a retrospective review of genomic testing patterns from pts with mCC diagnosed between January 2013 and December 2017 from 22 academic and community health centers in the United States who contributed to COTA’s de-identified Real World Evidence database. Results: 563 pts with mCC were identified in the Cota database over the 5-year period. 341 (61%) pts with mCC had testing for KRAS, 130 (23%) for NRAS, 177 (31%) for BRAF, and 297 (53%) for MMRd. Testing rates for NRAS, BRAF and MMRd all increased from 2013 to 2017 but remained suboptimal with undergenotyping rates of 48%, 45% and 12% respectively by 2017. 33 (6%) pts received cetuximab and/or panitumumab without prior testing for both KRAS and NRAS, despite established lack of benefit. Between 2016 and 2017, when NCCN recommended testing for KRAS, NRAS, BRAF and MMRd, only 24% of mCC pts were tested for all biomarkers. Conclusions: Although molecular genotyping rates increased over the described 5-year period, since 2016, less than one quarter of all mCC pts in this cohort received guideline recommended genotyping. Approaches that may improve complete genotyping rates in mCC, such as plasma-based comprehensive genomic profiling, may improve optimal treatment selection and should be tested in prospective trials.