Therapeutic implications of acquired high tumor mutational burden (TMB-H) after targeted therapy (TT) in metastatic colorectal cancer (mCRC).

person Celine Yeh Memorial Sloan Kettering Cancer Center, New York, NY info_outline Celine Yeh, Emily Harrold, Fergus Keane, Jenna Cohen Sinopoli, Michael Bonner Foote, Andrea Cercek, Rona Yaeger

Authors person Celine Yeh Memorial Sloan Kettering Cancer Center, New York, NY info_outline Celine Yeh, Emily Harrold, Fergus Keane, Jenna Cohen Sinopoli, Michael Bonner Foote, Andrea Cercek, Rona Yaeger Organizations Memorial Sloan Kettering Cancer Center, New York, NY Abstract Disclosures Research Funding NIH T32 Clinical Scholars Biomedical Research Training Program (MSK) Background: mCRC treated w/ TT can acquire a high number of mutations at progression. TMB is used as a proxy of neoantigen load for predicting benefit from immune checkpoint blockade (ICB). In TMB-H CRC, benefit to ICB is restricted to microsatellite unstable (MSI-H) hypermutated or POL-D/E ultramutated tumors. We asked whether TT can sensitize microsatellite stable (MSS) CRCs to ICB by inducing an acquired TMB-H state. Methods: We screened the MSK-IMPACT dataset for mCRC patients (pts) who received TT and underwent pre-treatment (pre-TT) and post-progression (post-TT) tissue or liquid biopsies (LBx). We included pts who received TT against EGFR, BRAF V600E, KRAS G12C, or HER2. DNA sequencing of tissue and LBx was performed w/ MSK-IMPACT and Guardant 360. MSI status was assessed with algorithms measuring somatic changes in length of repetitive sequences (e.g., MSIsensor). Due to known discrepancies between tissue TMB (tTMB) and blood TMB (bTMB), we defined acquired TMB-H as: 1) tTMB <10 mut/Mb or bTMB <20 mut/Mb pre-TT; 2) ≥5-fold increase in TMB post-TT; 3) tTMB ≥10 mut/Mb or bTMB ≥40 mut/Mb post-TT. Results: We identified 26 pts w/ paired samples that had been sequenced w/ determination of TMB. 4/26 (15%) met criteria for acquired TMB-H. None had detectable pathogenic mismatch repair or DNA damage repair alterations pre-TT. All 4 post-TT samples were LBx only. One pt’s tumor acquired a somatic MLH1 mutation and converted from MSS to MSI-H. In all cases, the majority of acquired mutations were subclonal w/ many at variant allele frequencies (VAF) of ≤1% (Table). After detection of acquired TMB-H, 2/4 pts received ICB. Pt 1 had RAS wild type mCRC w/ pre-TT tTMB 6.9. She received irinotecan + panitumumab for 25.8 months. Post-TT bTMB was 40.19 and met criteria for MSI-H. She received pembrolizumab and experienced clinical + radiographic progression of disease (POD) within 49 days. Pt 2 had KRAS G12C-mutant mCRC w/ pre-TT bTMB 13.4 and tTMB 4.9. She received sotorasib + panitumumab for 3.2 months w/ post-TT bTMB 145.34. She received pembrolizumab + regorafenib and experienced clinical + radiographic POD within 37 days. Conclusions: In CRC, acquisition of TMB-H at progression w/ predominantly low VAF mutations suggests activation of enhanced mutagenesis due to therapeutic pressure from exposure to TT. In this context, TMB may not be an accurate reflection of tumor immunogenicity as it does not distinguish between clonal and subclonal alterations. Our data suggest that pts w/ acquired TMB-H following TT are unlikely to benefit from ICB. Pt # Pre-TT tTMB Pre-TT bTMB TT Regimen(s) Time on TT (months) Post-TT bTMB # Acquired Mutations % Acquired Mutations w/ VAF ≤1% 1 6.9 - Irinotecan + panitumumab 25.8 40.19 14 64 2 4.9 13.4 Sotorasib + panitumumab 3.2 145.34 44 100 3 7 - Adagrasib, sotorasib + trametinib, adagrasib + cetuximab 15.9 44.98 19 84 4 7.4 - Sotorasib + panitumumab 3.3 40.19 11 100