LAUSR

Abstract Background We previously showed pTMB selects patients (pts) deriving benefit from PD-L1 and CTLA-4 inhibition with durvalumab and tremelimumab (D+T) compared to best supportive care (BSC) in refractory mCRC in CCTG CO.26. Methods We investigated somatic variants contributing to pTMB in MSS pts using cell-free DNA (cfDNA) analysis performed with the GuardantOMNITMassay (2.1Mb) on 166 pts from CO.26. Results Median pTMB was 16.3 mutations/megabase (mts/Mb). Using a minimum p-value approach, pts with pTMB>28 mts/Mb (21% of MSS pts) had the greatest overall survival (OS) benefit for D + T (HR 0.34, 90% CI 0.18-0.63, p-interaction=0.070) and worse OS in the BSC arm (HR 2.59, 90% CI 1.46-4.62). Of 4044 mts detected, 2718 (67.2%) were subclonal ( 10.6 mts/Mb (14.1% pts). Similarly, pTMB calculated using only subclonal mts associated with D+T OS benefit (HR 0.32, 90% CI 0.16-0.67, p-interaction=0.057). Though many genes were associated with pTMB and improved efficacy (see Table), only BRCA1 mts predicted treatment effect (p-interaction=0.035). Most DNA repair mts were subclonal and not associated with increased clonal pTMB. Table . 528PD Gene Mutation Median pTMB MT vs WT Median Clonal pTMB MT vs WT HR (90% CI, p) Among Pts with Selected MT MSS Study Population (n = 166) 16.3 (95% CI: 14.4-20.1) 5.8 (95% CI: 4.8-5.8) 0.66 (0.49-0.89, 0.024) BRCA1 (n = 21) 23.0 vs 15.3 (P = 0.0006) 5.8 vs 5.8 (P = 0.64) 0.20 (0.07-0.58, 0.013) BRCA2 (n = 29) 28.9 vs 14.6 (P  5.8 vs 5.8 (P = 0.52) 0.39 (0.18-0.85, 0.047) ATM (n = 31) 31.6 vs 13.9 (P  6.7 vs 4.8 (P = 0.023) 0.66 (0.37-1.18, 0.24) Polymerase E/D1/Q/H (n = 20) 36.4 vs 14.7 (P  5.3 vs 5.8 (P = 0.57) 0.95 (0.44-2.05, 0.91) TP53 (n = 126) 17.2 vs 13.1 (P = 0.024) 4.8 vs 6.7 (P = 0.017) 0.63 (0.45-0.88, 0.024) MMR Gene without MSI (n = 31) 24.9 vs 14.6 (P  5.8 vs 5.8 (P = 0.85) 0.38 (0.17-0.85, 0.050) Conclusions Both clonal and subclonal mts causing an elevated pTMB associated with benefit from D+T and mts in select DNA repair genes may correlate with benefit. cfDNA provides an opportunity to evaluate tumor evolution potentially missed in archival tissue. Clinical trial identification NCT02870920. Legal entity responsible for the study Canadian Cancer Trials Group. Funding Canadian Cancer Society. Disclosure P.Z. Brohawn: Shareholder / Stockholder / Stock options, Full / Part-time employment: AstraZeneca. K. Banks: Shareholder / Stockholder / Stock options, Licensing / Royalties, Full / Part-time employment: Guardant Health. K. Quinn: Shareholder / Stockholder / Stock options, Licensing / Royalties, Full / Part-time employment: Guardant Health. All other authors have declared no conflicts of interest.