65PDHigh tumor mutational burden (TMB) and PD-L1 have similar predictive utility in 2L+ NSCLC patients (pts) treated with anti-PD-L1 and anti-CTLA-4

Background: PD-L1 can predict benefit from anti-PD-1/L1 therapies; TMB is an emerging biomarker that may have additional predictive utility. We compared the utility of high TMB and PD-L1 to predict outcomes in NSCLC non-squamous pts treated with anti-PD-L1 (durvalumab) and anti-CTLA-4 monoclonal antibodies (tremelimumab, D+T). Methods: NCT02000947 was a ph 1b study in advanced NSCLC; this exploratory study focused on a cohort of IO-naive non-squamous NSCLC who progressed after 1 prior platinum-based therapy. Pts were treated with D 20 mg/kg Q4W up to 12 months and T 1 mg/kg Q4W with the first 4 cycles of D. By 20 OCT 2017, 214 pts received D+T. DNA sequencing was conducted with the FoundationOne CDxTM on 106 tumors and IHC was performed on tumoral PD-L1 on 193 biopsies using the validated SP263 Ventana assay, all of sufficient quality. NSCLC with ≥25% tumor cells stained for PD-L1 at any intensity were PD-L1+ and high TMB was the top tertile of nonsynonymous mutations per Mb (11.41). KM, Cox PH models and Spearman’s rho were calculated; RECIST v1.1 was used for response. Results: Of 200 biopsies evaluated for TMB, 106 (53%) provided usable data. PD-L1 was modestly correlated with TMB (r = 0.3, p = 0.002). PD-L1+ pts had higher ORR, PFS and OS compared to PD-L1-. TMB high pts had higher PFS than TMB low. HRs for PD-L1+ and TMB high were similar. TMB high/PD-L1+ pts did not have substantially better ORR than single positive pts, and TMB low/PD-L1- had the lowest response.Table: 65PD# Pts (# events [PFS])ORR % (95% CI)Median PFS, mo (95% CI)PFS HR (95% CI); pTMB high37 (27)29.7%7.1 (1.7,9.1)0.60 (0.38,0.96); 0.032TMB low69 (59)10.1%1.7 (1.6,3.6)PD-L1+57 (39)35.1%7.1 (3.5,9.2)0.55 (0.38,0.80); 0.0018PD-L1-136 (114)11.8%3.3 (1.7,3.6) Open table in a new tab Conclusions: The predictive utility of TMB was similar to that of PD-L1 in this cohort of NSCLC pts treated with D+T. This abstract was withdrawn from ASCO. Clinical trial identification: NCT02000947. Legal entity responsible for the study: Medimmune LLC. Funding: Medimmune LLC. Disclosure: B.W. Higgs, C.A. Morehouse, P.Z. Brohawn, S. Sridhar, R. Raja, G. Gao, J. Englert, K. Ranade: Employment: Medimmune/AstraZeneca; Stock ownership: AstraZeneca.