LBA50 Analysis of pathological features and efficacy outcomes with neoadjuvant nivolumab (N) plus platinum-doublet chemotherapy (C) for resectable non-small cell lung cancer (NSCLC) in CheckMate 816

Abstract

A post hoc analysis from the phase 3 CheckMate 816 study showed that residual viable tumor (RVT) in the primary tumor (PT) was associated with event-free survival (EFS) with N + C (2y EFS rates: 90%, 60%, 57%, and 39% in patients [pts] with 0–5, > 5–30, > 30–80, and > 80% RVT). Here, we report analyses in pts with or without (w/o) pathologic evidence of lymph node involvement (LN-I). Adults with resectable NSCLC were randomized to receive N 360 mg + platinum-doublet C Q3W or C alone Q3W for 3 cycles, followed by resection. Primary endpoints were pathologic complete response (pCR) and EFS (both met). Analyses included efficacy by LN-I and assessment of prespecified histopathologic features (%RVT, regression, and necrosis) in PT and LN. Pts with LN-I could have had either 0% (pathologic evidence of completely regressed tumor) or > 0%RVT in the resected LN. A time-dependent ROC curve analysis assessed the predictive ability of %RVT in PT for 2y EFS. pCR rates were improved with N + C vs C in both pts with or w/o LN-I (table); similar EFS benefit was seen in pts with or w/o LN-I (HR: 0.69 and 0.74). In the N + C arm, pts with LN-I who had 0% RVT in both PT and LN showed the best EFS outcomes numerically, followed by those with 0% RVT in either PT or LN; those with > 0% RVT in both PT and LN showed shorter EFS (2y EFS rates: 92%, 76%, and 49%). Overall, lower %RVT and higher % regression (appears to be inversely related to RVT) was seen with N + C vs C in PT and LN; % necrosis did not differ between the treatment arms (table). In both pts with or w/o LN-I (N + C), %RVT in PT was predictive of EFS at 2y (AUC: 0.76 and 0.73). These findings further support neoadjuvant N + C vs C as a novel treatment option for pts with resectable NSCLC, regardless of LN-I. %RVT in PT in pts with or w/o LN-I was associated with EFS with N + C. Further assessment of clinically relevant %RVT cutoffs to predict long-term outcomes with immunotherapy is warranted.Table: LBA50Path-evaluableaW/o LN-IWith LN-IN + C (n = 72)C (n = 51)N + C (n = 68)C (n = 74)EfficacyMedian EFS (95% CI), moNR (30.6–NR)NR (22.4–NR)31.6 (22.2–NR)22.7 (14.8–NR)EFS HR (95% CI)0.74 (0.39–1.41)0.69 (0.42–1.13)pCR, % 0% RVT, PT 0% RVT, LN 0% RVT, PT and LN 0% RVT, PT or LN40 – – –6 – – –24 34 19 383 5 1 7Path features, median (Q1–Q3)%RVT PT LN1 (0–48) –65 (12–90) –35 (2–82) 40 (0–100)82 (50–94) 96 (61–100)% Regression PT LN80 (18–99) –20 (4–70) –40 (6–82) 20 (0–85)13 (4–44) 1 (0–15)% Necrosis PT LN4 (0–18) –2 (0–10) –1 (0–16) 0 (0–15)1 (0–6) 0 (0–2)a140 (78%) and 125 (70%) of concurrently randomized pts in N + C and C arms had path-evaluable samples from both PT and LN. NR, not reached. Open table in a new tab