Pathologic response as a surrogate endpoint for event-free survival in neo-adjuvant immunotherapy trials of resectable non-small cell lung cancer: A subgroup analysis

BackgroundA recent study demonstrated a robust correlation between pathologic complete response (pCR) and major pathologic response (MPR) with 2-year event-free survival (EFS) in neoadjuvant immune checkpoint inhibitor (ICI) trials of resectable non-small cell lung cancer (NSCLC). We analysed the impact of PDL1 status and histology on the strength of these associations.MethodsWe searched for randomised controlled trials (RCTs) examining neoadjuvant ICI in early stage NSCLC. Study eligibility criteria were: ≥80 patients, an ICI arm (including combination ICI plus chemotherapy), pCR/MPR and EFS outcomes reported for subgroups categorised by PDL1 status and histology. Using weighted linear regression, correlations between pCR and MPR against 2-year EFS were assessed. For trial level analysis, odds ratio (OR) for pathological responses and hazard ratio (HR) for EFS were assessed. Coefficient of determination (R2) of <0.6 was considered as poor correlation and >0.7 as strong correlation.ResultsWe identified five RCTs with a total of 2098 patients available for analysis. In the PDL1≥1% group, correlations between pCR and MPR vs 2-year EFS were strong (R2=0.82 and 0.96 respectively) but moderate in the PDL1<1% group (R2=0.67 and 0.64 respectively). In squamous NSCLC, correlations between pCR and MPR vs 2-year EFS were strong (R2=0.95 and 0.98). In non squamous NSCLC, correlation between pCR and 2-year EFS was weak (R2=0.55) but strong between MPR and 2-year EFS (R2= 0.79). For trial level analysis, correlation between pCR OR and HR EFS was strong in the PDL1<1% group but poor in the PDL1≥1% group (R2= 0.87 vs 0.46 respectively). Correlation was poor for pCR OR and HR EFS in both squamous and nonsquamous NSCLC (R2= 0.02 and 0.54 respectively). Limited observations for MPR and OS precluded these analyses.ConclusionsThere is strong correlation and surrogacy between pCR/MPR and EFS in resectable NSCLC with PDL1≥1%. pCR and MPR were prognostic for 2-year EFS outcomes in PDL1<1% and squamous subgroups but did not demonstrate good surrogacy at the trial level. Similarly pCR was a poor surrogate for EFS in nonsquamous NSCLC. Mature follow-up data will help to further validate pathological response as a surrogate endpoint.Legal entity responsible for the studyThe authors.FundingHas not received any funding.DisclosureC.K. Lee: Financial Interests, Personal, Advisory Board: AstraZeneca, Amgen, Roche, Janssen, Gilead, MSD, GSK, Norvatis, Merck kGA; Financial Interests, Institutional, Coordinating PI: AstraZeneca, Amgen, Merck kGA, Norvatis, Roche. All other authors have declared no conflicts of interest.