Systematic review finds postoperative immunotherapy adds months free of disease

Patients with completely resected non-small cell lung cancer gain an average of 1.3 to 2.0 additional months free from disease events during the first three years after starting postoperative immunotherapy, according to a systematic review published in Lung Cancer. Zhiheng Yao and colleagues reconstructed data from four randomized phase III trials involving 4,336 total participants to evaluate the actual magnitude of benefit provided by postoperative PD-1 or PD-L1 blockade. Rather than relying solely on traditional hazard ratios that show relative risk reductions, the investigators measured absolute disease-free survival benefits to help clinicians and patients discuss realistic outcomes.

Trial Policies and 4,336 Patients Evaluated

The systematic review evaluated four completed randomized phase III trial families encompassing 4,336 participants across different clinical policies. Those four trials consisted of CCTG BR.31, ALCHEMIST/ANVIL EA5142, IMpower010, and PEARLS/KEYNOTE-091. Those individual studies differed significantly in their molecular eligibility criteria, programmed death-ligand 1 selection thresholds, stage distributions, postoperative chemotherapy policies, comparators, masking methods, analysis populations, and disease-free survival definitions.

Treatment Improves Disease-Free Survival over 36 Months

By reconstructing published Kaplan–Meier curves, the authors estimated two primary outcomes at a 36-month time horizon. The estimated improvement in restricted mean disease-free survival ranged from 1.3 to 2.0 additional months across the trial policies. In terms of absolute probability, the treatment resulted in approximately 2.3 to 6.1 additional patients remaining free of a trial-defined disease-free survival event per 100 patients randomized.

Substantial Statistical Uncertainty Across Clinical Trials

Despite the average gains observed, the analysis revealed substantial statistical uncertainty when looking at individual trial arms. Confidence intervals crossed zero for two of the restricted mean survival time estimates and for three of the four disease-free survival probability estimates. Official hazard ratios were retained for secondary descriptive synthesis while overall survival and toxicity outcomes were evaluated separately due to differing definitions and follow-up windows across the studies.

Previous meta-analyses evaluating resectable non-small cell lung cancer mainly reported relative disease-free and overall survival effects without calculating trial-specific fixed-horizon absolute outcomes. Maturity data for overall survival remained unestablished across the evaluations.