Final RATIONALE-315 Analysis Supports Perioperative Tislelizumab
Key Takeaways
- Perioperative tislelizumab remained associated with longer overall survival, and event-free survival also remained improved at final follow-up.
- Survival patterns were generally similar across reported subgroups, while exploratory tumor mutational burden analyses did not show clear overall survival separation.
- Surgery, adjuvant treatment, adverse events, and immune-mediated toxicity were reported across the perioperative treatment course.
This randomized, double-blind, placebo-controlled phase III trial enrolled adults at 50 sites in China. Participants had histologically confirmed squamous or nonsquamous stage II-IIIA disease, ECOG performance status 0 or 1, and planned curative resection. Patients received three or four neoadjuvant cycles of tislelizumab or placebo with platinum-based chemotherapy, then surgery, then adjuvant tislelizumab or placebo. At the 7 March 2025 cut-off, median follow-up was 38.5 months, and endpoints included BICR-assessed event-free survival and BIPR-assessed major pathological response, with overall survival among secondary outcomes. This final analysis provided a mature survival readout for the intent-to-treat population.
Overall survival favored tislelizumab, with a hazard ratio of 0.65, a 95% confidence interval of 0.45 to 0.93, and P = 0.009. The 36-month overall survival rates were 79.3% with tislelizumab and 69.3% with placebo, and deaths occurred in 52 of 226 and 70 of 227 patients. Event-free survival also favored tislelizumab, with a hazard ratio of 0.58, a 95% confidence interval of 0.43 to 0.79, and median event-free survival not reached versus 30.6 months. The 36-month event-free survival rates were 64.7% and 48.0%, with 170 BICR-assessed events overall, including 72 and 98 events, respectively. At the final analysis, both overall survival and event-free survival favored the perioperative tislelizumab regimen.
Benefit was generally consistent across prespecified and post hoc subgroups, including histology, disease stage, PD-L1 expression, and central node status. An exploratory tumor mutational burden analysis did not show clear overall survival separation between TMB-high and TMB-low groups, and sample size limited that assessment. Study-defined surgery was performed in 190 of 226 patients in the tislelizumab arm and 173 of 227 in the placebo arm. Adjuvant treatment was received by 168 and 147 patients, while treatment-emergent adverse events leading to surgery cancellation occurred in 5 and 2 patients and those leading to surgery delay in 17 and 8. Postoperative complications occurred in 125 of 190 and 108 of 173 surgical patients, with 22 and 28 grade 3 or higher events, adding perioperative safety context.
Among treated patients, treatment-related adverse events occurred in 224 of 226 patients with tislelizumab and 225 of 226 with placebo, including grade 3 or higher events in 165 and 152 patients. Serious treatment-related adverse events occurred in 35 and 20 patients, discontinuations due to treatment-related adverse events in 29 and 21, and treatment-related deaths in 4 and 2. Immune-mediated adverse events occurred in 91 of 226 and 41 of 227 patients, with grade 3 or higher immune-mediated events in 21 and 7. Common immune-mediated events with tislelizumab included skin reactions, hypothyroidism, and pneumonitis, while serious immune-mediated events occurred in 24 and 5 patients. Two immune-mediated deaths were reported with tislelizumab and none with placebo, alongside longer survival and higher immune-related toxicity in the perioperative setting.