LBA16 KEYNOTE-355: Final results from a randomized, double-blind phase III study of first-line pembrolizumab + chemotherapy vs placebo + chemotherapy for metastatic TNBC

A prior interim analysis of KEYNOTE-355 (NCT02819518) showed that first-line pembrolizumab (pembro) + chemotherapy (chemo) significantly improved PFS vs placebo (pbo) + chemo in patients (pts) with metastatic TNBC whose tumors expressed PD-L1 (CPS ≥10) (HR, 0.65, 95% CI, 0.49–0.86; one-sided P=0.0012 [P-value boundary, 0.00411]). We present final results for the dual primary endpoint of OS and other study endpoints. 847 pts with de novo metastasis or ≥6 mo DFI were randomized 2:1 to pembro + chemo (nab-paclitaxel, paclitaxel, or gemcitabine-carboplatin) or pbo + chemo for up to 35 administrations of pembro/pbo or until progression/intolerable toxicity. Pts were stratified by chemo type (taxane or gemcitabine-carboplatin), PD-L1 status (CPS ≥1 or <1) and prior (neo)adjuvant treatment with same-class chemo (yes or no). Dual primary endpoints are PFS (RECIST v1.1 by BICR) and OS in pts with PD-L1+ tumors (CPS ≥10 and ≥1) and all pts (ITT). ORR is a secondary endpoint. AEs were monitored throughout the study and graded per NCI CTCAE v4.0. As of June 15, 2021, median follow-up was 44.1 mo. Pembro + chemo significantly improved OS vs chemo alone in pts with CPS ≥10 tumors (Table). The p-value boundary for a significant OS benefit of pembro + chemo in pts with CPS ≥1 tumors was not met and formal testing in ITT was not performed. The benefit of pembro + chemo on PFS was consistent with the prior results. Pembro + chemo improved ORR in pts with CPS ≥10 tumors. For all endpoints, the pembro treatment effect increased with PD-L1 enrichment. Grade 3-5 treatment-related AE rates were 68.1% with pembro + chemo (2 deaths) vs 66.9% with chemo (0 deaths). Pembro + chemo showed a statistically significant and clinically meaningful improvement in OS vs chemo alone in pts with previously untreated locally recurrent inoperable or metastatic TNBC whose tumors expressed PD-L1 (CPS ≥10). No new safety signals were identified.Table: LBA16CPS ≥10CPS ≥1ITTP + Cn = 220Cn = 103P + Cn = 425Cn = 211P + Cn = 566Cn = 281OS, mo, mediana(95% CI)23.0 (19.0 – 26.3)16.1 (12.6 – 18.8)17.6 (15.5 – 19.5)16.0 (12.8 – 17.4)17.2 (15.3 – 19.0)15.5 (13.9 – 17.2)OS, HRb(95% CI)0.73c (0.55 – 0.95)0.86d (0.72 – 1.04)0.89 (0.76 – 1.05)PFS, mo, mediana(95% CI)9.7 (7.6 – 11.3)5.6 (5.3 – 7.5)7.6 (6.6 – 8.0)5.6 (5.4 – 7.4)7.5 (6.3 – 7.7)5.6 (5.4 – 7.2)PFS, HRb(95% CI)0.66 (0.50 – 0.88)0.75 0.62 – 0.910.82 0.70 – 0.98ORR, %(95% CI)52.7 (45.9 – 59.5)40.8 (31.2 – 50.9)44.9 (40.1 – 49.8)38.9 (32.2 – 45.8)40.8 (36.7 – 45.0)37.0 (31.4 – 42.9)aFrom Kaplan-Meier method.bFrom a stratified Cox model.cP=0.0093; boundary of P=0.0113 met.dP=0.0563; boundary of P=0.0172 not met. Open table in a new tab