Performance of PD-L1 immunohistochemistry (IHC) assays in unresectable locally advanced or metastatic triple-negative breast cancer (mTNBC): Post-hoc analysis of IMpassion130

Abstract

Conclusions: Vel þ C/P demonstrated significant improvement in PFS over C/P alone. Median PFS for both arms was over 12 months. Pts on the Vel arm had durable benefit compared to control, with 26% of pts on Vel arm alive and progression-free at 3 years vs. 11% of pts on Pbo arm. Vel did not substantially alter the toxicity profile of C/P. Clinical trial identification: NCT02163694. Editorial acknowledgement: Medical writing support was provided by Ana Mrejeru, Ph.D., of AbbVie. Legal entity responsible for the study: AbbVie. Funding: AbbVie. Background: IMpassion130 is a Ph 3 study evaluating atezolizumab (A) þ nab-pacli-taxel (nP) vs placebo (P) þ nP as 1L treatment for patients (pts) with mTNBC. AþnP significantly improved PFS in PD-L1þ pts (PD-L1 positivity defined as PD-L1-stained immune cells [IC] 1% of the tumour area by VENTANA PD-L1 SP142 assay). PD-L1þ pts also had clinically meaningful OS benefit with AþnP (25 vs 18 mo; HR 0.71 [95% CI: 0.54, 0.93]; median follow-up, 18 mo; Schmid, ASCO 2019). In this exploratory post-hoc analysis, we evaluated the analytical concordance of SP142 with 2 other PD-L1 IHC assays, and their ability to predict clinical activity. Methods: Available samples from IMpassion130 were evaluated for PD-L1 status using VENTANA SP142 or SP263 IHC assay (IC 1%, SP142þ or SP263þ) or Dako PD-L1 IHC 22C3 assay (combined proportion score [CPS] 1, 22C3þ) by central laboratory in a biomarker-evaluable population (BEP). Results: A BEP of 614 pts (68% of ITT) was evaluable for PD-L1 status using the 3 assays. PD-L1þ prevalence was 46% for SP142þ, 81% for 22C3þ, and 75% for SP263þ. The overall percentage agreement (OPA) of SP142 with 22C3 and SP263 was 69% and 63%, respectively. PPAs of 98% for both assays suggest that SP142þ pts are captured by the other two tests, while NPAs were < 45%. The PFS HR (95% CI) was 0.60 (0.47, 0.78) in SP142þ pts, 0.68 (0.56, 0.82) in 22C3þ pts, and 0.64 (0.53, 0.79) in SP263þ pts. The OS HR (95% CI) was 0.74 (0.54, 1.01) in SP142þ pts, 0.78 (0.62, 0.99) in 22C3þ pts, and 0.75 (0.59, 0.96) in SP263þ pts. Subgroup outcomes of SP142þ and SP263þ or 22C3þ indicate that the PFS and OS benefit with AþnP in SP263þ/SP142-or 22C3þ/SP142-subgroups was smaller than in double-positive subgroups (table). Conclusions: At the evaluated cutoffs, 22C3 and SP263 assays identified more pts with PD-L1þ tumours. The pts in the SP142 PD-L1þ population, which is nested within the 22C3 and SP263 PD-L1þ populations, derived the greatest clinical benefit with AþnP.