Abstract B033: Individual Patient Analysis of Clinical Efficacy for PD-1/PD-L1 Inhibitors as First-Line Standard Treatment in Advanced Solid Tumors Exhibiting Low PD-L1 Expression

Abstract Background: Immune checkpoint inhibitors (ICIs), which improve overall survival (OS) among patients with advanced solid tumors, have been approved by the US Food and Drug Administration (FDA) as standard first-line treatment in several cancers, regardless of programmed death ligand 1 (PD-L1) status. However, the benefits of ICIs for patients with low or negative PD-L1 expression remain unclear. Methods: This analysis encompassed cancers eligible for first-line ICI treatment without PD-L1-based restrictions. We included randomized, parallel-group phase III clinical trials that compared immunotherapy with conventional care for advanced solid malignancies until December 2023; all studies reported OS and performed PD-L1-based stratification. Summary effects were estimated using random-effects models. Reconstructed individual patient data (IPD) from reported Kaplan-Meier (KM) curves of OS and PFS using the "IPDfromKM" method, while also inferring unreported data for the PD-L1 low-expression group through the application of "KMSubtraction". Pooled analysis was conducted using the KM method with the log-rank test and restricted mean survival time difference (RMST-D) to assess ICI benefit under various cutoff values and scoring methods. This study is registered with PROSPERO (CRD42023441048). Results: In the study-level analysis, 39 randomized controlled trials involving 25,536 patients revealed that both high and low PD-L1-expressing subgroups derived a significant clinical benefit from ICIs (hazard ratios [HRs] were 0·69 and 0·81, respectively). There was no survival benefit for patients with low PD-L1 expression in the following types of cancers: esophageal squamous cell carcinoma (ESCC), biliary tract cancer (BTC), hepatocellular carcinoma (HCC), and melanoma. Conversely, patients with human epidermal growth factor receptor 2 (HER2)-negative gastroesophageal adenocarcinoma (GEA) exhibited a marginal benefit (HR=0·89, 95% confidence interval [CI]=0·80-0·99). In patient-level analysis, the ESCC combined positive score (CPS) <10 subgroup derived a significant OS benefit (HR 0·78, 95% CI=0·63-0·96, P=0·02; RMST-D=2·34 months); the ESCC tumor proportion score (TPS) <1% subgroup showed no significant improvement in OS benefit (HR=0·87, 95% CI=0·72-1·06, P=0·16). Among patients with HER2-negative GEA, no significant OS benefit was observed in the CPS <1, CPS <5, and CPS 1-4 subgroups (HRs=0·95, 0·89, and 0·97, respectively). There were significant OS benefits in the subgroups with CPS <10 (HR=0·867, 95% CI=0·754-0·996, P=0·0482; RMST-D=1·78 months, P=0·038) and CPS 1-9 (HR=0·83, 95% CI=0·73-0·95, P=0·0085; RMST-D=2·21 months, P=0·007). Interpretation: Study-level data suggest that first-line ICI benefits for low or negative PD-L1 expression subgroups with BTC, HCC, and melanoma require further evidence. Patient-level data indicate that, contrary to FDA guidelines, ESCCs with TPS <1% and HER2-negative GEAs with CPS <5 do not benefit from the addition of ICIs to conventional chemotherapy. More nuanced clinical trials and predictive biomarkers are warranted. Citation Format: Peng Wu, Chaoqi Zhang, Dongyu Li, Xuanyu Gu, Dexin Shang, Nan Sun, Jie He. Individual Patient Analysis of Clinical Efficacy for PD-1/PD-L1 Inhibitors as First-Line Standard Treatment in Advanced Solid Tumors Exhibiting Low PD-L1 Expression [abstract]. In: Proceedings of the AACR IO Conference: Discovery and Innovation in Cancer Immunology: Revolutionizing Treatment through Immunotherapy; 2025 Feb 23-26; Los Angeles, CA. Philadelphia (PA): AACR; Cancer Immunol Res 2025;13(2 Suppl):Abstract nr B033.