Performance analysis of a novel PD-L1 CAL10 assay in esophageal cancer

Background: Esophageal cancer (ESCC) remains a highly lethal disease due to late diagnosis and limited responsiveness to conventional therapies. Recent studies have highlighted the role of immune checkpoint pathways in its pathogenesis, particularly the PD-1/PD-L1 interaction, which enables tumor cells to escape immune surveillance. We have developed a PD-L1 CAL10 assay for immunohistochemical (IHC) detection of PD-L1 in ESCC specimens. Here, we have compared this novel PD-L1 CAL10 assay for IHC detection of PD-L1 in ESCC specimens with two common research-utilized PD-L1 IHC assays. Design: The primary objective was to compare the PD-L1 CAL10 assay with other commercially available assays, SP263 and 22C3, at ≥1% tumor proportion score (TPS) and ≥50% TPS cutoffs. Also, we measured interobserver/interpathologist agreement (secondary objective) for each assay type at ≥1% TPS and ≥50% TPS cutoffs. Seventy (70) whole tissue formalin fixed paraffin embedded (FFPE) unique ESCC cases were included in the study. Each case was serially sectioned at 4 µm. Each of these sections was stained with H&E and anti-PD-L1 antibodies: CAL10 on Leica Biosystems BOND RX, SP263 on Ventana BenchMark ULTRA, and Agilent/Dako's 22C3 optimized for use on a Ventana Discovery Ultra. These tissue-mounted glass slides were scored manually by 3 pathologists, with a washout period of two weeks between reads for each clone. In comparison to other assays, the PD-L1 CAL10 stained slides had no significant background or artifacts. Each pathologist scored independently, and the majority score for the PD-L1 status (positive or negative) was recorded. Results: The PD-L1 CAL10 assay showed comparable interassay and interreader variability (Table 1). Additionally, its scoring pattern was similar to the other two PD-L1 clones. Table 1. TPS agreement rates Interassay CAL10 vs. SP263 CAL10 vs. 22C3 SP263 vs. 22C3 ≥1% TPS cutoff Almost perfect (k= 0.82) Moderate (k=0.59) Substantial (k=0.67) OPA-91.2% OPA-79.7% OPA - 84.1% ≥50% TPS cutoff Moderate (k= 0.42) Slight (k=0.19) Substantial (k=0.65) OPA-92.6% OPA-89.9% OPA - 97.1% Interpathologist CAL10 SP263 22C3 ≥1% TPS cutoff Moderate (k=0.54) Substantial (k=0.71) Moderate (k=0.53) ≥50% TPS cutoff Moderate (k=0.48) Fair (k=0.37) Moderate (k=0.42) k, Cohen’s kappa statistic; OPA, Overall percent agreement Conclusion: The initial results suggest that the PD-L1 CAL10 assay is comparable to the other tested PD-L1 assays for ESCC specimens and can be a suitable option for further translational research evaluations.