High interobserver and intraobserver reproducibility among pathologists assessing PD-L1 CPS across multiple...

Area of Science:

• Oncology
• Pathology
• Immunohistochemistry

Background:

• Interobserver and intraobserver variability in scoring PD-L1 (programmed death-ligand 1) immunohistochemical staining is a common concern.
• Accurate PD-L1 scoring is crucial for patient selection in various cancer therapies.

Purpose of the Study:

• To assess the interobserver and intraobserver reproducibility of PD-L1 scoring using the combined positive score (CPS) among trained pathologists.
• To evaluate the effectiveness of a standardized training program for PD-L1 CPS scoring.

Main Methods:

• Data from 456 pathologists worldwide collected between 2017-2019.
• Digital and face-to-face training provided, followed by scoring of blinded samples stained with PD-L1 IHC 22C3 pharmDx.
• Reproducibility assessed for gastric cancer (GC), cervical cancer (CC), urothelial cancer (UC), oesophageal cancer (OC), and head and neck squamous cell carcinoma (HNSCC) at specific CPS cutoffs.

Main Results:

• High average interobserver agreement (91.0%–94.1%) and intraobserver agreement (90.2%–96.6%) were observed across different cancer types and CPS cutoffs.
• For example, GC samples at CPS ≥1 showed 91.5% interobserver and 90.2% intraobserver agreement.
• HNSCC samples at CPS ≥20 demonstrated 86.5% interobserver and 90.5% intraobserver agreement.

Conclusions:

• The study demonstrates that standardized, face-to-face training is highly effective in improving PD-L1 CPS scoring reproducibility.
• Consistently high concordance rates support the reliability of trained pathologists in assessing PD-L1 expression across multiple cancer indications.
• The findings validate the training methodology for consistent PD-L1 biomarker assessment.