POSTER: Clinical Validation of an Image Analysis Assay for Determining Programmed Death-Ligand 1 (22C3) in Non-small Cell Lung Cancer


Determination of programmed death-ligand 1 (PD-L1) level in tumor by immunohistochemistry (IHC) is widely used to predict response to check point inhibitor therapy. In particular, the Dako PD-L1 (22C3) antibody is a common companion diagnostic to the monoclonal antibody drug pembrolizumab in non-small cell lung cancer (NSCLC). However, for the practicing pathologist, interpretation of the PD-L1 (22C3) assay is cumbersome and time consuming.

Manual pathologist scoring also suffers from poor intra- and inter-pathologist precision. In this clinical validation study, we developed an image analysis (IA) based solution to accurately and precisely score digital images obtained from PD-L1 (22C3) stained NSCLC tissues for making clinical enrollment decisions. As a follow-up, we tested the concordance of the validated IA solution with manual pathology assessment.

Results and Conclusions:

The clinical validation study results show that IA can yield high analytical sensitivity, specificity, accuracy, and precision in the determination of the PD-L1 score as > 90% of the tissue cohort met the passing criteria for all validation parameters.

The concordance testing results show that digital TPS significantly correlated with the manual TPS evaluation obtained from 3 different pathologists (Pearson’s r > 0.8, p < 0.001). The concordance of the digital scoring across 3 repeated scoring of the same samples was also higher than the intra- and inter-pathologist concordance.

In conclusion, we have developed a clinically validated PD-L1 (22C3) IA assay that performs equivalently to manual pathologist scoring, requires less labor, and bypasses intra- and inter-pathologist variability by producing standardized and reproducible results with greater precision across repeated scoring.