A Multi-Institutional Study to Evaluate Automated Whole Slide Scoring of Immunohistochemistry for Assessment of Programmed Death-Ligand 1 (PD-L1) Expression in Non–Small Cell Lung Cancer

Abstract

Assessment of programmed death-ligand 1 (PD-L1) expression is a critical part of patient management for immunotherapy. However, studies have shown that pathologist-based analysis lacks reproducibility, especially for immune cell expression. The purpose of this study was to validate reproducibility of the automated machine-based Optra image analysis for PD-L1 immunohistochemistry for both tumor cells (TCs) and immune cells. We compared conventional pathologists’ scores for both tumor and immune cell positivity separately using 22c3 antibody on the Dako Link 48 platform for PD-L1 expression in non–small cell lung carcinoma. We assessed interpretation first by pathologists and second by PD-L1 image analysis scores. Lin’s concordance correlation coefficients (LCCs) for each pathologist were measured to assess variability between pathologists and between pathologists and Optra automated quantitative scores in scoring both tumor and immune cells. Lin’s LCCs to evaluate the correlation between pathologists for TC was 0.75 [95% confidence interval (CI), 0.64-0.81] and 0.40 (95% CI, 0.40-0.62) for immune cell scoring. Pathologists were highly concordant for tumor scoring, but not for immune cell scoring, which is similar to previously reported studies where agreement is higher in TCs than immune cells. The LCCs between conventional pathologists’ read and the machine score were 0.80 (95% CI, 0.74-0.85) for TCs and 0.70 (95% CI, 0.60-0.76) for immune cell population. This is considered excellent agreement for TCs and good concordance for immune cells. The automated scoring methods showed concordance with the pathologists’ average scores that were comparable to interpathologist scores. This suggests promise for Optra automated assessment of PD-L1 in non–small cell lung cancer.

Improving the Diagnostic Accuracy of the PD-L1 Test with Image Analysis and Multiplex Hybridization

“…It is important to recognize that a sensitivity and specificity analysis used to assess the suitability of a test (multiplex) to a gold standard (PD-L1 DAB IHC) can only be as reliable as the reference test is capable of determining sample status without error [15]. A recent publication assessed the sensitivity and specificity of image analysis to the pathologist gold standard in 100 cases and their findings showed, as have other studies, that automated scoring was no worse than the concordance between pathologists [14,16]. As PD-L1 IHC is an imperfect test, where no other reference test or standards are available that fully confirm the pathologist's subjective score [9], a full comprehensive validation is required to verify multiplexing accuracy.…”

“…The complexity and ambiguity of the assessment of the PD-L1 diagnostic test is well reported [5,13,14]. The large variation in antibody clones, staining platforms, and assessment criteria plague pathology departments globally.…”

The path to a better biomarker: application of a risk management framework for the implementation of PD‐L1 and TILs as immuno‐oncology biomarkers in breast cancer clinical trials and daily practice

“…Evaluation of TILs in solid tumors is a highly suitable application for computational assessment; automated quantification by computer-based image analysis provides accurate and reproducible results that can aid pathologists, especially for borderline cases surrounding the clinically relevant 1% cut-off that are challenging to distinguish by eye. In the basic retrospective research realm, image analysis algorithms have shown better or comparable concordance between the automated algorithm score and the mean pathologist score than between pathologists [9,93]. Like any biomarker, computer-based image analysis algorithms would need to be analytically and clinically validated with demonstrated clinical utility such that results are consistent with trial materials used to established cutpoints for clinical decision-making and approved by corresponding regulatory agencies before they can be applied in the daily practice.…”

Deep learning approach for automated cancer detection and tumor proportion score estimation of PD-L1 expression in lung adenocarcinoma

“…Utility of computer-assisted diagnostics showed the improvement of quantitative and qualitative pathologic interpretation of IHC staining, such as algorithms for IHC scoring of HER2, indicating the enormous potential of AI in assisting the pathologist with objective IHC scoring for stratified medicine 23 . Digital pathology technique may show advantages at the quantification analysis of PD-L1 IHC status and reproducibility and accuracy for interpretations of PD-L1 IHC scoring in a combination with AI and deep learning algorithms, as demonstrated in recent studies [24][25][26] .…”