Interobserver Reproducibility of PD-L1 Biomarker in Non-small Cell Lung Cancer: A Multi-Institutional Study by 27 Pathologists

Clinical Validation of Artificial Intelligence-Powered PD-L1 Tumor Proportion Score Interpretation for Immune Checkpoint Inhibitor Response Prediction in Non-Small Cell Lung Cancer

PURPOSE

Evaluation of PD-L1 tumor proportion score (TPS) by pathologists has been very impactful but is limited by factors such as intraobserver/interobserver bias and intratumor heterogeneity. We developed an artificial intelligence (AI)-powered analyzer to assess TPS for the prediction of immune checkpoint inhibitor (ICI) response in advanced non-small cell lung cancer (NSCLC).

MATERIALS AND METHODS

The AI analyzer was trained with 393,565 tumor cells annotated by board-certified pathologists for PD-L1 expression in 802 whole-slide images (WSIs) stained by 22C3 pharmDx immunohistochemistry. The clinical performance of the analyzer was validated in an external cohort of 430 WSIs from patients with NSCLC. Three pathologists performed annotations of this external cohort, and their consensus TPS was compared with AI-based TPS.

RESULTS

In comparing PD-L1 TPS assessed by AI analyzer and by pathologists, a significant positive correlation was observed (Spearman coefficient = 0.925; P < .001). The concordance of TPS between AI analyzer and pathologists according to TPS ≥50%, 1%-49%, and <1% was 85.7%, 89.3%, and 52.4%, respectively. In median progression-free survival (PFS), AI-based TPS predicted prognosis in the TPS 1%-49% or TPS <1% group better than the pathologist's reading, with the TPS ≥50% group as a reference (hazard ratio [HR], 1.49 [95% CI, 1.19 to 1.86] v HR, 1.36 [95% CI, 1.08 to 1.71] for TPS 1%-49% group, and HR, 2.38 [95% CI, 1.69 to 3.35] v HR, 1.62 [95% CI, 1.23 to 2.13] for TPS <1% group).

CONCLUSION

PD-L1 TPS assessed by AI analyzer correlates with that of pathologists, with clinical performance also being comparable when referenced to PFS. The AI model can accurately predict tumor response and PFS of ICI in advanced NSCLC via assessment of PD-L1 TPS.

A Deep Learning-Based Assay for Programmed Death Ligand 1 Immunohistochemistry Scoring in Non-Small Cell Lung Carcinoma: Does it Help Pathologists Score?

Several studies have developed various artificial intelligence (AI) models for immunohistochemical analysis of programmed death ligand 1 (PD-L1) in patients with non-small cell lung carcinoma; however, none have focused on specific ways by which AI-assisted systems could help pathologists determine the tumor proportion score (TPS). In this study, we developed an AI model to calculate the TPS of the PD-L1 22C3 assay and evaluated whether and how this AI-assisted system could help pathologists determine the TPS and analyze how AI-assisted systems could affect pathologists' assessment accuracy. We assessed the 4 methods of the AI-assisted system: (1 and 2) pathologists first assessed and then referred to automated AI scoring results (1, positive tumor cell percentage; 2, positive tumor cell percentage and visualized overlay image) for final confirmation, and (3 and 4) pathologists referred to the automated AI scoring results (3, positive tumor cell percentage; 4, positive tumor cell percentage and visualized overlay image) while determining TPS. Mixed-model analysis was used to calculate the odds ratios (ORs) with 95% CI for AI-assisted TPS methods 1 to 4 compared with pathologists' scoring. For all 584 samples of the tissue microarray, the OR for AI-assisted TPS methods 1 to 4 was 0.94 to 1.07 and not statistically significant. Of them, we found 332 discordant cases, on which the pathologists' judgments were inconsistent; the ORs for AI-assisted TPS methods 1, 2, 3, and 4 were 1.28 (1.06-1.54; P = .012), 1.29 (1.06-1.55; P = .010), 1.28 (1.06-1.54; P = .012), and 1.29 (1.06-1.55; P = .010), respectively, which were statistically significant. For discordant cases, the OR for each AI-assisted TPS method compared with the others was 0.99 to 1.01 and not statistically significant. This study emphasized the usefulness of the AI-assisted system for cases in which pathologists had difficulty determining the PD-L1 TPS.

Concordance of assessments of four PD-L1 immunohistochemical assays in esophageal squamous cell carcinoma (ESCC)

OBJECTIVE

Given real-world limitations in programmed death-ligand 1 (PD-L1) testing, concordance studies between PD-L1 assays are needed. We undertook comparisons of PD-L1 assays (DAKO22C3, Ventana SP263, Ventana SP142, E1L3N) among observers in esophageal squamous cell carcinoma (ESCC) to provide information on the analytical and clinical comparability of four PD-L1 IHC assays.

METHODS

Paraffin embedded samples of 50 cases of esophageal squamous cell carcinoma were obtained, satined with all four PD-L1 assays. PD-L1 was evaluated by 68 pathologists from 19 different hospitals. PD-L1 expression was assessed for combined positive score (CPS).

RESULTS

The expression sensitivity of SP263 was the highest in ESCC, followed by 22C3, E1L3N and SP142. Taking CPS 10 as the critical value, inter-observer concordance for CPS scores among 68 physicians was assessed for the 22C3, SP263, SP142, and E1L3N assays, yielding values of 0.777, 0.790, 0.758, and 0.782, respectively. In the comparison between assays, the overall CPS scores concordance rates between 22C3 and SP263, SP142, and E1L3N were 0.896, 0.833, and 0.853, respectively. 22C3 and SP263 have high concordance, with OPA of 0.896, while E1L3N and SP142 have the highest concordance, with OPA of 0.908.

CONCLUSION

In ESCC, the concordance of PD-L1 evaluation among observers is good, and the immune cell score is still an important factor affecting the concordance of interpretation among observers. Cases near the specific threshold are still the difficult problem of interpretation. SP263 had the highest CPS score of the four assays. SP263 cannot identify all 22C3 positive cases, but had good concordance with 22C3.E1L3N and SP142 showed high concordance.

Weakly Supervised Deep Learning Predicts Immunotherapy Response in Solid Tumors Based on PD-L1 Expression

Programmed death-ligand 1 (PD-L1) IHC is the most commonly used biomarker for immunotherapy response. However, quantification of PD-L1 status in pathology slides is challenging. Neither manual quantification nor a computer-based mimicking of manual readouts is perfectly reproducible, and the predictive performance of both approaches regarding immunotherapy response is limited. In this study, we developed a deep learning (DL) method to predict PD-L1 status directly from raw IHC image data, without explicit intermediary steps such as cell detection or pigment quantification. We trained the weakly supervised model on PD-L1-stained slides from the non-small cell lung cancer (NSCLC)-Memorial Sloan Kettering (MSK) cohort (N = 233) and validated it on the pan-cancer-Vall d'Hebron Institute of Oncology (VHIO) cohort (N = 108). We also investigated the performance of the model to predict response to immune checkpoint inhibitors (ICI) in terms of progression-free survival. In the pan-cancer-VHIO cohort, the performance was compared with tumor proportion score (TPS) and combined positive score (CPS). The DL model showed good performance in predicting PD-L1 expression (TPS ≥ 1%) in both NSCLC-MSK and pan-cancer-VHIO cohort (AUC 0.88 ± 0.06 and 0.80 ± 0.03, respectively). The predicted PD-L1 status showed an improved association with response to ICIs [HR: 1.5 (95% confidence interval: 1-2.3), P = 0.049] compared with TPS [HR: 1.4 (0.96-2.2), P = 0.082] and CPS [HR: 1.2 (0.79-1.9), P = 0.386]. Notably, our explainability analysis showed that the model does not just look at the amount of brown pigment in the IHC slides, but also considers morphologic factors such as lymphocyte conglomerates. Overall, end-to-end weakly supervised DL shows potential for improving patient stratification for cancer immunotherapy by analyzing PD-L1 IHC, holistically integrating morphology and PD-L1 staining intensity.

SIGNIFICANCE

The weakly supervised DL model to predict PD-L1 status from raw IHC data, integrating tumor staining intensity and morphology, enables enhanced patient stratification in cancer immunotherapy compared with traditional pathologist assessment.

PD-L1 Assessment in Needle Core Biopsies of Non-Small Cell Lung Cancer: Interpathologist Agreement and Potential Associated Histopathological Features

OBJECTIVE

Immune checkpoint inhibitors are used in the treatment of non-small cell lung cancer (NSCLC). Programmed cell death-ligand 1 (PD-L1) immunohistochemistry (IHC) assessed by pathologists is subject to interobserver variability. In advanced/metastatic disease and inoperable patients, PD-L1 assessment relies on biopsy specimens, commonly needle core biopsies (NCB). We aimed to determine the interobserver agreement for PD-L1 tumor proportion score (TPS) in NSCLC NCBs and identify histopathological features that may be related to interobserver variability.

MATERIAL AND METHODS

Sixty NSCLC NCBs with PD-L1 IHC were evaluated independently by four pathologists from different institutions. PD-L1 TPS was evaluated in three categories: no/low expression ( < 1%), intermediate expression (1%49%), and high expression (≥50%). Histological tumor type, necrosis, tumor-infiltrating lymphocytes, tumor length/percentage in the biopsy, and crush/squeeze artifact was evaluated.

RESULTS

The statistical analysis of the three PD-L1 TPS categories demonstrated moderate agreement (Fleiss Kappa 0.477) in the no/low category, fair agreement (Fleiss Kappa 0.390) in the intermediate category, and almost perfect agreement (Fleiss Kappa 0.952) in the high category. A significant correlation (p=0.003) was found between the crush/squeeze artifact in NCB and rate of discordant TPS categories. There was no significant correlation between pathologists' agreement in the TPS categories and histological tumor type, tumor length, tumor ratio, necrosis, and tumor-infiltrating lymphocytes.

CONCLUSION

Our results demonstrated moderate agreement among pathologists for the PD-L1 TPS 1% cut-off in NSCLC NCB, which is lower than that reported in resection materials. The presence of crush/squeeze artifact in NCBs is significantly related to the rate of discordant TPS categories, suggesting that PD-L1 assessment of pulmonary NCBs requires an awareness of this artifact.

Impact of Prolonged Ischemia on the Immunohistochemical Expression of Programmed Death Ligand 1 (PD-L1)

Antibodies targeting programmed death receptor 1 or programmed death ligand 1 (PD-L1) have become a standard of care to treat different cancers; for some of these tumors, there is a correlation between tissue expression of PD-L1 and response rates in patients. Although most of the analytical challenges in the evaluation of PD-L1 expression have been standardized, preanalytical issues have been less explored. The objective of this study was to evaluate the impact of time of ischemia on the performance of 2 commonly used antibodies against PD-L1. Sixteen tonsillectomy samples were kept in ischemia for <30 minutes from sample obtention (control) and 1, 3, 6, 12, and 24 hours at room temperature before formalin fixation and paraffin embedding. Selected areas were inserted into TMA paraffin recipient blocks stained with SP142 and SP263 antibodies and evaluated by 2 blind observers. The proportion of suboptimally stained samples was significantly higher for samples with cold ischemia times 6 hours or over ( P <0.0001). False-negative results were 25% in samples exposed to 6 hours of ischemia and raised to 34% for samples remaining in ischemia for 12 or 24 hours. When all observations were pooled, SP142 provided suboptimal results in 24% of observations and SP263 in 12.5%; this is a statistically significant difference ( P =0.042). In conclusion, the quality of staining for PD-L1 in tonsil samples varies with the time of cold ischemia. The SP142 antibody presented a significantly lower tolerance to prolonged cold ischemia than SP263. These results reveal the relevance of controlled preanalytical processing of samples.

EZH2 and POU2F3 Can Aid in the Distinction of Thymic Carcinoma from Thymoma

Thymic carcinoma is an aggressive malignancy that can be challenging to distinguish from thymoma using histomorphology. We assessed two emerging markers for these entities, EZH2 and POU2F3, and compared them with conventional immunostains. Whole slide sections of 37 thymic carcinomas, 23 type A thymomas, 13 type B3 thymomas, and 8 micronodular thymomas with lymphoid stroma (MNTLS) were immunostained for EZH2, POU2F3, CD117, CD5, TdT, BAP1, and MTAP. POU2F3 (≥10% hotspot staining), CD117, and CD5 showed 100% specificity for thymic carcinoma versus thymoma with 51%, 86%, and 35% sensitivity, respectively, for thymic carcinoma. All POU2F3 positive cases were also positive for CD117. All thymic carcinomas showed >10% EZH2 staining. EZH2 (≥80% staining) had a sensitivity of 81% for thymic carcinoma and a specificity of 100% for thymic carcinoma versus type A thymoma and MNTLS but had poor specificity (46%) for thymic carcinoma versus B3 thymoma. Adding EZH2 to a panel of CD117, TdT, BAP1, and MTAP increased cases with informative results from 67/81 (83%) to 77/81 (95%). Overall, absent EZH2 staining may be useful for excluding thymic carcinoma, diffuse EZH2 staining may help to exclude type A thymoma and MNTLS, and ≥10% POU2F3 staining has excellent specificity for thymic carcinoma versus thymoma.

A practical approach for PD-L1 evaluation in gastroesophageal cancer

PD-L1 is an established predictive immunohistochemical biomarker of response to immune checkpoint inhibitors. At present, PD-L1 is routinely assessed on biopsy samples of advanced gastroesophageal cancer patients before initiating first-line treatment. However, PD-L1 is still a suboptimal biomarker, due to changing cut-off values and scoring systems, interobserver and interlaboratory variability.

This practical illustrated review discusses the range of staining patterns of PD-L1 and the potential pitfalls and challenges that can be encountered when evaluating PD-L1, focusing on gastric and gastroesophageal adenocarcinoma (G/GEA) and esophageal squamous cell carcinoma (ESCC).

Modifying factors of PD-L1 expression on tumor cells in advanced non-small-cell lung cancer

BACKGROUND

Programmed death ligand-1 (PD-L1) expression predicts immunotherapy utility in nononcogenic addictive lung adenocarcinoma (ADC). However, its reproducibility and reliability may be compromised outside clinical trials. This study aimed to evaluate factors associated with PD-L1 expression in lung ADC.

METHODS

This observational study assessed 547 tumor samples with advanced lung ADC from January 2016 to December 2020 in a single cancer institution. Tumor samples were stained by at least one approved PD-L1 clone, SP263 (Ventana) or 22C3 (Dako), and stratified in tumor proportion score (TPS) <1%, 1-49%, or ≥50%.

RESULTS

Of all the tumor samples, positive PD-L1 staining was higher in poorly differentiated tumors (67.3% vs. 32.7%, p < 0.001). Analytical factors associated with a PD-L1 high expression (TPS ≥ 50%) were the SP263 clone (19.6% vs. 8.2%, p < 0.001), time of archival tumor tissue <12 months (15.3% vs. 3.8%, p = 0.024), whenever the analysis was performed in the most recent years (2019-2020) (19.0% vs. 8.3%, p < 0.001), and whenever the analysis was performed by pathologists in the academic setting (Instituto Nacional de Cancerologia, INCan) (19.9% vs. 11.9%, p = 0.001). In the molecular analysis, EGFR wild-type tumors had an increased proportion of PD-L1 positive and PD-L1 high cases (60.2% vs. 47.9%, p = 0.006 and 17.4% vs.8.5%, p = 0.004). A moderate correlation (r = 0.69) in the PD-L1 TPS% was observed between the two different settings (INCan vs. external laboratories).

CONCLUSION

Clinicopathological factors were associated with an increased PD-L1 positivity rate. These differences were significant in the PD-L1 high group and associated with the academic setting, the SPS263 clone, time of archival tumor tissue <12 months, and a more recent period in the PD-L1 analysis.

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

AIMS

A common concern among pathologists scoring PD-L1 immunohistochemical staining is interobserver and intraobserver variability. We assessed interobserver and intraobserver reproducibility of PD-L1 scoring among trained pathologists using combined positive score (CPS; tumour cell and tumour-associated immune cell staining).

METHODS AND RESULTS

Data were collected for 2 years (2017-2019) from 456 pathologists worldwide. Digital training encompassed unique, tumour-specific training and test sets. Samples were stained using PD-L1 IHC 22C3 pharmDx and evaluated at specific CPS cut-offs for gastric cancer (GC), cervical cancer (CC), urothelial cancer (UC), oesophageal cancer (OC), and head and neck squamous cell carcinoma (HNSCC). Pathologists underwent expert-to-peer training and scored 20 blinded samples on day 1 and 25 blinded samples on day 2 (including 15 of the day 1 samples). Interobserver and intraobserver reproducibility were assessed. For GC (120 observers) and CC (32 observers) samples assessed at CPS ≥1, average interobserver agreement was 91.5% and 91.0%, respectively, and average intraobserver agreement was 90.2% and 96.6%, respectively. For UC (139 observers) and OC (52 observers) samples measured at CPS ≥10, average interobserver agreement was 93.4% and 93.7%, respectively, and average intraobserver agreement was 92.0% and 92.5%, respectively. For HNSCC samples (113 observers), average interobserver agreement was 94.1% at CPS ≥1 and 86.5% at CPS ≥20; intraobserver agreement was 94.7% at CPS ≥1 and 90.5% at CPS ≥20.

CONCLUSION

The consistently high interobserver and intraobserver concordance rates support the effectiveness of face-to-face training of many global pathologists for scoring PD-L1 CPS across multiple indications at several specific cut-offs.

Artificial Intelligence-Assisted Score Analysis for Predicting the Expression of the Immunotherapy Biomarker PD-L1 in Lung Cancer

Programmed cell death ligand 1 (PD-L1) is a critical biomarker for predicting the response to immunotherapy. However, traditional quantitative evaluation of PD-L1 expression using immunohistochemistry staining remains challenging for pathologists. Here we developed a deep learning (DL)-based artificial intelligence (AI) model to automatically analyze the immunohistochemical expression of PD-L1 in lung cancer patients. A total of 1,288 patients with lung cancer were included in the study. The diagnostic ability of three different AI models (M1, M2, and M3) was assessed in both PD-L1 (22C3) and PD-L1 (SP263) assays. M2 and M3 showed improved performance in the evaluation of PD-L1 expression in the PD-L1 (22C3) assay, especially at 1% cutoff. Highly accurate performance in the PD-L1 (SP263) was also achieved, with accuracy and specificity of 96.4 and 96.8% in both M2 and M3, respectively. Moreover, the diagnostic results of these three AI-assisted models were highly consistent with those from the pathologist. Similar performances of M1, M2, and M3 in the 22C3 dataset were also obtained in lung adenocarcinoma and lung squamous cell carcinoma in both sampling methods. In conclusion, these results suggest that AI-assisted diagnostic models in PD-L1 expression are a promising tool for improving the efficiency of clinical pathologists.

Artificial intelligence-assisted system for precision diagnosis of PD-L1 expression in non-small cell lung cancer

Standardized programmed death-ligand 1 (PD-L1) assessment in non-small cell lung cancer (NSCLC) is challenging, owing to inter-observer variability among pathologists and the use of different antibodies. There is a strong demand for the development of an artificial intelligence (AI) system to obtain high-precision scores of PD-L1 expression in clinical diagnostic scenarios. We developed an AI system using whole slide images (WSIs) of the 22c3 assay to automatically assess the tumor proportion score (TPS) of PD-L1 expression based on a deep learning (DL) model of tumor detection. Tests were performed to show the diagnostic ability of the AI system in the 22c3 assay to assist pathologists and the reliability of the application in the SP263 assay. A robust high-performance DL model for automated tumor detection was devised with an accuracy and specificity of 0.9326 and 0.9641, respectively, and a concrete TPS value was obtained after tumor cell segmentation. The TPS comparison test in the 22c3 assay showed strong consistency between the TPS calculated with the AI system and trained pathologists (R = 0.9429-0.9458). AI-assisted diagnosis test confirmed that the repeatability and efficiency of untrained pathologists could be improved using the AI system. The Ventana PD-L1 (SP263) assay showed high consistency in TPS calculations between the AI system and pathologists (R = 0.9787). In conclusion, a high-precision AI system is proposed for the automated TPS assessment of PD-L1 expression in the 22c3 and SP263 assays in NSCLC. Our study also indicates the benefits of using an AI-assisted system to improve diagnostic repeatability and efficiency for pathologists.

Gastric Cancer: Mechanisms, Biomarkers, and Therapeutic Approaches

Gastric cancer (GC) remains one of the most common deadly malignancies worldwide. Recently, several targeted therapeutics for treating unresectable or metastatic GC have been developed. Comprehensive characterization of the molecular profile and of the tumor immune microenvironment of GC has allowed researchers to explore promising biomarkers for GC treatment and has enabled a new paradigm in precision-targeted immunotherapy. In this article, we review established and promising new biomarkers relevant in GC, with a focus on their clinical implications, diagnostic methods, and the efficacy of targeted agents.

Association of artificial intelligence-powered and manual quantification of programmed death-ligand 1 (PD-L1) expression with outcomes in patients treated with nivolumab ± ipilimumab

Assessment of programmed death ligand 1 (PD-L1) expression by immunohistochemistry (IHC) has emerged as an important predictive biomarker across multiple tumor types. However, manual quantitation of PD-L1 positivity can be difficult and leads to substantial inter-observer variability. Although the development of artificial intelligence (AI) algorithms may mitigate some of the challenges associated with manual assessment and improve the accuracy of PD-L1 expression scoring, use of AI-based approaches to oncology biomarker scoring and drug development has been sparse, primarily due to the lack of large-scale clinical validation studies across multiple cohorts and tumor types. We developed AI-powered algorithms to evaluate PD-L1 expression on tumor cells by IHC and compared it with manual IHC scoring in urothelial carcinoma, non-small cell lung cancer, melanoma, and squamous cell carcinoma of the head and neck (prospectively determined during the phase II and III CheckMate clinical trials). 1,746 slides were retrospectively analyzed, the largest investigation of digital pathology algorithms on clinical trial datasets performed to date. AI-powered quantification of PD-L1 expression on tumor cells identified more PD-L1-positive samples compared with manual scoring at cutoffs of ≥1% and ≥5% in most tumor types. Additionally, similar improvements in response and survival were observed in patients identified as PD-L1-positive compared with PD-L1-negative using both AI-powered and manual methods, while improved associations with survival were observed in patients with certain tumor types identified as PD-L1-positive using AI-powered scoring only. Our study demonstrates the potential for implementation of digital pathology-based methods in future clinical practice to identify more patients who would benefit from treatment with immuno-oncology therapy compared with current guidelines using manual assessment.

Comparison of Semi-Quantitative Scoring and Artificial Intelligence Aided Digital Image Analysis of Chromogenic Immunohistochemistry

Semi-quantitative scoring is a method that is widely used to estimate the quantity of proteins on chromogen-labelled immunohistochemical (IHC) tissue sections. However, it suffers from several disadvantages, including its lack of objectivity and the fact that it is a time-consuming process. Our aim was to test a recently established artificial intelligence (AI)-aided digital image analysis platform, Pathronus, and to compare it to conventional scoring by five observers on chromogenic IHC-stained slides belonging to three experimental groups. Because Pathronus operates on grayscale 0-255 values, we transformed the data to a seven-point scale for use by pathologists and scientists. The accuracy of these methods was evaluated by comparing statistical significance among groups with quantitative fluorescent IHC reference data on subsequent tissue sections. The pairwise inter-rater reliability of the scoring and converted Pathronus data varied from poor to moderate with Cohen’s kappa, and overall agreement was poor within every experimental group using Fleiss’ kappa. Only the original and converted that were obtained from Pathronus original were able to reproduce the statistical significance among the groups that were determined by the reference method. In this study, we present an AI-aided software that can identify cells of interest, differentiate among organelles, protein specific chromogenic labelling, and nuclear counterstaining after an initial training period, providing a feasible and more accurate alternative to semi-quantitative scoring.

Development and validation of a supervised deep learning algorithm for automated whole-slide programmed death-ligand 1 tumour proportion score assessment in non-small cell lung cancer

AIMS

Immunohistochemical programmed death-ligand 1 (PD-L1) staining to predict responsiveness to immunotherapy in patients with advanced non-small cell lung cancer (NSCLC) has several drawbacks: a robust gold standard is lacking, and there is substantial interobserver and intraobserver variance, with up to 20% discordance around cutoff points. The aim of this study was to develop a new deep learning-based PD-L1 tumour proportion score (TPS) algorithm, trained and validated on a routine diagnostic dataset of digitised PD-L1 (22C3, laboratory-developed test)-stained samples.

METHODS AND RESULTS

We designed a fully supervised deep learning algorithm for whole-slide PD-L1 assessment, consisting of four sequential convolutional neural networks (CNNs), using aiforia create software. We included 199 whole slide images (WSIs) of 'routine diagnostic' histology samples from stage IV NSCLC patients, and trained the algorithm by using a training set of 60 representative cases. We validated the algorithm by comparing the algorithm TPS with the reference score in a held-out validation set. The algorithm had similar concordance with the reference score (79%) as the pathologists had with one another (75%). The intraclass coefficient was 0.96 and Cohen's κ coefficient was 0.69 for the algorithm. Around the 1% and 50% cutoff points, concordance was also similar between pathologists and the algorithm.

CONCLUSIONS

We designed a new, deep learning-based PD-L1 TPS algorithm that is similarly able to assess PD-L1 expression in daily routine diagnostic cases as pathologists. Successful validation on routine diagnostic WSIs and detailed visual feedback show that this algorithm meets the requirements for functioning as a 'scoring assistant'.

Immunohistochemistry as predictive and prognostic markers for gastrointestinal malignancies

Biomarkers play a key role in the comprehensive pathologic evaluation of gastrointestinal malignancies. These biomarkers can be predictive, indicating whether a tumor is likely to respond to a particular therapy, or prognostic, providing information about the likely course and outcome of a disease. This review article will discuss available immunohistochemical stains for assessing these markers, including staining rationale, scoring criteria, associated systemic therapies, and pictorial examples. PD-L1, HER2, and mismatch repair status can be evaluated via immunohistochemistry for esophageal, gastric, and colorectal carcinomas. Biomarkers currently play a more limited role in evaluation of pancreatic and small bowel malignancies. Immunohistochemistry can also be used to evaluate biomarker status in gastrointestinal stromal tumors, gastrointestinal malignancies with NTRK gene fusions, and undifferentiated carcinomas with switch-sucrose non-fermentable complex abnormalities.

Programmed cell death-ligand 1 assessment in urothelial carcinoma: prospect and limitation

Programmed cell death protein 1/programmed death-ligand 1 (PD-1/PD-L1) inhibition has revolutionized the treatment paradigm of urothelial carcinoma (UC). Several PD-L1 assays are conducted to formulate appropriate treatment decisions for PD-1/PD-L1 target therapy in UC. However, each assay has its own specific requirement of antibody clones, staining platforms, scoring algorithms, and cutoffs for the determination of PD-L1 status. These prove to be challenging constraints to pathology laboratories and pathologists. Thus, the present article comprehensively demonstrates the scoring algorithm used and differences observed in each assay (22C3, SP142, and SP263). Interestingly, the SP142 score algorithm considers only immune cells and not tumor cells (TCs). It remains controversial whether SP142 expressed only in TCs truly accounts for a negative PD-L1 case. Moreover, the scoring algorithm of each assay is complex and divergent, which can result in inter-observer heterogeneity. In this regard, the development of artificial intelligence for providing assistance to pathologists in obtaining more accurate and objective results has been actively researched. To facilitate efficiency of PD-L1 testing, several previous studies attempted to integrate and harmonize each assay in UC. The performance comparison of the various PD-L1 assays demonstrated in previous studies was encouraging, the exceptional concordance rate reported between 22C3 and SP263. Although these two assays may be used interchangeably, a clinically validated algorithm for each agent must be applied.

Automated PD-L1 Scoring for Non-Small Cell Lung Carcinoma Using Open-Source Software

PD-L1 expression in non-small cell lung cancer (NSCLC) is predictive of response to immunotherapy, but scoring of PD-L1 immunohistochemistry shows considerable interobserver variability. Automated methods may allow more consistent and expedient PD-L1 scoring. We aimed to assess the technical concordance of PD-L1 scores produced using free open source QuPath software with the manual scores of three pathologists. A classifier for PD-L1 scoring was trained using 30 NSCLC image patches. A separate test set of 207 image patches from 69 NSCLC resection cases was used for comparison of automated and manual scores. Automated and average manual scores showed excellent correlation (concordance correlation coeffecient = 0.925), though automated scoring resulted in significantly more 1–49% scores than manual scoring (p = 0.012). At both 1% and 50% thresholds, automated scores showed a level of concordance with our ‘gold standard’ (the average of three pathologists’ manual scores) similar to that of individual pathologists. Automated scoring showed high sensitivity (95%) but lower specificity (84%) at a 1% threshold, and excellent specificity (100%) but lower sensitivity (71%) at a 50% threshold. We conclude that our automated PD-L1 scoring system for NSCLC has an accuracy similar to that of individual pathologists. The detailed protocol we provide for free open source scoring software and our discussion of the limitations of this technology may facilitate more effective integration of automated scoring into clinical workflows.

The Immunohistochemical Expression of Programmed Death Ligand 1 (PD-L1) Is Affected by Sample Overfixation

Humanized antibodies targeting programmed death receptor 1 (PD-1) or its ligand (PD-L1) have been approved for the treatment of different cancers. Some of these antibodies show a correlation between the tissue expression of PD-L1 and response. Evaluation of PD-L1 expression presents multiple challenges, but some preanalytical issues such as tissue fixation have been scarcely evaluated. With the hypothesis that immunohistochemical staining of PD-L1 may be impacted by the time of specimen fixation, we evaluated differences in its expression in tonsil samples exposed to predefined fixation times. Random nontumoral tonsillectomy specimens were blindly evaluated in tissue microarray slides after staining with SP142 and SP263 antibodies. With fixation times ranging from 12 to 72 hours, between 2.8% and 6.1% of the samples were considered to be suboptimally stained, with no differences between the 2 antibodies within these fixation times. A significantly higher proportion of samples exposed to a fixation time of 96 hours presented suboptimal immunostaining (15.6%, P<0.0001). In addition, suboptimally stained spots were 20.8% using SP142 and 10.4% using SP263 after 96 hours of fixation (P=0.046). In conclusion, the quality of staining for PD-L1 in tonsil samples decreased with overfixation of the specimen at times >72 hours. Samples exposed to formaldehyde for longer periods presented suboptimal results for both clones, but the SP142 antibody presented a significantly lower tolerance to formalin overexposure than SP263. These results indicate the relevance of a controlled preanalytical processing of samples and particularly the length of fixation of tumor specimens.

Interobserver agreement in programmed cell death-ligand 1 immunohistochemistry scoring in nonsmall cell lung carcinoma cytologic specimens

BACKGROUND

The evaluation of PD-L1 expression in nonsmall cell lung carcinoma (NSCLC) is becoming increasingly important given the effectiveness of PD-L1 inhibitors. Although cytologic specimens have been shown to be compatible with surgical specimens to evaluate PD-L1 immunohistochemistry (IHC), evidence of the reproducibility of PD-L1 in cytologic specimens is lacking. The aim of this study is to evaluate interobserver agreement in PD-L1 IHC in cytologic specimens.

METHODS

PD-L1 IHC was performed on 86 NSCLC cytology specimens using Dako PD-L1 IHC 22C3 pharmDx. The digitally scanned whole slide images (WSI) were read by five pathologists. Each case was given a Tumor Proportion Score (TPS) and the results were compared between the observers. The interobserver concordance was assessed using 1% and 50% as cutoffs.

RESULTS

TPSs were highly correlated among observers (Spearman correlation coefficient, 0.86-0.94). Using greater than 1% as a cutoff, interobserver agreement measured by Fleiss Kappa was 0.74 for all pathologists and Cohen's Kappa coefficient ranged from 0.49 to 0.83, consistent with moderate to substantial agreement. With a cutoff of greater than 50%, Fleiss Kappa was 0.79 for all pathologists and the kappa values ranged from 0.63 to 0.90, consistent with substantial to almost perfect agreement. Several pitfalls were identified by reviewing discordant cases, including staining in macrophages, stromal cells, and intratumoral heterogeneity.

CONCLUSION

Our data suggest that TPS of PD-L1 IHC on cytology specimens is reproducible, with a better agreement when using 50% as the cutoff value. However, special attention is required when the TPS is near the 1% cutoff.

Comparability of laboratory-developed and commercial PD-L1 assays in non-small cell lung carcinoma

PD-L1 expression in non-small cell lung cancer (NSCLC) is predictive of response to treatment with PD-1 and PD-L1 inhibitors. Different inhibitors have been developed with different PD-L1 assays, which use different PD-1 antibody clones on different immunohistochemistry platforms. Depending on instrument and reagent availability, laboratory-developed tests with cross-platform use of PD-L1 antibodies may have practical benefits over commercial assays. The 22C3 pharmDx Assay (referred to as 22C3 DAKO), the VENTANA PD-L1 SP263 Assay (referred to as SP263 VENTANA) and a lab-developed test using the 22C3 antibody on the VENTANA BenchMark ULTRA IHC/ISH system (referred to as 22C3 VENTANA) were performed on whole sections of 85 NSCLC surgical resections. All sections were independently scored by three pathologists using tumor proportion scores. Correlation coefficients for continuous scores in pairwise comparisons between assays ranged from 0.976 to 0.978. When using a 1% positivity threshold (dichotomous scores), the 22C3 DAKO assay and 22C3 VENTANA assays showed the greatest agreement (93% agreement, κ = 0.86, 95% CI 0.75-0.97), and the 22C3 DAKO and SP263 VENTANA assays tended to show slightly less agreement (84% agreement, κ = 0.66, 95% CI 0.50-0.82). When using a 50% positivity threshold (dichotomous scores), all pairwise comparisons showed similar agreement (96-99% agreement, κ = 0.89-0.97). Overall, there was no significant difference between assays at 1% or 50% thresholds (P = .77). These data are consistent with potential interchangeability of these assays, which may widen the scope of PD-L1 assays available to laboratories and reduce logistical barriers to testing.

Selected highlights of the 2019 Pulmonary Pathology Society Biennial Meeting: PD-L1 test harmonization studies

Immune checkpoint inhibitors (ICI) including programmed death 1 (PD-1) inhibitors, such as nivolumab and pembrolizumab, or programmed death ligand 1 (PD-L1) inhibitors, such as atezolizumab and durvalumab, have recently emerged in advanced stage lung cancer as new standards of care. They are now indicated in first- line and second- or later-line treatment of metastatic or locally-advanced stage III non-small cell lung cancer (NSCLC), as well as for metastatic small cell lung cancer (SCLC), as single agent immunotherapy or in combination with chemotherapy. Four PD-L1 immunohistochemistry (IHC) assays have been established and validated in randomized trials, each for a specific ICI. They use different primary monoclonal antibodies, platforms and detection systems, as well as different scoring systems to assess PD-L1 expression either by tumor cells (TCs) and/or by infiltrating immune cells (ICs). Most studies have shown a close analytical performance of three of these clinically-validated standardized assays, but their use restricted to dedicated platforms, which are not all available in most laboratories, questions their applicability. In addition, the relative high costs of the assays have led to the development of in-house protocols in many pathology laboratories. Their use in clinical practice to assess the predictive value of PD-L1 expression for prescription of ICI raises the issue of their reliability and their validation as compared to standardized assays. This article discusses the main comparative studies available between LDT and assays, with clear evidence that LDT can reach a performance equivalent to the trial-validated assays. The requirements are an adequate validation as compared to an appropriate standard, and the participation to external quality assurance programs and training programs for PD-L1 IHC assessment for pathologists.

Immunization against ROS1 by DNA Electroporation Impairs K-Ras-Driven Lung Adenocarcinomas ​

Non-small cell lung cancer (NSCLC) is still the leading cause of cancer death worldwide. Despite the introduction of tyrosine kinase inhibitors and immunotherapeutic approaches, there is still an urgent need for novel strategies to improve patient survival. ROS1, a tyrosine kinase receptor endowed with oncoantigen features, is activated by chromosomal rearrangement or overexpression in NSCLC and in several tumor histotypes. In this work, we have exploited transgenic mice harboring the activated K-Ras oncogene (K-Ras^G12D) that spontaneously develop metastatic NSCLC as a preclinical model to test the efficacy of ROS1 immune targeting. Indeed, qPCR and immunohistochemical analyses revealed ROS1 overexpression in the autochthonous primary tumors and extrathoracic metastases developed by K-Ras^G12D mice and in a derived transplantable cell line. As proof of concept, we have evaluated the effects of the intramuscular electroporation (electrovaccination) of plasmids coding for mouse- and human-ROS1 on the progression of these NSCLC models. A significant increase in survival was observed in ROS1-electrovaccinated mice challenged with the transplantable cell line. It is worth noting that tumors were completely rejected, and immune memory was achieved, albeit only in a few mice. Most importantly, ROS1 electrovaccination was also found to be effective in slowing the development of autochthonous NSCLC in K-Ras^G12D mice.