The reproducibility of PD-L1 scoring in lung cancer: can the pathologists do better?

Troncone, Giancarlo; Gridelli, Cesare

doi:10.21037/tlcr.2017.10.05

Cited by 26 publications

(15 citation statements)

References 10 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Since the propose of classifying tumor microenvironment base on TIL and PD-L19, quite a few study set out to unravel the clinical significance of such classification10, 11, 25, 26. But the yielded results were inconsistent with one another, which may owe to the modest sample size or the diverse detecting methods as well as various cutoff values adopted in the PD-L1 detection27, 28. At the same time, such classification fails to take PD-1 expression status into consideration, which plays an equal important role in immune evasion.…”

Section: Discussionmentioning

confidence: 99%

Deciphering Microenvironment of NSCLC based on CD8+ TIL Density and PD-1/PD-L1 Expression

Lin¹,

Gu²,

Cui³

et al. 2019

J. Cancer

View full text Add to dashboard Cite

Purpose: To determine whether distinct tissue immune microenvironments differentially impact on clinical outcome in non-small cell lung cancer (NSCLC), an extended analysis of PD-1/PD-L1 and Tumor Infiltrating Lymphocytes (TILs) was performed.Materials and Methods: 1016 NSCLC mRNA-sequence samples from The Genome Data Analysis Center (TCGA) and 275 NSCLC mRNA-microarray samples from Gene Expression Omnibus (GEO) were included as testing cohort and validation cohort respectively. Enrichment scores of CD8+ T cells' metagene were used for quantifying its infiltrating density. Based on the median values of CD8+ T cell density and PD-1/PD-L1 mRNA expression, the samples were classified into four Tumor Immune Microenvironment types (TIMTs). Overall survival, as well as clinicopathological features, mutational profiles, mismatch repair score etc. were compared across the four types.Results: Neither PD-1 expression nor PD-L1 expression was associated with outcome in the overall NSCLC. Classification of TIMT based on PD-1/PD-L1 and CD8+ TIL could efficiently classify patients of different survival in ADC but not SCC, with the best overall survival achieved in TIMT3 (high CD8+ TIL and low PD-1/PD-L1), whereas TIMT2 (low CD8+ TIL and high PD-1/PD-L1) manifested the worst outcome. TIMT classification based on PD-1/ CD8+ TIL could better stratify patient of different prognosis than PD-L1/ CD8+ TIL based classification. EGFR wide type and IFNγ overexpression were associated with TIMT4 (high PD-1/PD-L1 and high CD8+ TIL), whereas tumor mutational burden (TMB) manifested no significant difference across four TIMTs.Conclusion: The classification of tumors into four microenvironment subtypes based on PD-1/PD-L1 status and CD8+ TIL is an appropriate approach to stratify patients of different clinical outcome and better guide the practical use of immunotherapy.

show abstract

Section: Discussionmentioning

confidence: 99%

Deciphering Microenvironment of NSCLC based on CD8+ TIL Density and PD-1/PD-L1 Expression

Lin¹,

Gu²,

Cui³

et al. 2019

J. Cancer

View full text Add to dashboard Cite

show abstract

“…Pathologists have difficulties in interpreting an accurate PD-L1 status on cancer sections which is a time-consuming clinical practice procedure. Additionally, inter-observer variation among pathologists may contribute to inaccurate patient stratification and limit the applications of PD-L1 IHC staining as an effective biomarker in clinical procedure [10][11][12][13] .…”

Section: Introductionmentioning

confidence: 99%

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

Liu

et al. 2020

Preprint

View full text Add to dashboard Cite

BackgroundThis study proposed a computational method to detect the cancer areas and calculate the tumor proportion score (TPS) of PD-L1 immunohistochemistry (IHC) expression for lung adenocarcinoma based on deep learning and transfer learning. Patients and methodsPD-L1 22C3 and SP142 IHC slides of lung adenocarcinoma samples on digitized whole-slide images (WSI) database were employed. We build a deep convolutional neural network (DCNN) to automatically segment cancer regions. TPS was calculated based on segmented areas and then compared with the interpretations of pathologists. ResultsWe trained a DCNN model based on 22C3 dataset and fine-tuned it with SP142 dataset. We obtain a robust performance on cancer region detection on both datasets, with a sensitivity of 93.36% (22C3) and 92.80% (SP142) and a specificity of 93.97% (22C3) and 89.25% (SP142). With all the coefficient of determinations larger than 0.9 and Fleiss' and Cohen's Kappa larger than 0.8 (between mean or median of pathologists and TPS calculated by our method), we also found out the strong correlation between the TPS estimated by our computational method and estimation from multiple pathologists' interpretations of 22C3 and SP142 respectively. ConclusionWe provide an AI method to efficiently predict cancer region and calculate TPS in PD-L1 IHC slide of lung adenocarcinoma on two different antibodies. It demonstrates the potential of using deep learning methods to conveniently access PD-L1 IHC status. In the future, we will further validate the AI tool for automated scoring PD-L1 in large volume samples.3

show abstract

“…To complicate matters, PD-L1 positivity thresholds vary [ 80 ]. Inter-observer variability of PD-L1 assessment by pathologists is a known problem [ 17 , 77 ]. This may contribute to inaccurate patient stratification and the misinterpretation of the impact of PD-L1 expression on clinical outcome: Subjective decisions can lead to radically different therapeutic stratification when scoring around cutoffs.…”

Section: Introductionmentioning

confidence: 99%

Precision immunoprofiling by image analysis and artificial intelligence

et al. 2018

View full text Add to dashboard Cite

Clinical success of immunotherapy is driving the need for new prognostic and predictive assays to inform patient selection and stratification. This requirement can be met by a combination of computational pathology and artificial intelligence. Here, we critically assess computational approaches supporting the development of a standardized methodology in the assessment of immune-oncology biomarkers, such as PD-L1 and immune cell infiltrates. We examine immunoprofiling through spatial analysis of tumor-immune cell interactions and multiplexing technologies as a predictor of patient response to cancer treatment. Further, we discuss how integrated bioinformatics can enable the amalgamation of complex morphological phenotypes with the multiomics datasets that drive precision medicine. We provide an outline to machine learning (ML) and artificial intelligence tools and illustrate fields of application in immune-oncology, such as pattern-recognition in large and complex datasets and deep learning approaches for survival analysis. Synergies of surgical pathology and computational analyses are expected to improve patient stratification in immuno-oncology. We propose that future clinical demands will be best met by (1) dedicated research at the interface of pathology and bioinformatics, supported by professional societies, and (2) the integration of data sciences and digital image analysis in the professional education of pathologists.

show abstract

The reproducibility of PD-L1 scoring in lung cancer: can the pathologists do better?

Cited by 26 publications

References 10 publications

Deciphering Microenvironment of NSCLC based on CD8+ TIL Density and PD-1/PD-L1 Expression

Deciphering Microenvironment of NSCLC based on CD8+ TIL Density and PD-1/PD-L1 Expression

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

Precision immunoprofiling by image analysis and artificial intelligence

Contact Info

Product

Resources

About