Domain Adaptation-Based Deep Learning for Automated Tumor Cell (TC) Scoring and Survival Analysis on PD-L1 Stained Tissue Images

Abstract

We report the ability of two deep learningbased decision systems to stratify non-small cell lung cancer (NSCLC) patients treated with checkpoint inhibitor therapy into two distinct survival groups. Both systems analyze functional and morphological properties of epithelial regions in digital histopathology whole slide images stained with the SP263 PD-L1 antibody. The first system learns to replicate the pathologist assessment of the Tumor Cell (TC) score with a cut-point for positivity at 25% for patient stratification. The second system is free from assumptions related to TC scoring and directly learns patient stratification from the overall survival time and event information. Both systems are built on a novel unpaired domain adaptation deep learning solution for epithelial region segmentation. This approach significantly reduces the need for large pixel-precise manually annotated datasets while superseding serial sectioning or re-staining of slides to obtain ground truth by cytokeratin staining. The capacity of the first system to replicate the TC scoring by pathologists is evaluated on 703 unseen cases, with an addition of 97 cases from an independent cohort. Our results show Lin's concordance values of 0.93 and 0.96 against pathologist scoring, respectively. The ability of the first and second system to stratify anti-PD-L1 treated patients is evaluated on 151 clinical samples. Both systems show similar stratification powers (first system: HR =0.539, p=0.004 and second system: HR=0.525, p=0.003) compared to TC scoring by pathologists (HR=0.574, p=0.01).

Automated PD-L1 Scoring Using Artificial Intelligence in Head and Neck Squamous Cell Carcinoma

“…A completely different approach attempts to directly map PD-L1 scores using Deep Learning [16,35], which might allow it to be less vulnerable to artifacts. Using an auxiliary classifier generative adversarial network for tumor detection, the TPS could be calculated based on the ratio of the pixel numbers of positive and negative tumor areas.…”

“…Using an auxiliary classifier generative adversarial network for tumor detection, the TPS could be calculated based on the ratio of the pixel numbers of positive and negative tumor areas. Generative adversarial networks basically consist of two neural networks (discriminator and generator), which compete with each other and allow the model to learn [16,35]. Although this deep learning approach was able to determine the TPS in entire WSIs, the interpretability of the underlying decision criteria was not provided.…”

“…The effectiveness of artificial intelligence (AI) in medical image analysis (IA) of histopathological images has been widely demonstrated [25], suggesting that automated determination bases on AI/IA could enable reliable and reproducible PD-L1 scoring [14]. Attempts have already been made to determine PD-L1 scores using AI/IA [16,[26][27][28][29][30][31][32][33][34][35][36]. However, an open-source and automated determination of all three PD-L1 scores simultaneously in whole slides images (WSI) has not yet been described in the available literature.…”

Machine learning in computational histopathology: Challenges and opportunities

“…147 For kidney cancers, segmenting glomeruli 126,148 or different subtypes of kidney tissues 127,149 are key tasks of interest. Researchers have also made progress on segmenting normal and abnormal parts from histopathological images for breast, 90,99,150 lung, 110,151 bladder, 134 stomach, 123 prostate 90 cancer.…”

“…These are then clustered to automate the identification of subgroups in lung, 159 brain, 159 breast, 98,181 colon cancer 113 WSIs. In addition, research has begun to use ML to learn associations between gene expression and pathological images, 110,130,182 perform stain normalization for histopathological images, 106,[183][184][185][186] generate synthetic data, 106,130,148,151,170 compress images 187 and automate histopathological captioning and diagnosis generation. 135…”