Predicting the Cell-level PD-L1 Expression Status from H&E-stained Histopathological Slide

Area of Science:

• Computational pathology
• Biomedical imaging analysis
• Cancer diagnostics

Background:

• Tumor Proportion Score (TPS) is vital for PD-L1 immunohistochemistry (IHC) treatment eligibility.
• Current IHC slide assessment is time-consuming and expensive for pathologists.

Purpose of the Study:

• To develop a cost-effective method for predicting PD-L1 expression using Hematoxylin and Eosin (H&E) stained slides.
• To reduce diagnostic time and costs associated with PD-L1 IHC staining.

Main Methods:

• Extracting features from PD-L1 stained slides.
• Mapping PD-L1 positive cell data onto H&E slides to create pseudo-labels.
• Developing a deep correlation model linking PD-L1 expression to H&E morphological features.

Main Results:

• The model successfully predicts PD-L1 markers directly from H&E slides.
• Achieved high accuracy in cell-level PD-L1 detection.
• Demonstrated the feasibility of assessing PD-L1 status using H&E staining.

Conclusions:

• The proposed method offers a promising, cost-effective alternative for PD-L1 assessment in cancer.
• Simplifies PD-L1 detection in clinical practice by leveraging H&E slides.
• Potential to significantly aid pathologists in cancer diagnostics and treatment planning.