Assessing the Performance of Deep Learning for Automated Gleason Grading in Prostate Cancer

TL;DR

This study evaluates 11 deep neural network architectures for automated Gleason grading in prostate cancer, highlighting ConvNeXt as the top performer and demonstrating the potential of AI to improve diagnostic accuracy.

Contribution

It compares traditional and recent deep learning architectures for Gleason grading, providing a standardized evaluation framework and identifying ConvNeXt as a promising model.

Findings

ConvNeXt achieved the highest sensitivity among tested architectures.

Newer architectures outperformed traditional models in grading accuracy.

Challenges remain in differentiating closely related Gleason grades.

Abstract

Prostate cancer is a dominant health concern calling for advanced diagnostic tools. Utilizing digital pathology and artificial intelligence, this study explores the potential of 11 deep neural network architectures for automated Gleason grading in prostate carcinoma focusing on comparing traditional and recent architectures. A standardized image classification pipeline, based on the AUCMEDI framework, facilitated robust evaluation using an in-house dataset consisting of 34,264 annotated tissue tiles. The results indicated varying sensitivity across architectures, with ConvNeXt demonstrating the strongest performance. Notably, newer architectures achieved superior performance, even though with challenges in differentiating closely related Gleason grades. The ConvNeXt model was capable of learning a balance between complexity and generalizability. Overall, this study lays the groundwork for enhanced Gleason grading systems, potentially improving diagnostic efficiency for prostate cancer.