Hitchhiker's guide to cancer-associated lymphoid aggregates in histology images: manual and deep learning-based quantification approaches

TL;DR

This paper reviews methods for quantifying lymphoid aggregates in cancer histology images, introducing a deep learning algorithm called HookNet-TLS for automated detection and providing guidelines for its use.

Contribution

It presents a new deep learning-based algorithm for automated quantification of lymphoid aggregates and offers practical guidelines for training and implementation.

Findings

HookNet-TLS effectively detects lymphoid aggregates across tissue types

Automated quantification reduces subjectivity and variability

Guidelines improve reproducibility of lymphoid aggregate analysis

Abstract

Quantification of lymphoid aggregates including tertiary lymphoid structures with germinal centers in histology images of cancer is a promising approach for developing prognostic and predictive tissue biomarkers. In this article, we provide recommendations for identifying lymphoid aggregates in tissue sections from routine pathology workflows such as hematoxylin and eosin staining. To overcome the intrinsic variability associated with manual image analysis (such as subjective decision making, attention span), we recently developed a deep learning-based algorithm called HookNet-TLS to detect lymphoid aggregates and germinal centers in various tissues. Here, we additionally provide a guideline for using manually annotated images for training and implementing HookNet-TLS for automated and objective quantification of lymphoid aggregates in various cancer types.