Challenging mitosis detection algorithms: Global labels allow centroid localization

TL;DR

This paper introduces a weakly supervised deep learning approach for mitosis detection that uses only image-level labels, simplifying the process and achieving competitive results compared to more complex, multi-stage methods.

Contribution

The study presents a novel weakly supervised method for mitosis localization that avoids complex multi-stage algorithms and uses only image-level labels, simplifying the process.

Findings

Achieves an F1-score of 0.729 on TUPAC16 dataset

Performs competitively with state-of-the-art methods

Uses only one training phase without pixel-level labels

Abstract

Mitotic activity is a crucial proliferation biomarker for the diagnosis and prognosis of different types of cancers. Nevertheless, mitosis counting is a cumbersome process for pathologists, prone to low reproducibility, due to the large size of augmented biopsy slides, the low density of mitotic cells, and pattern heterogeneity. To improve reproducibility, deep learning methods have been proposed in the last years using convolutional neural networks. However, these methods have been hindered by the process of data labelling, which usually solely consist of the mitosis centroids. Therefore, current literature proposes complex algorithms with multiple stages to refine the labels at pixel level, and to reduce the number of false positives. In this work, we propose to avoid complex scenarios, and we perform the localization task in a weakly supervised manner, using only image-level labels on patches. The results obtained on the publicly available TUPAC16 dataset are competitive with state-of-the-art methods, using only one training phase. Our method achieves an F1-score of 0.729 and challenges the efficiency of previous methods, which required multiple stages and strong mitosis location information.