Weakly supervised multiple instance learning histopathological tumor segmentation
This addresses the challenge of clinical translation in medical imaging by reducing reliance on costly annotations, though it appears incremental as it builds on existing weakly supervised methods.
The paper tackles the problem of histopathological tumor segmentation by proposing a weakly supervised framework that uses standard clinical annotations instead of hand-crafted ones, achieving promising results compared to expert annotations on public datasets like The Cancer Genome Atlas and PatchCamelyon.
Histopathological image segmentation is a challenging and important topic in medical imaging with tremendous potential impact in clinical practice. State of the art methods rely on hand-crafted annotations which hinder clinical translation since histology suffers from significant variations between cancer phenotypes. In this paper, we propose a weakly supervised framework for whole slide imaging segmentation that relies on standard clinical annotations, available in most medical systems. In particular, we exploit a multiple instance learning scheme for training models. The proposed framework has been evaluated on multi-locations and multi-centric public data from The Cancer Genome Atlas and the PatchCamelyon dataset. Promising results when compared with experts' annotations demonstrate the potentials of the presented approach. The complete framework, including $6481$ generated tumor maps and data processing, is available at https://github.com/marvinler/tcga_segmentation.