Cytology is required for cancer screening and cancer diagnosis, yet most artificial intelligence systems assist cytology at individual tasks and specific specimen types. We developed CROWN, a universal foundation model for cytology that has the potential to support a broad spectrum of diagnostic applications across organs, diseases and clinical settings.

Fewer than 1% of the spatial transcriptomics papers we analysed had a first author in a low- or lower-middle-income economy.

Dr. Amit Sethi, Indian Institute of Technology Bombay, presents Toward Trustworthy AI for Computational Pathology: From Robust Learning to Clinically Relevant Biomarkers.

Colorectal cancer histopathological grading relies on the accurate segmentation of glandular structures. Current deep learning–based methods depend heavily on large-scale pixel-level annotations that are labor-intensive and not amenable to clinical practice. Weakly supervised semantic segmentation offers a promising alternative; yet, existing class activation map–based weakly supervised semantic segmentation approaches often produce incomplete, low-quality pseudo-masks that overemphasize discriminative regions and fail to provide reliable supervision for unannotated glandular structures, limiting their suitability for dense histopathology segmentation under sparse supervision. We propose a novel weakly supervised teacher–student framework that leverages sparse pathologists’ annotations and an Exponential Moving Average–stabilized teacher network to generate refined pseudo-masks.

Training foundational medical imaging models using multi-task learning. - FraunhoferMEVIS/MedicalMultitaskModeling