IIIT Hyderabad, in collaboration with Nizam's Institute of Medical Sciences (NIMS), has announced the release of digitised histopathological datasets as part of the India Pathology Dataset (IPD) project. This collaborative initiative brings together academic institutions, hospitals, industries, and the government to digitise tissue biopsy slides, a step that promises to revolutionise the way medical conditions such as brain cancer and kidney disease (lupus nephritis) are diagnosed and managed.
The primary goal of this project is to digitise tissue biopsy slides to reduce the risk of damage to physical slides, enable AI-driven research, and improve clinical decision-making. The digitisation of these slides allows for remote collaboration among pathologists, facilitating faster and more accurate diagnoses. Prof Vinod P K, who leads the dataset curation for various cancers, emphasized the advantages of using computers for visualising slides, which allows for collaborative diagnoses, even across different locations.
One of the first datasets, the IPD-Brain dataset, focuses on brain cancer and has been published in Nature Scientific Data. This open-access dataset contains 547 high-resolution slides from 367 patients, making it one of Asia's largest collections of its kind. The dataset is particularly valuable for exploring regional and ethnic variations in diseases and is set to aid in training machine learning models to improve diagnostic accuracy.
The project is expanding beyond brain tumours, with ongoing efforts to build datasets for other cancers, including breast, lung, colorectal, oral, and cervical cancers. A dedicated dataset for lupus nephritis, a kidney disease, has also been compiled, given its significant incidence in India, particularly in Telangana.
The open-source nature of these datasets makes them a valuable resource for researchers and educators, including pathology students who can leverage these resources for enhanced learning.
