Written by : Jayati Dubey
January 24, 2025
The project digitizes biopsy slides to prevent damage, enhance clinical decisions, and speed up diagnoses while enabling AI-driven pathology research.
IIIT Hyderabad (IIITH), in partnership with Nizam’s Institute of Medical Sciences (NIMS), Hyderabad, has launched publicly available datasets featuring digitized histopathological images of brain cancer and lupus nephritis (a kidney disease).
This initiative is part of the India Pathology Dataset (IPD) project, which aims to revolutionize pathology in India by bringing together academia, hospitals, industry, and government.
The project focuses on digitizing tissue biopsy slides to safeguard them from physical damage, improve clinical decision-making, and accelerate diagnosis turnaround times.
Additionally, the datasets aim to create opportunities for artificial intelligence (AI)-driven research in pathology.
Supported by iHub-Data, IIITH installed a whole-slide digital scanner at the NIMS campus as part of the IPD initiative.
Prof Vinod P.K., who leads the project, explained, “Traditionally, tissue samples are examined under a microscope.
By digitizing these slides, they can be visualized on computers and shared across locations for collaborative diagnosis with other pathologists.”
One of the key datasets, the IPD-Brain dataset, has been published in Nature Scientific Data.
This open-access resource comprises 547 high-resolution H&E slides from 367 patients, representing one of Asia’s largest datasets focused on Indian demographics.
Dr Megha Uppin from NIMS highlighted the importance of the IPD-Brain dataset in brain tumor research.
“Effective cancer management relies on precise typing, sub-typing, and grading. This dataset is a critical first step in training machine learning models to explore regional and ethnic disease variations, enhancing diagnostic precision,” she said.
With brain tumor diagnosis increasingly dependent on molecular genetics, AI can play a pivotal role in bridging gaps in cost-effective and accurate diagnosis by predicting molecular abnormalities.
Furthermore, AI can address the shortage of specialized neuropathologists, enabling collaborative diagnoses between peripheral institutes and experts via digital pathology.
Efforts are underway to expand the datasets to include other cancers, such as breast, lung, colorectal, oral, and cervical cancers. NIMS is already contributing to the lung cancer dataset.
The project also addresses kidney diseases, with a dataset on lupus nephritis, a condition caused by the immune system attacking the kidneys.
According to Prof Vinod, lupus nephritis disproportionately affects women in India, with a high incidence in Telangana.
The dataset is designed to assist nephropathologists in classifying diseases and guiding effective treatment plans.
Dr Uppin noted that AI could mitigate interobserver variations in subtyping lupus nephritis, improving diagnostic consistency and outcomes.
The team is exploring AI’s potential to predict molecular markers from H&E slide images—capabilities that traditional histopathology cannot achieve.
Prof Vinod explained, “Pathologists can’t see underlying molecular changes at the DNA level reflected in tissue morphology.”
For instance, the team has developed AI models to predict IDH mutations, which are crucial for diagnosing and prognosing brain tumors.
The open-source nature of these datasets offers valuable tools for researchers, educators, and healthcare professionals.
“This is one of the first few instances of open-source medical data from India for ‘human good’,” said Prof Vinod.
The datasets are also instrumental in helping pathology students gain deeper insights into histopathological images.
To support broader access, a second whole-slide scanner has been installed at the IIITH campus, making the technology available to dental colleges and corporate hospitals.
Prof Vinod also emphasized that the IPD project is tailored to India’s unique healthcare challenges, offering an alternative to datasets based on foreign populations.
As the project expands to include datasets on breast cancer and other conditions, it aims to advance teaching, learning, and research in histopathology, fostering healthcare innovation and improving patient outcomes in India.
Stay tuned for more such updates on Digital Health News.