|
The Department of Biotechnology (DBT) is planning to create a network of image biobanks and use the potential of artificial intelligence (AI) algorithms to improve medical screenings, precision medicine and assess patient risk factors better.
The Department has called for eligible research institutions, private institutions and non-governmental organisations to set up Image Informatic Centres for the purpose of collection, storage, and retrieval of diagnostic images and associated clinical data.
"This initiative calls for innovative proposals to create Imaging biobanks that focus on curating high-quality, annotated image databases. This dataset serves as the cornerstone for building powerful AI algorithms. Projects that combine diverse clinical expertise to improve image metadata and develop sophisticated AI tools for image analysis will be prioritised," said the DBT. The deadline for submission of Letter of Interest (LoI) is August 31, 2025.
Proposals focused on creating a detailed dataset of images showing healthy and diseased states, annotated comprehensively by clinical experts, will be favoured over those merely collecting images and devising AI algorithms for their analysis.
The Department plans to provide grants-in-aid for up to 10 ambitious and focused, multi-institutional projects of up to five years duration, which has a budgetary requirement per project not exceeding Rs. 15 crore. The financial support will be provided for capital investment on equipment and other accessories and general investment into manpower, consumables, contingencies, travel, among others as per the guidelines. Grants-in-aid support to the private R&D institutions shall be up to Rs 2.5 crore, subject to conditions.
The scope of these centres will include collection and retrieval of a sufficient number of images of the human body, both healthy and pathological, along with associated metadata such as demographics, habits, clinical tests, outcomes, treatments, and genetic information relevant to the target disease, and establishment of secure, accessible databases and IT infrastructure at participating institutes.
The scope would also include making the data available to the research community through a user-friendly online interface, and driving the development of algorithms for disease prediction, diagnosis, and prognosis using the collected data and publishing the results.
The department said that the imaging bio-banks, defined as organised databases of medical images and associated imaging biomarkers of radiology and beyond, linked to other biorepositories for sharing among the multiple researchers. A long-term scope should be the creation of a network of such repositories integrated with the already existing bio-banking network.
Deep learning applications and related AI models, clinical information, and medical imaging investigation can be potential elements for making a positive, enduring impact on human lives in a moderately short measure of time.
Whether it is for a cardiac event, fracture, neurological condition, cancer or thoracic complications, AI can quickly diagnose and provide treatment options. By implementing AI into medical imaging, the technology can enhance medical screenings, improve precision medicine and assess patient risk factors, it added.
"AI-Assisted physicians can make more accurate diagnosis accessible to the bottom of the economic pyramid. In addition, machine learning algorithms, after adequate validations, have the potential for developing a primary level screening system to identify abnormal images/tissues from normal images/tissues or triaging cases which otherwise need the expertise of an experienced radiologist/pathologist," said DBT.
Imaging data also emerges in enabling technologies such as biological samples, electron microscopy, plant leaves, grain quality check and other multi-scale biological systems.
Proposals to develop data banks with direct relevance for medical application will also be considered, it added.
|