New Delhi, 26 Nov: Scientists in India have developed an artificial intelligence (AI) framework that could transform the understanding and treatment of cancer by analyzing its molecular “personality,” rather than relying solely on traditional tumor staging systems.
The study, conducted by researchers from the S N Bose National Centre for Basic Sciences, an autonomous institute under the Department of Science and Technology (DST), in collaboration with Ashoka University, introduces OncoMark, the first AI system capable of decoding the complex molecular signals driving cancer progression.
Unlike conventional staging systems, such as TNM, which focus on tumor size and spread, OncoMark examines the hallmarks of cancer—the hidden biological programs that control tumor growth, metastasis, immune evasion, and therapy resistance. This approach helps explain why patients with seemingly similar cancer stages can experience very different outcomes.
The framework analyzed 3.1 million single cells across 14 cancer types, creating synthetic “pseudo-biopsies” to model hallmark-driven tumor states. OncoMark demonstrated over 99% accuracy in internal testing and maintained above 96% accuracy across five independent cohorts. Validation on 20,000 real-world patient samples confirmed its broad applicability.
“This AI model allows us, for the first time, to visualize how hallmark activity evolves with cancer progression,” said Dr. Shubhasis Haldar, one of the lead researchers. Dr. Debayan Gupta, co-lead, added that OncoMark could guide doctors toward therapies targeting the most active molecular processes in a patient’s tumor, enabling more precise and early interventions for aggressive cancers.
The research has been published in Communications Biology (Nature Publishing Group).
Source: Ministry of Science & Technology, Study published in Communications Biology (Nature Publishing Group).
