Indian Scientists Develop AI Tool ‘OncoMark’ to Decode Cancer

In a major breakthrough in cancer research, Indian scientists have unveiled ‘OncoMark’, an AI-powered diagnostic model that analyses cancer at the molecular level, potentially revolutionising how the disease is classified and treated. Developed by the S N Bose National Centre for Basic Sciences in collaboration with Ashoka University, this tool goes far beyond traditional cancer staging by identifying biological hallmarks that drive tumour growth, aggressiveness, and resistance to therapy.

This marks a significant advancement in India’s contribution to precision medicine, where treatments are tailored based on a tumour’s molecular behaviour rather than broad-stage categorisations.

What Is OncoMark and How It Works

OncoMark is a machine learning framework trained to detect hallmark-based biological programs — the underlying processes that regulate how cancers behave and evolve. These include,

• Metastasis (spread of cancer)
• Immune system evasion
• Genomic instability
• Uncontrolled proliferation

While conventional methods like TNM staging rely on tumour size, lymph node involvement, and metastasis, OncoMark provides granular insights into tumour behaviour at a cellular and molecular level — allowing for individualised treatment strategies.

Key Achievements of OncoMark

• Trained on 3.1 million single cells from 14 different cancer types
• Uses synthetic pseudo-biopsies to model tumour progression dynamically
• Achieved over 99% accuracy during internal testing
• Maintained 96%+ accuracy across independent patient datasets
• Validated on 20,000 patient samples from eight major cancer datasets
• Can map hallmark activity visually as cancer progresses

These features make OncoMark a powerful tool for early detection, aggressiveness prediction, and drug response analysis.