Software that uses AI to detect or diagnose cancer by analysing X-rays, CT scans, and other medical images will now be regulated as Class C medical devices, according to a notification issued by the Central Drugs Standard Control Organisation (CDSCO). This category is reserved for moderate-to-high risk products.
The decision will require AI-based cancer detection and diagnostic tools — already in use at some hospitals and diagnostic centres — to obtain regulatory approval, undergo safety validation and be subject to continuous monitoring before they can be widely deployed in clinical settings. Developers will also be required to comply with prescribed quality standards and report any patient safety–related issues.
Dr Pallavi Shukla, Associate Professor of Preventive Oncology at AIIMS New Delhi welcomed the move and said a defined regulatory framework is critical to ensure the ethical use of AI in healthcare and safeguard patients. While AI holds promise for cancer screening — especially for cervical, lung and breast cancers — she said most such tools in India are still at the research stage due to the lack of large, representative Indian datasets and are not yet suitable for routine clinical use.
She added that the regulatory step is timely, as it introduces oversight in a rapidly evolving area before large-scale adoption.
Health experts described the move as significant, noting that AI tools are increasingly influencing decisions related to cancer screening and diagnosis. Officials said it could also pave the way for other AI-based medical software and devices to be brought under formal regulatory oversight as their role in clinical care expands.
India has witnessed a surge in startups offering AI-driven cancer screening solutions, often making early-detection claims. Regulators said the new framework is intended to ensure such claims are scientifically validated and do not compromise patient safety. The CDSCO added that the risk classification list will be reviewed periodically as technology evolves.
(With inputs from TOI)