Niramai screens for breast cancer with body heat and AI — no radiation, no radiologist

Breast cancer is a leading cause of cancer deaths among Indian women, and most cases are detected late: screening is rare in rural areas because mammography machines and radiologists are scarce, the test is costly, and women under 45 get little benefit from it due to denser breast tissue. Dr. Geetha Manjunath, a machine-learning researcher, left her corporate career in 2016 after a family member's diagnosis to build a different way.

The result, Niramai Health Analytix and its Thermalytix platform, uses a high-resolution thermal sensor to capture heat patterns from the chest. Cancer's abnormal vasculature shows up as localised heat, and proprietary AI converts the images into a risk score — contactless, radiation-free, privacy-conscious, and operable by a trained technician rather than a radiologist.

By 2024 Niramai had partnered with more than 150 hospitals and diagnostic centres across India and screened over 100,000 women, including through mobile screening camps run with NGOs and public health departments. A multi-reader clinical study of 258 symptomatic women found Thermalytix reached 95.2% sensitivity and 97.7% negative predictive value, with an AUROC of 0.85 — substantially better than manual interpretation of the same thermal images.

The technology has received clearance in the US and UK and won the World Bank–IFC Global Women Health Award in 2022. Because the test is cheap, portable and technician-run, it is designed for exactly the settings — rural clinics, camps, workplaces — where mammography has never scaled.