Bengaluru researcher creates AI model for early cervical cancer risk detection

Researcher Lalasa Mukku of CHRIST (Deemed to be University) in Bengaluru has developed a suite of artificial‑intelligence models to predict which women are at high risk of progressing to cervical cancer years before a tumor forms. The flagship model, named CMT‑CNN, combines sequential colposcopy images with clinical data and achieved a classification accuracy of 92.3% for detecting Cervical Intraepithelial Neoplasia (CIN). A later quantum‑convolutional architecture reported about 98.6% accuracy on public screening datasets.

Cervical cancer remains a leading malignancy among women worldwide, with roughly 660,000 new cases and 350,000 deaths in 2022, and over 120,000 new cases and 80,000 deaths annually in India. Early identification of precancerous lesions can prevent most invasive cancers, but access to specialist screening is limited in many low‑ and middle‑income settings. Mukku’s AI approach aims to support clinicians by flagging high‑risk patients for closer monitoring and preventive treatment, potentially strengthening national screening programmes.

The models are still at the research stage and will require extensive clinical validation before deployment in hospitals. If validated, the technology could complement existing screening methods, improve diagnostic accuracy, and contribute to India's goal of reducing the cervical cancer burden through earlier detection.