VA3D: Presurgical AI for Tumour Sizing in 3D Breast Scans

TP16

Home > Our Research > Algorithms & AI/ML Models > VA3D: Presurgical AI for Tumour Sizing in 3D Breast Scans
Dr. Saritha Unnikrishnan

Dr. Saritha Unnikrishnan

Lead Translational Researcher

Thematic area:

Algorithms and AI/ML Models

Product:

A volumetric attention neural network that accurately quantifies tumour size from 3D breast imaging (e.g., MRI). The model includes explainability modules to support radiologist trust and integrates into clinical imaging workflows.

Market:

Targets breast imaging markets globally, where precise tumour sizing affects millions of surgeries annually. Potential for ~$280M per year in incremental revenue when deployed as an add-on to breast MRI workflows. Customers include radiology departments, imaging equipment providers, and AI diagnostics companies.

Fit:

Addresses a well-documented gap: radiologists’ tumour size estimates have 35–64% sensitivity. VA3D cuts reexcision rates, improves surgical planning, and supports value-based breast cancer care. Strong synergy with existing research teams and clear commercial pathways through spinout or licensing.

Other Projects