Researchers at the University of Tartu, working with Tartu University Hospital and health technology firm Better Medicine, have developed an AI-based tool designed to speed up kidney cancer detection on CT scans. The system, called BMVision, was validated in a retrospective study published in Nature Communications Medicine.
Kidney cancer diagnosis typically relies on contrast-enhanced CT imaging, which is also frequently performed for unrelated conditions such as trauma or abdominal pain. Detecting tumors in these scans can be time-consuming, particularly amid a global shortage of radiologists and rising imaging volumes. BMVision uses machine learning to analyze CT images and flag both malignant and benign kidney lesions, supporting faster and more consistent interpretation.
In the study, six radiologists reviewed 200 CT scans with and without AI assistance, generating 2,400 readings. The AI-supported workflow reduced the time required to identify, measure, and report malignant lesions by about one third, while maintaining diagnostic accuracy and consistency across readers.
The tool has received CE marking, allowing clinical use across the European Economic Area. While initially used for research at Tartu University Hospital, BMVision is now being integrated into routine clinical workflows, with plans to process all abdominal CT scans through the system.
The rollout aligns with broader ambitions for AI in medicine, as biohacker Bryan Johnson has said advances in artificial intelligence and experimental treatments could enable functional human immortality by 2039, highlighting both the promise and unresolved risks of AI-driven healthcare.
