AI-Based Tool Accelerates Detection of Kidney Cancer

Tumors are not always searched for deliberately, as many scans are performed for unrelated reasons such as trauma or abdominal pain, increasing the risk that early cancers are missed. At the same time, radiology faces a global shortage of specialists while imaging volumes continue to rise, placing growing pressure on already overstretched clinicians. Now, researchers have validated an artificial intelligence (AI) tool that supports radiologists by accelerating kidney cancer detection and improving consistency in image interpretation.

The machine-learning-based system, developed at the University of Tartu Institute of Computer Science (Tartu, Estonia), in collaboration with engineers from Better Medicine (Tartu, Estonia), is designed to assist radiologists in analyzing abdominal CT scans and identifying kidney lesions more efficiently. The tool, called BMVision, analyses CT images and highlights both malignant and benign kidney lesions. It is designed to function as a supportive second reader, helping clinicians interpret complex imaging data without replacing human expertise.

In a retrospective study to evaluate the effectiveness of the AI tool, six radiologists reviewed 200 CT scans, both with and without AI assistance, resulting in 2,400 individual image readings. The assessments compared diagnostic sensitivity, tumor measurement accuracy, reporting speed, and agreement between radiologists. The findings, published in Nature Communications Medicine, showed that AI assistance reduced the time required to identify, measure, and report malignant lesions by approximately one third, while maintaining diagnostic accuracy.

“This study adds to the growing body of evidence that modern AI tools developed in research labs can make a real impact in clinical practice and support doctors in their daily work,” said Associate Professor Dmytro Fishman. “We are very encouraged by these results, which show that AI research in medicine is not only meaningful, but it can truly be used for good.”

Related Links:
University of Tartu Institute of Computer Science
Better Medicine