Artificial Intelligence versus cancer: new algorithm reveals tumor weaknesses

Researchers from the University of Cambridge have made an important step in the fight against oncological diseases. Thanks to their efforts, PRRDetect has emerged - an algorithm capable of detecting tumors with impaired DNA repair ability. This discovery could radically change the approach to cancer treatment.

PRRDetect works by analyzing patients' complete genomes. Thanks to this tool, it becomes possible to determine with high accuracy which tumors respond to certain types of therapy, including immunotherapy, which previously was not always applied effectively.

The algorithm was tested on a data array from the "100,000 Genomes" project, covering 4775 cases of various types of cancer. PRRDetect successfully identified unique mutational signatures inherent to tumors with post-replicative repair defects - a condition in which cells lose the ability to properly "repair" their DNA after copying.

One of the key results - the algorithm identified 37 characteristic mutational patterns, including 19 new ones that were not previously associated with known pathways of malignant neoplasm development. Particularly important is that tumors with these abnormalities often demonstrate increased susceptibility to immunotherapy - a type of treatment that stimulates the immune system to recognize and destroy cancer cells.

This achievement brings precision medicine closer to everyday practice: doctors will be able to select treatment based on the genetic profile of the tumor, increasing the chances of success and reducing the risk of side effects.