Artificial Intelligence Against Cancer: New Algorithm Reveals Tumor Weaknesses

Cambridge scientists have made a breakthrough in oncology: the PRRDetect algorithm has been created to identify tumors with impaired DNA repair. This discovery could revolutionize cancer treatment.

PRRDetect analyzes patients' complete genomes, allowing high-precision determination of which tumors will respond to specific treatment methods, including immunotherapy, whose effectiveness was previously difficult to predict.

The new algorithm underwent extensive testing on data from the "100,000 Genomes" project, covering 4775 cases of various cancer types. PRRDetect successfully detected unique mutational signatures characteristic of tumors with post-replicative repair defects — a condition where cells lose the ability to effectively repair their DNA after replication.

Among the most important achievements is the identification of 37 characteristic mutational patterns, including 19 completely new ones that were not previously associated with known mechanisms of malignant neoplasm development. Particularly significant is that tumors with such abnormalities often show increased sensitivity to immunotherapy — a treatment method that activates the body's own immune system to fight cancer cells.

This scientific achievement brings personalized medicine closer to clinical practice: specialists will be able to select therapy based on the genetic characteristics of each patient's tumor, significantly increasing the likelihood of successful treatment and minimizing unwanted side effects.