A method capable of predicting how breast cancer will evolve 5 years after diagnosis and therefore of providing important information to the oncologist to decide which therapy is the most appropriate for the individual patient. So as to spare her chemotherapy, for example, or to treat her aggressively if the likelihood of relapse is high. All with greater precision than the methods already used today thanks to the development of Artificial Intelligence algorithms. It was developed by the Institute of Informatics and Telematics of the National Research Council (Cnr-Iit) which published its results in Scientific Reports, a journal of the Nature group.
Breast cancer is a leading cause of death in Europe. The annual incidence of new cases in Europe in 2019 was 92.9 women per 100 thousand women; while the annual death rate was 23.1 per 100 thousand. For a patient with breast cancer who has undergone surgical removal of the tumor tissue, it is necessary to decide on a post-operative treatment path that prevents the recurrence of the tumor disease and the formation of metastases. For this purpose, oncologists collect a series of measurements of different parameters (clinical, histological, molecular) and evaluate them with the help of guidelines.
The high social and personal cost of chemotherapy and the evidence of overprescribing with standard methodologies7 have fueled the search for scientific and technological advances in this area, which could impact clinical practice. The need for better prognosis and prediction of therapy outcomes has led to the development of an important line of research on alternative biomarkers based on the molecular profile of breast cancer and on new prediction models and algorithms, which could overcome the inherent limitations of the approaches. previous. In particular, high-throughput sequencing technologies are considered key factors for the success of this new approach, as are efforts for systematic molecular data collection.
Currently, prognostic tools based on molecular biomarkers are considered valid tools to support clinical decisions, complementary to traditional histopathology. Molecular prognostic testing is affordable compared to the cost of chemotherapy for patients who would ultimately not benefit from it. They are considered complementary to more traditional methods based on histology. The tool conceived by the CNR, which consists in the use of a list of marker genes and a computational method to analyze them capable of predicting the survival of a patient 5 years after the removal of the tumor tissue, depending on the chosen therapeutic path, however, it promises to be more efficient. The measurements and analyzes carried out on a database of genetic sequences of biopsy exams of a group of 2000 patients, in fact, thanks to the application of artificial intelligence, indicate a predictive capacity superior to that of the methods currently in use. “Our methodological invention followed two directions,” explains Marco Pellegrini, research director of Cnr-Iit. “On the one hand we drew on the genetic sequencing and biomarkers of excised tissue samples, on the other hand we inserted and analyzed these data in a“ predictor ”an artificial intelligence tool based on a new algorithm. This made it possible to achieve a prediction accuracy of 80% and in some cases 90% ”.
The methodology of the CNR-IIT researchers, which has been the subject of the filing of the patent application in Italy, the United States and the European Community, can provide an important contribution to clinical decisions on breast cancer therapy and the ability to customize the cure with the highest chance of survival.