Artificial intelligence predicts efficacy of breast cancer treatment

Using a new artificial intelligence (AI) algorithm, engineers at the University of Waterloo in Canada have developed a method that could help predict the likelihood that women with breast cancer will respond well to chemotherapy given before surgery. aim Eliminates potential biases and high uncertainty in clinical judgment by avoiding unnecessary side effects in inappropriate treatments.

The application of the new technique is mainly aimed at the so-called. “neoadjuvant chemotherapy treatment” where doctors administer therapeutic agents before the main treatment. This treatment may result in a pathological complete response (pCR), in which previously inoperable tumors shrink, but there is no way to predict the patient’s course or the pathological response to chemotherapy.

How can AI predict the effectiveness of chemotherapy?

Developed as part of the open source Cancer-Net initiative, the new algorithm was developed in a project led by Amy Tai, a graduate student at the Image and Image Processing Lab. The researcher “trained” the AI ​​software with images of breast cancer obtained with a new magnetic resonance imaging method.

The technique, called synthetic associated diffusion imaging (CDI), was developed by Dr. Alexander Wong, who also guided the algorithm project. Artificial intelligence fed by data from these CDI images taken in old breast cancer cases and diagnoses, Make a reliable prediction of whether preoperative chemotherapy treatment will benefit new patients based on their current imaging..

baptized As Cancer-Net BCa, the English initials for the prediction of the complete pathological response for breast cancer, the new technology may be extended to other ICD applications in different types of cancer in the future. AI tends to be an important tool for oncologists to use, reducing anxiety and uncertainty.

ARTICLE – ArXiv – DOI: 10.48550/arXiv.2211.05308 (not peer reviewed).