Scientists at Northwestern University in Chicago have developed a tool to predict cancer patients’ chances of survival using artificial intelligence. The study showed that survival in breast, thyroid and pancreatic cancer is affected not only by the stage of the disease but also by many other factors. The work is available at: American College of Surgeons Clinical Congress (ACS).
The team used data from 259,485 breast cancer patients, 76,624 thyroid cancer patients and 84,514 pancreatic cancer patients. They found that survival was affected by the patient’s age, tumor size and the type of treatment given. The factors differ for different tumor types. For example, in case of thyroid cancer, it is important whether the patient has lymph node involvement or not, and in case of breast cancer, whether there is sensitivity to certain hormones.
Three quarters of the collected data were used to train machine learning algorithms. Using the remaining data, the researchers tested the calculator’s performance. It accurately assessed whether the patient would survive the next nine to ten months.
Currently, survival estimates for cancer patients depend primarily on the stage of the cancer. The new calculator includes certain biomarkers and treatment features that affect a patient’s expected prognosis and are not available in previous survival calculators.
happened before knownIt was revealed that living alone increases the risk of death from cancer by 32 percent.