Machine Learning and Cancer Prognosis
One of the most devastating diseases is Cancer, the problems around this particular disease are vast; from finding a cure to life expectancy of a patient. Often, doctors face the frustrating fact that when a patient asks how long they have, doctors are stumped and not sure of what to say. Throughout the medical history, the prediction of one’s life expectancy rely heavily on guesswork; that is based on a doctor’s experience and knowledge, along with simple statistics that are based on previous patients history.
However, a team of researchers recently has improved the accuracy of the predictions by using artificial intelligence. These new and developed algorithms lead to a much more accurate prediction in cancer prognosis and of course, better care for all patients.
Patient prognosis helps both the doctors and the patients with all the difficult choices they must face like; which course of treatment is the best option? Or whether they should give an aggressive course to a patient who doesn’t have much time? Should a patient want to try out experimental treatments? Unfortunately, there are no easy answers for any of these questions. Doctors simply can do better with more experience and cases. One problem is, not all doctors have the experience needed and not all patients can reach the ones who do. This is where AI comes in to make things better for everyone. An algorithm can improve both statistical data and the doctors’ expertise. By simply adding and storing data from both doctors experience as well as previous medical histories of patients, an algorithm can give better and improved data which doctors can use to make their decisions. This machine learning tools have combined all the data available in medical records which includes: doctors notes, radiology reports, lab test results and much more, with using this information, the algorithm uses variables to predict life expectancy as well as courses of treatments. This amount of work can’t be done by people, can you imagine how long the lists could be? The algorithm on the other hand, uses all the data in the patient chart and cross-reference them with its own stored data to make a prediction in a matter of minutes. These critical information is immediately available for the physician to help make his decision. With more time, more data will be stored and available to be used, therefore giving the algorithms more accuracy every time.