Cancer Treatment Can Become Less Toxic as AI Uses ‘Self-Learning’


The men and women of science from MIT are looking to incorporate state of the art machine-learning techniques in order to improve the lives of people diagnosed with glioblastoma, a terribly aggressive form of brain cancer. Currently, the treatment is performed through chemotherapy and radiotherapy dosing which can become extremely toxic.

What is glioblastoma?

This form of cancer represents a malignant tumor that grows in the spinal cord or in the brain and it comes with a terrible prognosis, five years left for adults. In order to prolong their lives and maintain a certain quality of life, patients have to endure radiation therapy combined with multiple drugs on a monthly basis.

The drugs are administered in high doses so that the tumor will shrink as much as possible. However, these drugs are incredibly potent and can debilitate patients.

What can be done to prevent this?

The researchers from the MIT Media Lab made a model which can decrease the treatment’s toxicity while still ensuring efficiency. This model uses a machine-learning technique which allows it to ‘self-learn’ and by looking at what treatments are currently being used it will try to adjust the doses. In the end it will come up with the ideal treatment plan which will continue to reduce the tumor’s size and lower the drug’s potency at the same time.

Has it been tested?

There were tests conducted on a group of 50 patients. The machine was able to design treatment cycles which were lower in debilitating risk (from 25% to 50% safer to use) while maintaining a tumor-shrinking potential similar to traditional treatments.

Pratik Shah, one of the investigators from Media Lab stated that the goal was to help the patients with reducing their tumors and also, they looked to increase their quality of life by lowering the toxicity.


Recommended For You

Leave a Reply

Your email address will not be published. Required fields are marked *