AI Learns to Target Tumors for Radiation

April 22, 2019 By IdeaConnection

crowdsourcing tumor identifyingThere are new weapons against lung cancer, the leading cause of cancer death among men and women - crowdsourcing and artificial intelligence.

A competition hosted on TopCoder tasked participants with creating a computer algorithm that could identify tumors and segment them for radiation treatment.

Tumor segmentation is a critical but time-consuming task that requires the expertise of highly trained professionals. It involves a CT scan of the patient’s tumor and surrounding tissues and oncologists determining where to send beams of radiation. They do this by drawing the outlines of tumors and surrounding organs.

Many parts of the world lack sufficient numbers of specialists to perform this kind of work which is where a novel AI approach can come in.  Hence the TopCoder competition which offered $50,000 in prize money.

Winning Algorithms

Ten winning algorithms were selected that can take over some of the medical duties.  The programs allow a computer to look at multiple expert-drawn tumors and thereby learn to draw them. Such an approach could help to fill a staffing shortage gap.

“One of the biggest challenges to targeting tumors is the lack of manpower,” said Dr. Sushil Beriwal, a professor of radiation oncology and deputy director of radiation services at the Hillman Cancer Center at the University of Pittsburgh Medical Center in Pennsylvania.

“The eventual goal is to come out with a product that can be used to help where an expert is not available and to use as a second check (of the radiation oncologist’s own work).”

By the end of the crowdsourcing contest the best algorithms were close to the variation that is seen between physicians.

Details of this crowdsourcing approach to finding novel AI solutions were published in an edition of JAMA Oncology.



Share on      
Next Post »