IBM’s Watson for Oncology, an artificial intelligence-based diagnostic tool, is making treatment recommendations more congruent to those made by doctors, a local hospital said Tuesday.

Watson for Oncology’s top recommendations achieved a 55.9 percent concordance rate with those made by physicians for 118 colon cancer patients, Gachon University Gil Medical Center said at a press conference.

The concordance rate signified a 7 percent increase from a retrospective study on 656 patients from 2009 to 2016, the hospital said.

“The increased concordance rate in the ‘strongly recommended’ category signifies an equal rise in conformity between doctors and Watson,” said Dr. Baek Jeong-heum from Gil Medical Center who presented the study findings.

Watson produces three treatment options – “strongly recommend,” “recommend,” and “do not recommend” – based on expert data compiled from 290 medical journals, 200 textbooks, and 12 million pages, the hospital said.

Compiling data for both Watson’s strongly recommended and recommended treatment suggestions raised its concordance rate with physicians’ recommendations to 78.8 percent for colon cancer patients, the hospital said.

The concordance rate for rectal cancer and gastric cancer were equally high with 77.8 and 72.7 percent, respectively, Gil said.

In December last year, Gachon University Gil Medical Center became the first Korean hospital to install IBM’s Watson for Oncology – an artificial intelligence-based diagnostic tool that recommends treatment based on patients’ clinical information.

Watson for Oncology is now running in seven Korean hospitals – Gachon University Gil Medical Center, Konyang University Hospital, Pusan National University Hospital, Daegu Catholic University Medical Center, Keimyung University Dongsan Medical Center, Chosun University Hospital, Chonnam National University Hospital.

Except for Chonnam National University Hospital, six hospitals have formed an AI healthcare consortium in October to strengthen the medical technology innovation and publicity of medical institutions.

Patient satisfaction with the AI interdisciplinary care at its center reached 94 percent, the hospital said.

"The Watson for Oncology’s multidisciplinary treatment involves six doctors, so it's no exaggeration to say that each patient gets up to 180 minutes of diagnosis and treatment,” said Lee Eon, director of Gil Medical Center. “Patient satisfaction must be high since Watson makes its decisions based on a treatment plan based on countless patient cases.”

Using artificial intelligence healthcare will help medical workers get better results at a lower cost, reducing the burden of massive medical expenses caused by a super-aging society while maximizing patient satisfaction, Lee added.

Copyright © KBR Unauthorized reproduction, redistribution prohibited