Jang Ji-seok, Professor of Radiation Oncology, Yonsei University Cancer Center
Normal Organ Compartmentalization During Radiation Therapy - Bringing AI to the Treatment Planning Stage
Participated in the development of OncoStudio,an AI software for radiation therapy
Terminal cancer patients with distant metastases also improve survival - keeping hope alive for a cure
“The diagnosis of biliary tract cancer was embarrassing enough, but when I found out it had metastasized to my lungs,I was blindsided. I never thought I would be able to attend my son's graduation ceremony. He is a lifesaver for me and my family.”
Seo Ji-hyeok (65, pseudonym)was diagnosed with stage 4 biliary tract cancer in July 2019 after visiting a hospital with unexplained indigestion and jaundice.
At the time, the doctors said, “The biliary tract cancer has metastasized to the lungs and he has six months to live at most,” but he is still alive with chemotherapy for nearly five years. He also sent his regards to Jang Ji-seok, a professor of radiation oncology at Yonsei Cancer Hospital, who has been a clinical instructor at BC Cancer Centre in Vancouver, Canada, since August 2022, along with a photo from his son's graduation ceremony. How is it possible for a patient with terminal cancer, the worst form of cancer with a survival rate of only 3%, to survive for such along period of time?
◇ Biliarytract cancer, which is difficult to diagnose early... 'only 3% survival rate'if it metastasizes remotely
담도암은 지방의 소화를 돕는 담즙(쓸개즙)이 배출되는 통로(담도)에 발생하는 암이다. 일반적으로 간에서 만들어진 담즙이 십이지장으로 흘러들어가기까지 경로인 담도뿐 아니라 담즙을 일차적으로 저장하는 장소인 담낭에 생긴 악성 종양을 통틀어 담도암이라고 부른다.
국가암등록통계를 보면 2021년 한해 동안 담낭 및 기타 담도암으로 새롭게 진단된 환자는 7619명으로 전체 암 발생자(24만7952건)의 3.9%를 차지했다. 국내에서 9번째로 많이 발생하는 암이지만 조기 발견이 어려워 예후가 나쁜 암으로 꼽힌다. 초기에는 황달 등의 증상이 전혀 없고, 진행되면 소화불량·복통·체중감소 등 특이하지 않은 증상이 동반되다 보니 건강검진을 통해 우연히 진단받는 경우가 대부분이다.
발견이 늦어지니 예후도 좋지 않다. 국가암등록통계를 보면 2017~2021년 담낭 및 기타 담도암 환자가 일반인과 비교해 5년간 생존할 확률을 뜻하는 ‘5년 생존율’은 50.0%에 그쳤다. 그나마 5년 생존율이 절반 정도인 것도 암이 처음 발생한 장기를 벗어나지 않은 국한 단계일 때에 한한다. 주위 장기나 인접한 조직 혹은 림프절을 침범한 국소 진행단계인 경우 5년 생존율이 34.2%, 멀리 떨어진 다른 부위로 전이된 원격 전이 단계에서는 3.2%까지 떨어진다.
◇ 방사선치료 정밀도 향상…수술 못하는 말기암 환자에 새로운 치료 대안으로
Surgery is currently the most reliable treatment to improve survival rates for biliary tract cancer patients. However, only about 40 to 50 percent of all patients are able to undergo radical resection (removal of the entire compartment along with the tumor), and those who cannot under go surgery, like Mr. Seo, are treated with chemotherapy or radiation to relieve symptoms. Seo first met Prof. Jang when he came to Yonsei Cancer Hospital for radiation therapy after being diagnosed with terminal cancer.
It wasn't always a pleasant encounter.After six to seven weeks of chemotherapy and radiation treatments once a day,Monday through Friday, the cancer cells had spread to every nook and cranny of her body. For Seo, the hardest part was hearing the results of the imaging tests after the planned treatment.
“You don't know how fast radiation therapy is advancing these days, it's almost as good as surgery for some cancers. You should be in your son's graduation picture.”
Whenever Professor Jang would report that a metastatic lesion had enlarged or that a new metastasis had appeared, he would introduce the non-medical scientist to the breakthroughs in radiotherapy.
◇ AI algorithm applied...“Reduced 3D long-term contouring time by 90%”
The process of using images from computed tomography (CT) and magnetic resonance imaging (MRI) scans to outline organs and plan radiation treatments, combined with cutting-edge technologies such as artificial intelligence (AI), has led to faster and more effective treatments.
For so-called “oligometastasis,” where the cancer is confined to a small number of metastases, even if it's stage 4,there's hope that stereotactic body radiation therapy (SBRT), a highly precise treatment that delivers high doses of radiation over three to five treatments,could lead to long-term survival and even a cure.
After trusting Professor Jang and continuing to undergo radiation therapy, Seo was told that in less than two years, “all the metastatic lesions had disappeared.” More than three years later, she continues to receive the same treatment with no new metastases.
OncoStudio, which OncoSoft began developing shortly after its launch and obtained a Class 2 medical device license from the Ministry of Food and Drug Safety, is an AI-based automatic contouring software.
It automates the contouring work that was done manually by medical staff, reducing the work time by 90%.
It's like a medical version of the image-generating AI Midjourney.
By utilizing OncoStudio's algorithm, it is possible to reduce the work time from 1-2 days to a few minutes, and it is highly accurate.
Prior to his overseas training, Prof. Jang was directly involved in the development of OncoStudio's AI-based contouring(compartmentalization) software for radiotherapy. Radiotherapy, which uses high doses of radiation to kill cancer cells, is all about protecting surrounding normal tissue and minimizing side effects. Precise radiotherapy planning requires technology that can contour normal organs on images such as CT-MRI and predict the radiation dose to each cancerous tissue and organ.
OncoStudio automates this process, which was previously done manually by medical staff, with AI. OncoStudio's AI algorithm reduces the time it takes to contour (segment) normal organs in three-dimensional (3D) images by 90%. This reduces the time from three to four hours to one to two days to a matter of minutes, while increasing accuracy.This effect is expected to be maximized as AI technology is applied to various stages of radiation therapy.
OncoStudio was approved as a Class 2medical device by the Ministry of Food and Drug Safety in February 2022 and is used in more than 50 radiation oncology departments in Korea, including Severance Hospital and Samsung Medical Center. Several overseas institutions are also showing interest, including a hospital in Canada where Prof. Jang is currently working.
“BC Cancer Center is testing several products to implement an AI contouring program,” said Prof. Jang. ”When we demonstrated OncoStudio to local doctors, they praised it for being faster and more accurate than other products.”
He will be returning to Yonsei Cancer Hospital in August this year and is working with OncoStudio and local researchers on a collaborative research project. The ultimate goal is to use AI technology to automatically locate the location of metastasized cancer and calculate the speed of the disease to identify when and under what indications body stereotactic radiotherapy can help.
He emphasized, “AI algorithms can help reduce variation between hospitals and improve care efficiency” and “I want to contribute to conquering incurable cancer as well as improving the overall quality of treatment by allowing cancer patients and doctors to communicate closely.”
