Science

Is Artificial Intelligence the future of radiotherapy?

physicsworld.com

Dec 27, 2024, 16:40

Is Artificial Intelligence the future of radiotherapy?

David Sher, Chief of Head and Neck Radiation Oncology Service, shared a recent article by Yu-Xian Yang on X:

“Is This The Future Of Radiotherapy?

Offline: immobilization, MRI simulation, AI then MD-verified GTV contours.

No Plan.

Online: At linac, fan-beam CT, AI-generated contours, verified by MD. AI-generated Plan, verified by MD, then treat.

Median time: 23 minutes.”

Title: Artificial Intelligence-Empowered Multistep Integrated Radiation Therapy Workflow for Nasopharyngeal Carcinoma

Authors: Yu-Xian Yang, Xin Yang, Xiao-Bo Jiang, Li Lin, Guang-Yu Wang, Wen-Zhao Sun, Kang Zhang, Bing-Huan Li, Hua Li, Le-Cheng Jia, Zi-Quan Wei, Yan-Fei Liu, Dan-Ning Fu, Jun-Xiang Tang, Wei Zhang, Jing-Jie Zhou, Wen-Chao Diao, Ya-Juan Wang, Xue-Mei Chen, Chen-Di Xu, Liu-Wen Lin, Jia-Ying Wu, Jia-Wei Wu, Li-Xia Peng, Jin-Fa Pan, Bing-Zhong Liu, Chi Feng, Xiao-Yan Huang, Guan-Qun Zhou and Ying Sun

Is Artificial Intelligence the future of radiotherapy?

“MD time at console: ~ 5 minutes.

DSC coefficients (especially GTVp) make it clear the MD is still needed, but overall the system-generated targets, OARs, and plans are acceptable.

Comparison of offline contours/plans (done 6 months later) show no meaningful difference to online workflow.”

“One can quibble with the specifics of the workflow (if there is an offline mask and contours, why not do it all offline so it’s even faster for the patient?), but the main takeaway is that the RT treatment process is changing dramatically.

Faster, with less human interaction.”

“In how many years will we do everything offline (contours based on diagnostic imaging, pre-plan if even necessary), and then bring the patient back, on Day 1, for mask and treatment?”

For more updates, follow OncoDaily.