Yu Lei, Qianxi Ni, Binbing Wang, Kang Zhang, Feng Shi, Shixiong Huang, Guoping Shan, Zhong Yang, Ying Guo, Zhen Zhang, Jiazhou Wang, Weigang Hu
Deep learning (DL) -based automated treatment planning (ATP) shows significant promise in streamlining radiotherapy workflow and reducing variability in plan quality. However, it often lacks the flexibility needed for achieving individualized trade-offs in real-world practice. Herein, we propose a hybrid strategy by integrating DL-based dose prediction with clinical-goal-guided inverse optimization to generate directly deliverable plans within five minutes. DL models for five disease sites were trained separately using datasets from a single institution and were tested retrospectively for clinical application among three institutions, with tailored prioritized clinical goals. We find that over 80% of the 250 auto-plans met clinical criteria, and 60% were preferred over manual plans in blinded reviews. Dosimetric analyses show that the auto-plans quantitatively matched or exceeded the quality of human-driven plans. This study highlights ATP’s potential to transform radiotherapy practice, with ongoing efforts aimed at refining its versatility and adoption across diverse clinical settings. Deep learning (DL)-based automated treatment planning has the potential to improve radiotherapy planning for patients in cancer but how best to implement them in the clinic is less clear. Here, the authors propose a hybrid strategy to develop a DL-based automated radiotherapy planning solution and demonstrate its real-world effectiveness in producing clinically deliverable plans.