Script-Based Automatic Intensity-Modulated Radiotherapy Planning with Robust Optimization for Craniospinal Irradiation: A Retrospective Study

Background Craniospinal irradiation (CSI) is essential for treating central nervous system malignancies. However, it is very challenging and time-consuming to design a robust treatment plan capable of delivering a uniform dose to the entire treatment volume for standard linear accelerators. This study proposes a script-based automatic planning method with robust optimization for CSI to reduce sensitivity to setup errors and increase planning efficiency. Methods Ten CSI patients with planning target volume (PTV) lengths between 49.8 and 85.0 cm were retrospectively studied. IMRT plans with robust optimization were generated by the automatic planning script. The plans’ sensitivities to positional inaccuracy were evaluated by deliberately shifting the middle beamsets ±5 mm longitudinally. Non-robust optimization plans were also created and compared to the robust ones. The planning time was recorded to evaluate the efficiency of the automatic CSI planning. Results For the ten patients enrolled in the study, homogeneous dose coverage were achieved for both robust and non-robust plans with average conformity index value of 0.827 (±0.028) and 0.841 (±0.023) (p = 0.047), homogeneity index values of 0.076 (±0.014) and 0.078 (±0.027) (p = 0.114), respectively. In robust plans, the variations of D1% and D99% were between -0.17–0.99% and -1.18–0.53% for cranial PTV, and -1.02–7.2% and -4.78–0.31% for spinal PTV, while in non-robust plans, the variations were much more significant with corresponding values of -0.53–8.31% and -5.17–0.57% for cranial PTV, and 0.53–40.14% and -46.06% to -2.01% for spinal PTV. The total planning time of a robust optimization plan was about 9-13 minutes by the automatic planning script. Conclusions The script-based automatic CSI planning with robust optimization could efficiently create a uniform dose coverage for the entire target volume regardless of PTV length, with the advantage of being insensitive to setup errors and standardizing the planning process.