A Biological Robustness Planning Method for Carbon Ion Radiation Therapy
Abstract
Purpose
Carbon ion radiotherapy (CIRT) offers high LET and RBE for treating radioresistant tumors but suffers from biological dose inhomogeneity due to RBE uncertainties. Techniques like LET painting focus on optimizing LET but do not directly address biological effects. Multi-gantry and multi-ion planning can improve dose uniformity and reduce exposure to healthy tissue. However, uncertainties in RBE models and parameters remain a concern, particularly as they may worsen dose inhomogeneity within the target. This study develops a robust optimization framework to improve biological dose stability in CIRT.
Methods
A probabilistic biologically robust optimization method was developed using three α/β variation scenarios within the target based on the LEM I model. An iterative Jacobian-based linearization algorithm addressed the nonlinear optimization arising from DVH, minimum monitor unit (MMU), and LQ-based biological dose constraints. The method was evaluated in a head-and-neck case with a prescription of 3 Gy RBE. Robustly optimized plan (Bio_rob) was compared with nominal-only plan Non_rob, and worst-case renormalized plans Bio_rob2.
Results
With a 10% α/β variation, nominal planning led to ~3% average biological dose variation, underestimation of hot spots (up to 15% in volume or 2% in dose), and reduced target coverage (to 80% in volume vs. planned 90%). The robust method reduced hot-spot volume by 12% (2% in dose) and improved target dose homogeneity. Even after renormalization to restore worst-case coverage, hot-spot dose was reduced by ~1%.
Conclusion
The proposed robust biological optimization improves biological dose homogeneity, hot-spot control, and worst-case target coverage in CIRT, enhancing plan reliability under biological uncertainties.