Knowledge-Based Planning for Ventricular Tachycardia SBRT
Abstract
Purpose
Stereotactic Body Radiation Therapy (SBRT) has gained attention for the treatment of patients with refractory ventricular tachycardia (VT). VT substrates are complex, multifocal, progressive, and may require urgent treatment. As a result, plan quality can vary widely due to differences in planner experience, time constraints, and planning priorities driven by limited literature and clinical trial data, as well as substantial variability in target size and location. This study sought to develop a standard knowledge-based planning (KBP) model and test its feasibility in producing quality treatment plans for heterogeneous VT cases.
Methods
The KBP model was trained using 33 high quality VT-SBRT cases that varied in terms of target size and proximity to critical structures. The model was evaluated retrospectively on 15 VT-SBRT cases with PTV volumes ranging from 26.3 to 200.7 cc. All plans were prescribed 25Gy/1fx and achieved similar target coverage. The KBP and clinically treated plans were compared using clinical goals from the literature and RAVENTA protocol, conformity index (CI), and gradient index (GI).
Results
The KBP model produced VT-SBRT plans that achieved all clinical goals. On average, KBP maximum doses to critical structures were as follows: aorta: 11.2 Gy; esophagus: 6.7 Gy; bronchial tree/trachea, 2.3 Gy; spinal canal: 2.8 Gy; and ICD: 0.108 Gy. The KBP model produced plans with average CI and GI of 0.92 and 3.14, respectively, 0.89 and 3.24 in the clinical plans. Overall, the KBP model was able to generate plans that satisfied the constraints from the literature and were comparable to the clinically approved plans.
Conclusion
A VT-SBRT KBP model is feasible for clinical integration and can standardize planning, potentially improving the efficiency for highly irregular VT anatomies and substrates.