Purpose: Tumor shrinkage during lung SBRT has been reported. Adaptive SBRT planning can potentially escalate dose to the target and mitigate dose to surrounding tissues. In this study, we investigated dosimetric effects of adaptive lung SBRT planning using cone‐bean CT (CBCT) images. Methods: 48 patients who had SBRT treatments for lung cancer in our institution were included in this study. All patients underwent a 3D helical CT scan and 4DCT scan for treatment planning. Treatments were delivered on Linac machines equipped with kV‐OBI, CBCT, and MV EPID. CBCT images of all patients were retrospectively analyzed to determine tumor size changes. 10 patients with the largest percentages of tumor shrinkage were selected for adaptive planning. Plans were created on CBCT images of each fraction using the same planning parameters as the original CT‐based plans. ITV was contoured on CBCT. PTV was generated by adding 5 mm margin to ITV. Two accumulative dose plans, non‐adaptive (P‐NON) and adaptive (P‐ADP), were determined for each patient using deformable image registration via Velocity AI. Dosimetric comparisons between PNON and PADP were performed for all OARs using RTOG0915 metrics. Statistical significances were tested using paired t‐test with a significance level of 0.05. Results: Comformality ranged from 1.03 to 1.41, with small intra‐subject variations (ranged from 0.01 to 0.06). Compared to PNON, PADP reduced all dosimetric metrics and significantly in the following (p‐value range: 0.02–0.04): V20Gy of lungs; D0.35cc, D1.2cc and maximum dose of cord, D5cc and maximum dose of esophagus, D30cc of chest wall, and D15cc of heart. Coverage to original GTV is comparable in two plans. Conclusions: Adaptive SBRT planning for selected patients resulted in significant dose reduction for adjacent OARs. It has the potential to further escalate dose to the target and mitigate dose to surrounding tissues.
ASJC Scopus subject areas
- Radiology Nuclear Medicine and imaging