Author: Mark E Artz, Julie Bradley, Hardev Singh Grewal, Perry B. Johnson, Christina Klassen, Raymond Mailhot Vega, Nancy Mendenhall, Jiyeon Park, Emma V. Viviers, Yawei Zhang 👨🔬
Affiliation: UF Health Proton Therapy Institute, University of Florida 🌍
Purpose: Verification CTs (VFCT) are used in radiotherapy to assess patient dose during treatment. However, they are time-consuming and contribute additional radiation exposure to the patient. This study evaluates the accuracy and effectiveness of CBCT-based synthetic CT imaging in monitoring the fractional dose distribution for breast radiotherapy.
Methods: A retrospective study was conducted on patients treated for unilateral breast cancer between 2022 and 2024. Of these, 26 patients received a CBCT on the same day as a VFCT. The ground truth CT (gtCT) was generated by deforming the VFCT to the same-day CBCT to minimize anatomical uncertainties. This gtCT was used as the dosimetric baseline. Two synthetic CT algorithms were evaluated: The corrected CBCT (corrCBCT), created using an iterative image correction method, and the virtual CT (virtCT), generated by deforming the planning CT to the CBCT with a low-density override. Clinical treatment plans were recalculated on the gtCT, corrCBCT, and virtCT for evaluation and the resultant dose metrics were compared.
Results: The corrCBCT and virtCT produced reliable dosimetric data compared with the gtCT. Target coverage, measured by the mean absolute dose difference in D95% to the breast or chestwall, was 0.1CGE(sd =0.1) for the corrCBCT and 0.2CGE(sd =0.1) for the virtCT. The left anterior descending (LAD) artery D0.1cc showed a mean dose difference of 0.5CGE(sd =0.6) for the corrCBCT and 0.9CGE(sd =1.4) for the virtCT. For lung V20CGE (standard fractionation) or V16CGE (hypofractionation), the mean absolute percentage volume difference was 1.3%(s =1.4) for the corrCBCT and 0.7%(sd =0.8) for the virtCT.
Conclusion: This study demonstrates the feasibility of using synthetic CTs (corrCBCT or virtCT) to calculate dose differences after anatomic or positional changes, to indicate VFCT for re-planning. This approach eliminates the need for routine VFCTs while maintaining an effective method for adaptive re-planning when dosimetric changes are detected.