Delivered dose for preoperative single fraction partial breast radiation therapy

Purpose/Objective(s): To evaluate patient positioning and investigate the impact of changes in breast shape on planned versus actual delivered dose for preoperative single fraction partial breast radiation therapy. Materials/Methods: Thirty-two patients underwent preoperative single fraction partial breast radiation therapy for early stage breast cancer. A 0.3 cm margin was added from clinical target volume (CTV) to planning target volume (PTV). Patients were initially positioned using the skin marks placed at the time of planning CT simulation. The initial setup was corrected with on-line image-guidance using 2D kV images followed by 3D cone-beam CT (CBCT) prior to treatment. Post-treatment 2D or 3D images were acquired for all patients. The initial setup error was determined as the difference between the position based on skin markers and the position based on 2D kV images. The residual setup error was determined as the difference between 2D kV images and 3D CBCT. The intra-fractional motion was determined as the difference between pre-treatment and posttreatment images. Deformable registration was performed to register the planning CT to the pre-treatment CBCT retrospectively. The breast outlines from the deformed CT and CBCT were visually confirmed to be aligned after deformable registration. Dose was reconstructed based on the deformed planning CT to generate delivered plans. Delivered plans and original plans were compared. Results: The mean initial setup error, mean residual setup error, and mean intra-fractional motion were 1.00 ± 0.75 cm, 0.28 ± 0.42 cm and 0.21 ± 0.18 cm, respectively. The mean CTV volume receiving 95% of prescription dose (V95%) and PTV V95% were reduced by 0.9 ± 2.4% (p = 0.04) and 2.2 ± 3.9% (p < 0.01), respectively, in delivered plans compared to original plans. The differences in homogeneity index and conformity index were negligible. The mean dose to 10 cm3 (D10cm 3 ) and maximum dose (Dmax) for skin were greater in delivered plans than in original plans by 1.5 ±2.8% (p < 0.01) and 2.8 ± 8.0% (p = 0.06). Four patients had a planned skin Dmax higher than a suggested dose constraint based on previous studies (e.g., skin Dmax > 18 Gy). After accounting for breast deformation, only two more patients had a delivered skin Dmax exceeding 18 Gy. Other structures such as heart, lung, and breast showed non-significant dose differences (< 1%) between delivered plans and original plans. Conclusions: This study showed that a 0.3cm PTV margin was able to provide good CTVand PTV coverage (i.e., V95%) even after accounting for set up error, patient motion, and difference in breast shape with real-time IGRT. Breast deformation can result in a delivered skin dose greater than the planning dose. Because this may contribute to toxicity outcomes, the treatment plan should be carefully optimized to minimize dose to the skin.