2309 - Automatic Skin Flash Optimization in VMAT with Static-Angle Ports: Effect of Flash Margin Size on Breast and Chestwall Plan Quality and Robustness

E. Hubley¹, B. Koger¹, M. Salerno¹, R. M. Scheuermann¹, T. Li¹, L. Dong², and B. K. K. Teo¹; ¹Department of Radiation Oncology, University of Pennsylvania, Philadelphia, PA, ²Department of Radiation Oncology, Perelman School of Medicine, University of Pennsylvania, Philadelphia, PA

Purpose/Objective(s): A new optimizer creates VMAT plans with static-angle ports (VMATp), and allows for the addition of automatic skin (ASF) flash during arc delivery. ASF optimizes with an automatic virtual bolus and performs final dose calculation without, eliminating iterative optimizing/dose calculation. In this study, we compare the effects of differing flash margins used in the ASF tool on plan quality and robustness. Materials/

Methods: VMAT and VMATp plans were created for 13 breast and chestwall patients (9 left- and 4 right-sided). VMAT planning used 2 partial-arcs. VMATp plans used one partial-arc with 2 static ports at the position of standard tangent fields. For the VMATp plan, the collimator angle was chosen to maximize lung sparing for the static ports, and dynamic collimation rotation was allowed between ports. Plans were created with 6 MV, a maximum of 1000 iterations in the optimizer, and a static port weight of 50. The prescription dose was 4005 cGy in 15 fractions to the PTV (cropped 2-6 mm from the skin) and normalized to D_95% = 95%. The VMATp plan was reoptimized using 0-40 mm ASF with default HU (target median). Plan robustness was assessed by shifting the isocenter 5mm posteriorly to move the target into the flash region. CTV D_95% was compared between the shifted and unshifted plans. Plan quality was assessed with the DVH metrics presented in the table below. Statistical analysis was performed using a two-tailed paired t-test.
Results: After 5mm shift, CTV D_95% decreased by 5.9±3.9%, 4.9±3.1%, 1.1±2.3%, 0.4±0.4%, 0.4±0.5%, 0.2±0.4%, 0.4±0.9% for 0, 5, 7, 10, 15, 20, 40mm of flash. Adding flash beyond 10mm did not improve robustness. All DVH metrics except for contralateral lung were significantly increased (p<0.05) when ASF was used, compared to no ASF, though differences were small (table 1). OAR doses increased with increased ASF margin. Underlined values indicate statistical significance compared to 10mm ASF.
Conclusion: Plan quality decreased with a flash margin of 10mm or larger. Plans were robust to 5mm shifts with 7-10mm ASF. Given this, an ASF margin equal to the expected uncertainty (5mm in this study) plus the amount of target crop from skin provided the best compromise between robustness and plan quality. Abstract 2309 – Table 1

		Auto Skin Flash Margin (mm)
		0	5	7	10	15	20	40
PTV	V105%[%]	0.6±1.0	6.3±9.7	9.9±12.9	16.2±18.7	19.3±18.9	22.8±18.8	44.3±18.7
Breast_Contra	Dmean[Gy]	0.9±0.4	0.9±0.4	1.0±0.4	1.0±0.4	1.0±0.4	1.0±0.4	1.3±0.5
Heart	Dmean[Gy]	1.1±0.4	1.2±0.4	1.2±0.4	1.1±0.4	1.2±0.4	1.3±0.5	1.9±0.9
Lung_Contra	Dmean[Gy]	0.6±0.3	0.6±0.6	0.7±0.3	0.7±0.3	0.7±0.3	0.8±0.3	1.0±0.3
	V4Gy[%]	0.4±0.7	0.4±0.8	0.5±0.8	0.4±0.7	0.5±0.8	0.9±1.0	1.9±1.7
Lung_Ipsi	Dmean[Gy]	5.3±1.0	5.4±1.0	5.5±1.1	5.5±1.1	5.7±1.2	5.8±1.2	7.0±1.6
	V16Gy[%]	11.3±3.2	11.7±3.3	11.8±3.5	12.0±3.5	12.2±3.5	12.5±3.7	15.5±4.4
	V4Gy[%]	28±5.5	28.8±5.4	29.1±5.5	29.2±5.5	29.8±5.7	31.0±6.4	37.9±8.1