SU-G-201-14: Is Maximum Skin Dose a Reliable Metric for Accelerated Partial Breast Irradiation with Brachytherapy?




To evaluate the reliability of the maximum point dose (Dmax) to the skin surface as a dosimetric constraint, we investigated the correlation between Dmax at the skin surface and dose metrics at various definitions of skin thickness.


42 patients treated with APBI using a Strut Adjusted Volume Implant (SAVI) applicator between 2010 and 2014 were retrospectively reviewed. Target (PTV_EVAL) and organs at risk (OARs: skin, lung, and ribs) were delineated on a CT following NSABP B-39 guidelines. Six skin structures were contoured: a rind 3cm external to the body surface and 1, 2, 3, 4, and 5mm thick rinds deep to the body surface. Inverse planning simulated annealing optimization was used to deliver 32–34Gy in 8-10 fractions to the target while minimizing OAR doses. Dmax, D0.1cc, D1.0cc, and D2.0cc to the various skin structures were calculated. Linear regressions between the metrics were evaluated using the coefficient of determination (R2).


The average±SD PTV_EVAL volume and cavity-to-skin distances were 71.1±28.5cc and 6.9±5.0mm. The target V90 and V95 were 97.3±2.3% and 95.1±3.2%. The Dmax to the skin structures were 78.7±10.2% (skin surface), 82.2±10.7% (skin-1mm), 89.4±12.6% (skin-2mm), 97.9±15.4% (skin-3mm), 114.1±32.5% (skin-4mm), and 157.0±85.3% (skin-5mm). Linear regression analysis showed D1.0cc and D2.0cc to the skin 1mm and Dmax to the skin-4mm and 5mm were poorly correlated with other metrics (R2=0.413±0.204). Dmax to the skin surface was well correlated (R2=0.910±0.047) and D1.0cc to the skin-3mm was strongly correlated with all subsurface skin layers (R2=0.935±0.050).


Dmax to the skin surface is a relevant metric for breast skin dose. Contouring discontinuities in the skin with a 1mm subsurface rind and the active dwells in the skin 4 and 5mm introduced significant variations in skin DVH. D0.1cc, D1.0cc, and D2.0cc to a 3mm skin rind are more robust metrics in breast brachytherapy.