Skin Dose Variations during Tomotherapy for Head and Neck Cancer Patients: Need for Adaptive Replanning?

Abstract

An increase of skin dose during RT due to anatomy deformation may potentially translate into an increased risk of severe acute toxicity, likely to translate into an increase of late fibrosis, suggesting a potential benefit from planning adaptation (ART) to counteract this effect. Current study applied an original method for dose-of-the-day calculation in quantifying changes of the skin dose during Tomotherapy (HT) for head and neck (HN) cancer with the aims to identify those patients (pts) that could benefit of ART. Planning CTs of 18 HN pts previously treated with radical HT (SIB: 54/66Gy, 30fr) at our Institute were deformable registered to MVCT images taken at fraction (fr) 15 and 26 using a constrained intensity-based algorithm. Dose-of-the-day on the resulting deformed images (CTdef15 and CTdef26) were calculated; this approach (using planning CT-to-MVCT elastic registration) was previously validated by comparing the dose distributions on CTdef against the ones calculated on diagnostic kVCTs taken at the same day. In order to assess changes of the skin dose, superficial layers with thickness of 2 mm (SL2) were considered in the body cranial-caudal extension corresponding to the high-dose PTV. SL V95%-V107% of the prescribed PTV dose (i.e.: 66 Gy, 2.2 Gy/fr) were recalculated at fr 15 and 26 and compared against planning values: changes of SL2 were considered a surrogate of skin dose changes. Based on the previous validation study, the approach was found to be accurate in assessing V95%-V107% of SL2mm within ±1cc: a change ≥+1cc was preliminarily considered to define those pts who may benefit of ART. Changes of skin dose occur during HT with individually variable patterns likely to depend on several factors like the PTV location/dimension, the shape/homogeneity of the dose distribution and the body shrinkage at the level of the high-dose PTV. The average changes of V95%, V100% and V105% at fr 15 were -1.5 cc (1SD: 2.9cc), 0.0 (1,3) and +0.2 (0.5). Focusing on V100% (i.e.: 2.2 Gy/fr), most pts showed changes within ±1cc (13/18 and 11/18 at fr 15 and 26) and few had changes ≤-1cc (2/18 and 3/18). Two and 4 pts showed changes ≥ +1cc at fr 15 and 26 respectively, with values up to 3.4cc, corresponding to about 16 cm2 of skin surface. For 3 out of the 4 pts with a change ≥ +1cc, a pattern of increase from planning to fr 26 was visible. Notably, if considering V105%, only 1 pt showed a change ≥ +1cc at fr 26, likely due to the high homogeneity of the planned dose distribution. A method to carefully estimate changes of the skin dose during HT was able to individually select pts that may benefit of ART. First results suggest that 10-20% of HN pts treated at our Institute could be candidate to ART to counteract skin dose increase. Further investigation is necessary to corroborate these results on a larger population and to define sub-categories whose changes may be expected to be more likely and for which an ad-hoc skin dose-of-the-day check may be justified and clinically implemented.