Evaluation of radiotherapy planning approaches for head and neck patients with tumors close to the skin surface.

Abstract

In radiotherapy of the head and neck (H&N) it is common for the clinical target volume (CTV) to extend to the patient's skin. Adding a margin for set-up uncertainty and delivery creates a planning target volume (PTV) that extends beyond the patient surface. This can result in excessive fluence being delivered to the build-up region and therefore the skin. This study evaluates four different planning methods used when inverse-planning H&N radiotherapy treatments with CTV extending to the skin. The aim of the study was to determine which planning method gives superior plan quality. Ten H&N cancer patients with a CTV contoured to the skin were inverse-planned using four planning methods. The planning methods compared were: cropping the optimization PTV back from the skin surface by 5.0, 3.0, and 0.0mm and a virtual bolus method. For each planning method, the increased fluence at the skin surface was analyzed. The CTV coverage and skin doses were compared. Plan robustness was evaluated by applying an isocenter shift of ±3.0mm in the major axes. The planning method cropping the PTV 0.0mm from the skin surface results in an increased fluence in the build-up region. The average volume of CTV receiving 98% of the prescription dose was 89.6%±3.4%, 91.6%±2.4%, and 93.5%±1.7% when cropped 5.0, 3.0, and 0.0mm, respectively, and 93.4%±2.1% for the virtual bolus method. Introducing plan uncertainty affects CTV coverage the most when cropping 5.0mm. When plan uncertainties are considered the methods of cropping 5.0, 3.0mm, and the virtual bolus method have the same average skin dose within ±0.3%. This study shows that a virtual bolus planning method results in no increased fluence at the patient's surface, improves CTV coverage, and is the most robust to changes in setup and patient anatomy.