Target delineation in stereotactic body radiation therapy for recurrent head and neck cancer: a retrospective analysis of the impact of margins and automated PET-CT segmentation.

BACKGROUND: Few guidelines exist on stereotactic body radiation therapy (SBRT) treatment planning for recurrent head and neck cancer. We assessed the impact of retrospectively adding margins/automated PET volumes to the gross tumor volume (GTV) in patients with post-SBRT recurrences.
MATERIALS AND METHODS: We reviewed 89 patients with recurrent head and neck cancer treated with SBRT using no margin around the GTV. GTVs were recontoured with 1-5mm margins. PET-CT planned GTVs were also recontoured by adding PET-standardized uptake value (SUV)(3.5), SUV(4.5), SUV(40% max), and signal/background ratio (SBR) to the original GTV. We deformably registered recontoured GTVs to post-SBRT scans and assessed fraction of recurrence volume (RV) falling within the GTV, the "RV-GTV overlap."
RESULTS: With non-PET-CT planning, median RV-GTV overlap increased from 11.7% to 48.2% using 5mm margins, and median GTV size increased by 41.8 cc (156%). With PET-CT planning, RV-GTV overlap increased from 45% to 93.6% using 5mm margins, and GTV size increased by 34.8 cc (140%). Adding SUV(3.5) and SBR increased RV-GTV overlap from 45% to 73.3% and 73.6%, with GTV size increases of 0.8 (3%) and 3.1 cc (11%), respectively.
CONCLUSIONS: Recontouring increased recurrence coverage and also GTV size. Margins up to 5mm may reduce failures but could possibly increase toxicities. Automated PET contours may reduce near-miss failures with smaller increases in GTV size.