A retrospective, deformable registration analysis of the impact of PET-CT planning on patterns of failure in stereotactic body radiation therapy for recurrent head and neck cancer

Background: Stereotactic body radiation therapy (SBRT) has seen increasing use as a salvage strategy for selected patients with recurrent, previously-irradiated squamous cell carcinoma of the head and neck (rSCCHN). PET-CT may be advantageous for tumor delineation and evaluation of treatment failures in SBRT. We analyzed the patterns of failure following SBRT for rSCCHN and assessed the impact of PET-CT treatment planning on these patterns of failure. Methods: We retrospectively reviewed 96 patients with rSCCHN treated with SBRT. Seven patients (7%) were treated after surgical resection of rSCCHN and 89 patients (93%) were treated definitively. PET-CT treatment planning was used for 45 patients whereas non-PET-CT planning was used for 51 patients. Categories of failure were assigned by comparing recurrences on post-treatment scans to the planning target volume (PTV) from planning scans using the deformable registration function of VelocityAI (TM). Failures were defined: In-field (>75% inside PTV), Overlap (20-75% inside PTV), Marginal (<20% inside PTV but closest edge within 1cm of PTV), or Regional/Distant (more than 1cm from PTV). Results: Median follow-up was 7.4 months (range, 2.6-52 months). Of 96 patients, 47 (49%) developed post-SBRT failure. Failure distribution was: In-field-12.3%, Overlap-24.6%, Marginal-36.8%, Regional/Distant-26.3%. There was a significant improvement in overall failure-free survival (log rank p = 0.037) and combined Overlap/Marginal failure-free survival (log rank p = 0.037) for those receiving PET-CT planning vs. non-PET-CT planning in the overall cohort (n = 96). Analysis of the definitive SBRT subgroup (n = 89) increased the significance of these findings (overall failure: p = 0.008, Overlap/Marginal failure: p = 0.009). There were no significant differences in age, gender, time from prior radiation, dose, use of cetuximab with SBRT, tumor differentiation, and tumor volume between the PET-CT and non-PET-CT groups. Conclusions: Most failures after SBRT treatment for rSCCHN were near misses, i.e. Overlap/Marginal failures (61.4%), suggesting an opportunity to improve outcomes with more sensitive imaging. PET-CT treatment planning showed the lowest rate of overall and near miss failures and is beneficial for SBRT treatment planning.