Patterns of Failure After SBRT Reirradiation for Recurrent Head and Neck Cancer

Abstract

Stereotactic body radiotherapy (SBRT) is an emerging modality for the reirradiation of head and neck cancers (HNC). The need to use smaller reirradiation volumes to limit severe toxicity concerns represents a challenge in determining optimal target coverage. Herein, we investigate the associated patterns of failure after SBRT using previously defined spatial and dosimetric analysis (Mohamed et al, Radiat Oncol. 2017).SBRT reirradiation HNC cases treated between 2014-2019 and surveilled by CT, PET/CT, and/or MRI were reviewed retrospectively in this IRB approved study. Patients with radiographic evidence of recurrence within the head and neck region after SBRT were identified. Multifocal disease was defined as any non-contiguous FDG-avid or biopsy positive disease seen on imaging. Diagnostic scans were co-registered with planning CT scans (pCT) for manual segmentation of recurrent gross volumes (rGTV). rGTVs were then deformed to co-registered pCTs for failure classification. The classification followed a granular typology of 5 failure categories: central high-dose (type A or "in-field"), peripheral high-dose (type B), central low-dose (type C), peripheral low-dose (type D or "marginal"), and extraneous-dose (type E or "out of field"). Each type was determined based on the centroid location of rGTV relative to the planning target volumes plus the dose received by 95% of rGTV volume.A total of 106 SBRT cases were reviewed out of which 34 patients (32%) recurred after SBRT; 27 (79%) patients had squamous cell carcinomas, 7 (21%) received postoperative SBRT, and 24 (71%) received concurrent systemic therapy. Twenty-three (68%) patients received 42.5-45 Gy in 5 fractions (Fx) and 7 (21%) patients received 36 Gy in 6 Fx. Median time to failure was 5.5 months (range, 1-22 months). Twenty-two recurrences (65%) were out-of-field type E failures, 6 (18%) type A, 3 (9%) type C, and 3 (9%) type D. Fourteen of 16 (88%) multifocal SBRT reirradiation cases had Type E failures compared to 8 of 18 (44%) unifocal SBRT cases. Among the 6 in-field failures, 3 received 36 Gy in 6 Fx and 3 received 42.5-45 Gy in 5 Fx.Post SBRT reirradiation recurrences were seen in one-third of cases with the majority of failures being Type E (outside the treatment field), particularly in patients who were treated for multifocal disease. The use of 42.5-45 Gy in 5 Fx appears to be an effective biological dose for local tumor control. Marginal (Type C and D) failures may be reduced with larger margins, but limit feasibility of SBRT. Therefore, future reirradiation strategies should aim to improve patient selection, multimodal therapy, and margin definition in the HNC SBRT reirradiation setting.