SBRT For Head and Neck Skin Cancer: An Initial Experience in 106 Medically Unfit Patients

Abstract

Stereotactic Body Radiation Therapy (SBRT) is an alternative treatment strategy for patients with head and neck skin cancer (HNSC) unable to undergo a protracted radiotherapy course. We report outcomes of SBRT for HNSC patients treated at a tertiary high-volume institution. A retrospective review of patients who received SBRT from 2012-2019 was conducted. Kaplan-Meier method was used to estimate the one-year local control (LC), locoregional control (LRC) outside of SBRT field, overall survival (OS), progression-free survival (PFS) and late toxicity (LT) rates from time of SBRT completion. Univariate and multivariate Cox proportional hazard models were performed. Acute and late grade 3 and 4 toxicity were reported according to the Common Terminology Criteria for Adverse Events v4.0. A total of 106 patients were identified with 112 lesions treated. The cohort consisted of patients treated for primary (n = 51), nodal (n = 47) and both primary and nodal disease (n = 8). The T and N categories prior to SBRT were Tx-2: n = 78, T3-4: n = 28; N0-1: n = 67, N2-3: n = 39. Median age at diagnosis was 84 years with median ECOG performance status of 2. HNSC histologies included cutaneous squamous cell carcinoma (n = 77), basal cell carcinoma (n = 14), melanoma (n = 8), Merkel cell carcinoma (n = 6), and 1 unconfirmed. Fractionation schemes ranged from 32-50 Gy in 4-6 fractions, with the most common regimens being 45 Gy in 5 fractions [52% (n = 55)] and 40 Gy in 5 fractions [32% (n = 34)]. Median gross tumor volume (GTV) was 31 cm3 (17-56) and mean planning tumor volume (PTV) dose was 44 Gy. Median follow-up was 8 months and actuarial median OS was 14 months. The one-year LC, LRC outside of SBRT field, OS, PFS and LT rates were 78% (69-88), 72% (62-84), 53% (43-65), 52% (40-62), and 25% (16-38), respectively. Larger GTVs were associated with reduced PFS [HR 1.58, p = 0.004], OS [HR 1.37, p = 0.039], and LRC outside of SBRT field [HR 1.83, p = 0.006] but not LC [HR 0.87, p = 0.50]. Multivariable models revealed that larger GTVs [HR 1.47, p = 0.046] and higher biologically effective dose (BED10) [HR 1.05, p = 0.03] were associated with a higher rate of late toxicity. There was no statistical difference between proportions of grade ≥3 acute (p = 0.22) and late toxicities (p = 0.37) in patients treated with doses ≤40 Gy vs. >40 Gy. For acute toxicity, 41 patients had grade 3, and 1 had grade 4. The most common acute toxicity was dermatitis (n = 39). Twenty-eight patients (26%) experienced ≥ late grade 3 toxicity, with the most common being fibrosis (n = 19) and soft tissue ulceration (n = 6). No grade 5 toxicity was observed. SBRT in HNSC provides durable disease control with acceptable toxicity for select medically unfit patients unable to undergo extended radiotherapy courses. Predictive factors of treatment response and toxicity include GTV and BED10. Further prospective studies including quality of life measures and dosimetric correlations with larger cohorts are needed.