Comparison of Clinical Outcomes of Different Radiation Strategies in Postoperative Radiation Therapy for Patients with Head and Neck Squamous Cell Carcinoma: A Propensity-Score Matched Analysis

Abstract

Postoperative radiotherapy (PORT) may be useful for head and neck squamous cell carcinoma patients (HNSCC) with pathological risk factors. We previously reported an unfavorable loco-regional control rate (LRC) after limited field (LF)-PORT. Thus, we changed our strategy to cover the whole neck (WN-PORT) region and retrospectively analyzed the outcomes of HNSCC patients to compare the efficacy of the two different strategies. Before 2010, the clinical target volume covered the tumor bed without including the lymph node region for close/positive margin cases and only involved the neck region for multiple nodes and/or extracapsular extension (ECE). Since 2011, whole neck irradiation was planned with 3D conformal RT or intensity-modulated radiation therapy. Univariate analysis, multivariate analysis, and propensity-score matched analysis were performed. The study included 275 patients: 186 received LF-PORT and 89 received WN-PORT. The median follow-up for the entire cohort was 40.8 months (range, 2.3–172 months), 46.9 months for LF-PORT and 27.9 months for WN-PORT. Patient characteristics of the entire cohort were balanced in terms of age, sex, primary site, and pathological stage. However, the rates of ECE, positive surgical margin, and poor differentiation were significantly larger in the WN-PORT group. The 2-year overall survival rate (OS), progression-free survival rate (PFS), and LRC for the entire cohort were 66.8%, 49.7%, and 70.4%, respectively. In univariate analysis, favorable factors for OS were the tumor site, pTstage, pNstage, ECE, and pathological differentiation. Although the radiation strategy did not have a significant effect on either OS (P = .106) or PFS (P = .087), the WN-PORT group had a significantly better LRC (HR, 0.38; 95% CI, 0.21–0.67; P < .001). In multivariate analysis, unfavorable factors for OS were advanced pTstage and pNstage, high risk (ECE and/or close/positive margin), poorly differentiated tumor, having previous curative surgery, and LF-PORT; WN-PORT was a favorable factor for OS (HR, 0.49; 95% CI, 0.31–0.78; P = .003). Propensity-score matching resulted in a cohort consisting of 118 patients who were well matched and divided evenly between LF-PORT (n = 59) and WN-PORT (n = 59). The WN-PORT group had significantly better 2-year OS (56.4% vs 78.1%; HR, 0.410; 95% CI, 0.223-0.751; P = .003), 2-year PFS (34.7% vs 59.8%; HR, 0.523; 95% CI, 0.320-0.856; P = .009), and 2-year LRC (54.4% vs 83.2%; HR, 0.279; 95% CI, 0.136-0.573; P < .001). The WN-PORT group had a higher incidence of G3 late toxicity for any site (6.8% vs 17%), bone (3.4% vs 6.8%), and pharynx and esophagus (3.4% vs 6.8%) in the matched-pair cohort. WN-PORT is a more appropriate choice than LF-PORT for improving LRC, PFS, and OS, although it is associated with an acceptable increase in late toxicities of Grade 3 or greater. Thus, we recommend WN-PORT especially for HNSCC patients with ECE or a close/positive margin.