Patterns of Care and Survival of Patients With pT1-2N1M0 Head and Neck Cancers Receiving Postoperative Radiation or Chemoradiation

Abstract

To analyze the patterns of care and survival for pT1-2N1M0 head and neck cancer based on receipt of surgery alone (S), postoperative radiation (S+RT) or postoperative chemoradiation (S+CRT). The National Cancer Database was explored to identify patients diagnosed with head and neck squamous cell carcinoma of the oral cavity, oropharynx, hypopharynx, or larynx between 2004-2012. Patients were included if they underwent surgery (>local excision) and were staged as pT1-2N1M0 with negative margins and no extracapsular extension (ECE). Those who were identified as receiving chemotherapy and radiation therapy within 14 days of each other were included as S+CRT. Univariable and multivariable logistic regression was used to assess for predictors of S+RT compared to S alone and separate multivariable logistic regression was used to assess for predictors of S+CRT compared to S+RT. Univariable and multivariable Cox Regression was similarly performed to assess for covariates that had an impact on survival. Propensity matching was performed and the multivariable Cox Regression was repeated utilizing the propensity matched sample. There were 2,099 patients included in this study, from which 852 (40.6%) received surgery alone, 864 (41.2%) received postoperative radiation alone, and 383 (18.2%) received postoperative chemoradiation. Median follow up for living patients was 47.1 months. The 5-year overall survival was 61.8% for surgery alone, 70.7% for surgery + postoperative radiation, and 85.7% for surgery + postoperative chemoradiation (p<0.001 for each comparison on pairwise analysis). On multivariable logistic regression, academic centers were associated with a decreased likelihood of S+RT (OR 0.69) and S+CRT (OR 0.70). On multivariable Cox Regression, S+RT was associated with improved survival over S alone (HR 0.69, 95% CI 0.58-0.82, p=0.006). Propensity score matching identified a cohort of 1,250 patients (625 received S and 625 received S+RT). Multivariable Cox Regression on the matched sample revealed a persistent improvement in survival for postoperative radiation (HR 0.79, 95% CI 0.64-0.98, p<0.001). There was also a survival benefit favoring S+CRT over S+RT (HR 0.57, 95% CI 0.42-0.78, p<0.001). Propensity score matching identified a cohort of 604 patients (302 received S+RT and 302 received S+CRT). Multivariable Cox regression of the matched sample revealed a persistent survival benefit associated with S+CRT (HR 0.56, 95% CI 0.39-0.82, p=0.003). Nearly 60% of patients with pT1-2N1 head and neck cancer with negative margins and no ECE received S+RT or S+CRT. Patients treated in academic centers are more likely to receive S alone. There was a survival benefit noted for S+RT over S alone as well as S+CRT over S+RT on propensity matched analysis, though caution is still necessary interpreting these results given the inability for propensity matching to account for all causes of bias.