Surveillance Imaging with PET/CT and CT and/or MRI for Head and Neck Cancer and Mortality: A Population-based Study.

Background To the knowledge of the authors, no strong evidence supports surveillance imaging in patients with head and neck cancer (HNC). Purpose To investigate the association between surveillance imaging and mortality using a population-based study design with statewide cancer registry data, all-payer claims data, and health care facility data. Materials and Methods The retrospective population-based study identified patients with HNC diagnosed between January 2012 and December 2017. Current Procedural Terminology codes were used to search surveillance imaging procedures. Cox proportional hazards models were used to estimate hazard ratios (HRs) and 95% CIs for mortality with adjustment for sex, ethnicity, age, health insurance status, cancer site, stage, and treatment. Results The study identified 1004 patients (mean age, 61 years ± 12 [SD]; 753 men), including 902 patients with squamous cell carcinoma (SCC) HNC and 102 patients with non-SCC. The effect of imaging on mortality among patients with SCC was not statistically significant when the entire sample was analyzed (HR, 0.76; 95% CI: 0.57, 1.02; P = .07). However, in stratified analyses by cancer stage, surveillance imaging was associated with lower mortality among patients with SCC for regionalized cancer stage (HR, 0.55; 95% CI: 0.36, 0.83; P = .005) and distant cancer stage (HR, 0.40; 95% CI: 0.19, 0.83; P = .01). Among patients with non-SCC, surveillance imaging was associated with lower mortality versus no surveillance imaging (HR, 0.19; 95% CI: 0.04, 0.94; P = .04). PET/CT was associated with lower mortality for patients with SCC (HR, 0.29; 95% CI: 0.09, 0.94; P = .04), and CT and/or MRI was associated with lower mortality for patients with non-SCC (HR, 0.11; 95% CI: 0.01, 0.94; P = .04). Conclusion Surveillance imaging was associated with lower mortality among patients with head and neck squamous cell carcinoma with regionalized or distant disease. The surveillance imaging protective association was observed up to 2 years after treatment completion. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Branstetter in this issue.