Impact of distance to treatment center on tumor stage of nasopharyngeal cancer at presentation.

Nasopharyngeal carcinoma (NPC) is a relatively rare cancer that accounts for only 0.7% of new cancer cases worldwide, but it is unique in its disproportionate prevalence among East and Southeast Asian patients.1, 2 Although advances in chemoradiation have reduced NPC incidence and mortality,3 gains in screening have been limited due to NPC's nonspecific symptoms.4 Since screening for NPC requires endoscopic evaluation and/or magnetic resonance imaging/computed tomography imaging, travel distance to a treatment center may impact time to patient presentation. In patients with laryngeal cancer, living farther from a primary treatment center is associated with a higher stage at presentation, but this relationship has not been evaluated among patients with NPC.5 Thus, we aimed to analyze how distance to an accredited primary cancer treatment center impacts NPC tumor stage at presentation using data from the National Cancer Database (NCDB). In this study, we analyzed the NCDB, a nationwide oncologic database with data from over 1500 Commission on Cancer–accredited facilities. Patients diagnosed with primary NPC from 2004 to 2018 were included. Patients were excluded if they had unconfirmed, benign, borderline, or in-situ cases, nonsquamous cell carcinoma antigen histology, or unidentifiable primary tumors. Patients without travel or survival data or who traveled > 250 miles were also excluded. Patient demographics and tumor characteristics were collected. Patients were classified as living in a rural, urban, or city environment based on Rural-Urban Continuum Codes, which use population size and measures of urbanization to classify metropolitan areas. Travel distance, represented in the NCDB as the number of miles between a patient's home and hospital zip code, was stratified into short (< 12.5 miles), intermediate (12.5–49.9 miles), and long (50–250 miles) in accordance with previous studies.5 Patient and tumor characteristics were summarized and compared between distance groups. Ordered multivariable logistic regression analysis was performed to identify predictors of advanced stage disease (defined as T4 disease). Multivariate Cox regression analysis was performed to determine whether travel distance, tumor characteristics, or patient characteristics were predictors of survival. Analyses were conducted using STATA version 17.0 (StataCorp LLC). Statistical significance was defined as p < 0.05 and adjusted for multiple comparisons using the Simes false discovery rate. This study was exempt from institutional review board approval due to the publicly available and deidentified nature of NCDB. Our cohort consisted of 5652 patients with histologically confirmed NPC squamous cell carcinoma. Baseline characteristics of the cohort are outlined in Table 1. A total of 71% of patients were men, 72% identified as White , 41% had private insurance, 32% presented at stage T4 disease, and 37% were treated at academic or research programs. As shown in Table 2, those living in an urban setting were less likely to present at an advanced T stage. Patients with Medicaid, no insurance, or Medicare were more likely to present with T4 disease. <12.5 miles 12.5–49.9 miles 50–250 miles When controlling for patient characteristics, urban status, and insurance status, travel distance was associated with higher likelihood of T4 disease at presentation. Patients were more likely to present with T4 staging if they traveled between 50 and 250 miles to reach their treatment facility. Advanced-stage disease was a predictor of increased mortality (hazard ratio [HR], 2.15 [95% confidence interval (CI), 1.95–2.38], p < 0.001). Long travel distances were associated with an increased chance of overall survival (HR, 0.85 [95% CI, 0.75–0.98], p < 0.05) after adjustment for stage of disease, with no significant association between short and intermediate distances. There was no association between overall TNM (tumor, node, metastasis) staging and distance traveled. In this study, we reviewed the impact of distance to treatment center on tumor stage in patients with NPC at presentation. We found that increased travel distance was associated with a higher T stage at presentation. Patients with no health insurance, Medicaid, or Medicare were also more likely to present at an advanced T stage compared with patients with private health insurance. Our findings associating travel distance with stage of presentation are consistent with similar studies for patients with laryngeal, breast, lung, and colon cancers.5, 6 It reinforces the importance of geographically accessible care in early diagnosis and treatment, and our findings suggest a need to improve screening rates and access in areas with less available care. Patients may face several obstacles when treatment centers are far away. These include access to reliable transportation, taking time off work, access to childcare, and the expenses of traveling. While increasing the number of treatment centers and satellite clinics may help reduce geographic disparities, physician offices could offer more immediate alternative options to address these barriers such as telemedicine visits or travel vouchers as local treatment centers are developed. Our findings that patients who were uninsured or on Medicaid or Medicare were more likely to present with advanced T-stage disease is consistent with previous studies demonstrating decreased access to care among certain populations.7 There may be a need to increase screening rates among uninsured, low-income, and elderly patients as these groups likely have more difficulty in accessing cancer treatment facilities. Increasing health insurance access to specialist care as well as promoting awareness of these findings among specialists could improve screening rates as a nasal endoscopy is unlikely to be performed by a primary care physician and patients usually need to be examined by an otolaryngologist for definitive diagnosis of NPC. Our study is not without limitations. As a database study, we can only establish associations between variables. Furthermore, NCDB data are not wholly representative of the US population and thus may be skewed. Future directions include prospectively measuring the impact of travel distance on tumor stage at presentation and investigating social interventions that improve NPC management and survival. The authors declare no conflicts of interest.