Unplanned Interruptions during Head and Neck Radiotherapy prior to and during the COVID-19 Pandemic

PURPOSE/OBJECTIVE(S): Patients with head and neck (HN) cancers receiving radiotherapy (RT) often require complex coordination between providers and may be at increased risk for treatment interruptions during the ongoing pandemic. We sought to determine if patients with HN cancers receiving RT at our academic center experienced higher rates of RT interruption – defined as any unscheduled miss of fractionated treatment, except for holidays – during the initial 6 months after the March 13(th), 2020 pandemic declaration compared to the prior 6-month interval. We also sought to determine predictors associated with interruption risk. MATERIALS/METHODS: All patients receiving RT for HN cancers with first fraction between 9/2019 and 9/2020 were identified from institutional electronic medical records. Candidate predictive factors including demographics and RT details (dose and fractionation) were abstracted. Primary residence zip code was used to obtain median household income. Descriptive statistics were used to compare outcomes regarding unplanned RT interruptions pre (9-2019- 3/12/2020) and during COVID (3/13/2020 -9/2020). Multinomial logistic regression was used to identify factors associated with highest quartile of unplanned RT interruptions. Spearman's rho was used to assess associations between factors and number of unplanned interruptions. RESULTS: 119 patients were included (oropharynx (n=33), oral cavity (n=31), larynx (n=15), and other including salivary gland, nasopharynx, and thyroid (n= 40)). The cohorts pre and during COVID (N= 55 pre-COVID, 64 during COVID) were balanced. 87 patients (73.1%) had at least one unplanned RT interruption (n=38, 69.1% pre-COVID vs n=49, 76.6% post-COVID). There was not a difference in presence of any (yes/no) unplanned RT interruptions (p =.36) or total number of unplanned RT interruptions between pre vs post COVID subgroups (p=.85) by univariable analysis. The lowest quartile of patients had 0 interruptions, the 2nd quartile had 1 interruption, the 3(rd) quartile had 2-4 interruptions, and the highest quartile had > 4 interruptions. No demographic factors, nor RT dose or fraction number were associated with risk of being in the highest quartile of unplanned interruptions. No significant association between pre vs post COVID-19 timing and number of unplanned interruptions was found by Spearman's rho. CONCLUSION: No significant differences in presence or number of unplanned RT interruptions were detected at our institution during the pandemic. Although these results suggest some degree of health system-level resiliency during COVID-19, they unfortunately also confirm persistent shortfalls handicapping reliable on-treatment access to RT. Independent corroborative data will be necessary to conclusively answer whether HN patients are more broadly experiencing heightened barriers to RT during the public health emergency.