Risk and Predictors of Late Lower Cranial Neuropathy in Long-term Oropharyngeal Cancer Survivors Treated with Definitive Radiotherapy: A Retrospective Cohort Study among 1,988 Survivors

Abstract

Lower cranial neuropathy (LCNP) is a rare, but permanent, late effect of radiotherapy (RT) and other cancer therapies. The objective of this study was to quantify the cumulative incidence and identify clinical predictors of late LCNP. We hypothesized that 5-year incidence rate of late LCNP would approximate 5% and risk would correlate with tumor subsite and stage, RT dose, fractionation schedule, smoking status, and systemic therapy. This study included 1,988 OPC survivors (median survival: 6.85, range: 0.30-18.41 years) who received definitive RT at a single institution from 2000 to 2013. Late LCNP events defined by neuropathy of the glossopharyngeal (IX), vagus (X) and hypoglossal (XII) nerves ≥3-months after cancer therapy with normal baseline nerve function were abstracted from medical records. Cumulative incidence of LCNP was estimated using the Kaplan Meir method and multivariable Cox proportional hazard models were fit. About 4.4% (n = 87) of OPC survivors were diagnosed with late LCNP with median time to LCNP onset of 5.52 (range: 0.3-14.1) years post-treatment. Cumulative incidence of LCNP was 0.02 (95% CI: 0.02-0.03), 0.06 (95% CI: 0.05-0.08), and 0.10 (95% CI: 0.07-0.13) at 5, 10, and 15 years of follow-up, respectively. Multivariable Cox regression identified T4 vs T1 stage (HR: 3.90; 95%CI: 1.89-8.05, p<0.001) and accelerated vs standard RT fractionation (HR 2.20, 95%CI 1.37-3.53, p = 0.001) as independently associated with late LCNP status, adjusting for age, subsite, smoking, therapeutic modality, and solid food diet at baseline. Further, among all patients induction chemotherapy regimen as either TPF (docetaxel, cisplatin, and fluorouracil; HR: 2.37; 95%CI: 1.28-4.39; p = 0.006) or C-TPF (cetuximab, docetaxel, cisplatin, and fluorouracil; HR: 4.0; 95%CI: 1.22-13.15, p = 0.022) were identified in addition to T-stage (model with TPF; HR: 3.72, 95%CI: 1.81-7.65, p = <0.001; model with CTPF; HR: 3.97, 95%CI: 1.92-8.21, p = <0.001 ) and accelerated RT fractionation (model with TPF; HR: 2.56, 95%CI: 1.55-4.21, p = <0.001; model with CTPF; HR: 2.28, 95%CI: 1.41-3.68, p = 0.001) as significantly associated with late LCNP after adjustment. In our large cohort study, the risk of late LCNP progressed to exceed 10% cumulative risk over the lifetime of OPC survivors at 15 years follow-up. Our prediction model identified survivors with larger T4 tumors, those treated with more intensive treatment regimens including accelerated RT fractionation and induction TPF or CTPF as having a higher risk of late LCNP even after controlling for potential cofounders, suggesting that some LCNP risk may be modifiable. Further efforts are necessary to investigate other associations and effective risk reduction strategies for this disabling late effect experienced by growing numbers of long-term OPC survivors.