A Systematic Review and Meta-Analysis of Toxicity in Oropharyngeal Cancer Patients Treated with Proton Therapy

Abstract

Proton radiotherapy (PRT) is an emerging treatment modality for patients with oropharyngeal cancer (OPC). Due to its dosimetric properties, PRT is hypothesized to be associated with lower rates of acute and late toxicity than photon intensity modulated radiotherapy (IMRT). Current evidence supporting PRT in this setting is limited to small, single institution reports. We sought to examine pooled rates of PRT toxicity from the current literature. A systematic review of PubMed, Embase, and Web of Science was performed. Articles published from 1980-2022 that contained >10 patients treated with definitive or adjuvant PRT for OPC were included. From these studies acute and late clinician rated adverse events (CRAEs) were extracted. The primary outcomes were the pooled rates of grade 3 or higher (G3+) acute CRAEs. The secondary outcomes were the pooled rates of late CRAEs. Pooled rates were estimated using random effects models for CRAEs that were reported in ≥3 studies and graded using Common Terminology Criteria for Adverse Events (CTCAE). In studies that reported both PRT and IMRT, a mixed effects log odds ratios (ORs) was used to compare risk of CRAE between modalities. A total of 8 studies (5 retrospective, 3 prospective) with 291 patients treated with PRT for OPC were identified. The majority of patients were men (90%) with HPV-associated disease (93%) treated with definitive intent (56%). Primary tumor was most frequently tonsil (55%); 58% had T1-2, and 60% had N2-3 disease (AJCC7). Estimated pooled rates of G3+ acute CRAEs were as follows: dermatitis: 22%, mucositis: 38%, xerostomia: 1.6%, dysphagia: 15%, and weight loss: 1.9%. Estimated pooled rates of G2+ acute CRAEs were as follows: dermatitis: 73%, xerostomia: 12%, weight loss: 14%, and dysgeusia: 29%. Rate of acute hospitalization was 10%. In terms of late toxicities of PRT, the pooled rate of G2+and G3+ xerostomia were 22.5% and 1.2%, respectfully, while the rates G2+ and G3+ dysphagia were 15.6% and 2.6%, respectfully. Among the included studies, 4 studies reported CRAEs from patients treated with IMRT. Due to inconsistent reporting of CRAEs, only use of feeding tube (FT) could be compared between treatment modalities. PRT was associated with significantly lower FT use in the acute setting compared to IMRT - 18% versus 28%, respectively (log OR -0.88, P<0.001). Long term FT use was not significantly different between PRT and IMRT - 1.4% versus 2.7%, respectively (log OR -0.95, P = 0.24). In the largest pooled analysis of PRT for OPC to date, PRT was associated with a 3-fold reduction of acute, but not late, FT use compared to IMRT. The pooled rates of other PBT CRAEs appear to be at least similar to, if not less than, those reported on IMRT trials treating similarly selected patients. Ultimately, heterogeneity in reporting PRT toxicity outcomes greatly reduces the interpretability of these data and limits proton-photon comparisons. Maximizing consistency of CRAE reporting in future studies is critical.