头颈癌中顺铂诱导的耳毒性：对当前证据、临床意义及标准化监测需求的系统综述

文献信息

中文摘要

背景： 以顺铂为基础的放化疗仍是局部晚期头颈部鳞状细胞癌（HNSCC）的标准治疗，但顺铂诱导的耳毒性常见、通常不可逆且监测不一致。方法：检索MEDLINE、EMBASE和Cochrane Library中2016年1月至2026年1月发表的成人英文临床研究。报告至少一项听力、耳鸣、前庭或耳毒性结局的原始全文临床研究符合纳入条件；仅会议摘要、重复报告、综述和共识文件不予纳入。从原始数据库途径保留24项研究。2026年9月16日进行的针对性补充PubMed验证采用更广泛的听觉和前庭术语，识别出另外6项符合条件的全文研究，通过PRISMA 2020“其他方法”途径纳入，共30项研究。由单一综述作者使用RoB 2、ROBINS-I和Joanna Briggs Institute（JBI）工具对所有30项研究进行偏倚风险评估；未进行重复独立评估，并作为局限性报告。原计划进行荟萃分析，但因听力测量频率、分级系统、随访间隔和效应量不可比而未实施；这作为方案偏离报告。结果：30项研究共约10,500名参与者，但两项报告使用了重叠的匈牙利患者样本，且若干研究纳入了非HNSCC人群。报告的耳毒性范围为13%（每周顺铂联合耳蜗保护性放疗后严重、不可逆听力损失）至72%（生存者门诊筛查发现的听力损失），反映的是定义不同而非单一潜在发生率。比较性HNSCC队列一致报告，每三周100 mg/m2方案的听力损失多于每周40 mg/m2方案。30项研究中有10项为高、严重或极严重偏倚风险，另有一项比较性研究为中度至严重风险；长期随机比较报告每周30 mg/m2方案的局部区域控制较差。基线听力、耳蜗辐射暴露、监测依从性和药物遗传易感性是进一步的决定因素。在无眩晕的情况下记录到客观前庭功能障碍。成人中尚无已确立的药物性耳保护策略。结论：顺铂耳毒性是核心的生存期终点。每周给药可能在特定情况下减少听力毒性，但不得解读为普遍肿瘤学等效。标准化定义、基线和纵向检测、耳蜗剂量报告以及前庭结局和耳保护剂的前瞻性评估是优先事项。

英文摘要

Background: Cisplatin-based chemoradiotherapy remains a standard treatment for locally advanced head and neck squamous cell carcinoma (HNSCC), but cisplatin-induced ototoxicity is common, usually irreversible and inconsistently monitored. Methods: MEDLINE, EMBASE and the Cochrane Library were searched for English-language clinical studies in adults published between January 2016 and January 2026. Primary full-text clinical studies reporting at least one hearing, tinnitus, vestibular or ototoxicity outcome were eligible; conference-only abstracts, duplicate reports, reviews and consensus papers were not. Twenty-four studies were retained from the original database route. A targeted supplementary PubMed verification performed on 16 September 2026 with broader auditory and vestibular terminology identified six additional eligible full-text studies, which were included through the PRISMA 2020 'other methods' pathway, for a total of 30 studies. Risk of bias was appraised for all 30 studies with RoB 2, ROBINS-I and Joanna Briggs Institute (JBI) tools by a single review author; duplicate independent appraisal was not performed and is reported as a limitation. Meta-analysis had been planned but was not undertaken because audiometric frequencies, grading systems, follow-up intervals and effect measures were not comparable; this is reported as a protocol deviation. Results: The 30 studies comprised approximately 10,500 participants, although two reports used overlapping Hungarian patient samples and several studies included non-HNSCC populations. Reported ototoxicity ranged from 13% (severe, irreversible loss after weekly cisplatin with cochlear-sparing radiotherapy) to 72% (screen-detected loss in a survivorship clinic), reflecting different definitions rather than a single underlying rate. Comparative HNSCC cohorts consistently reported more hearing loss with 100 mg/m2 every three weeks than with weekly 40 mg/m2. Ten of the 30 studies were at high, serious or critical risk of bias, with one additional comparative study at moderate-to-serious risk; the long-term randomized comparison reported inferior locoregional control with weekly 30 mg/m2. Baseline hearing, cochlear radiation exposure, monitoring adherence and pharmacogenetic susceptibility were further determinants. Objective vestibular dysfunction was documented in the absence of dizziness. No pharmacological otoprotective strategy is established in adults. Conclusions: Cisplatin ototoxicity is a core survivorship endpoint. Weekly dosing may reduce hearing toxicity in selected settings but must not be read as universal oncological equivalence. Standardized definitions, baseline and longitudinal testing, cochlear-dose reporting and prospective evaluation of vestibular outcomes and otoprotective agents are the priorities.