Article record

我们是否对前庭神经鞘瘤监测过度？质疑追踪所有肿瘤的必要性。

Are We Overmonitoring Vestibular Schwannomas? Challenging the Necessity of Tracking Them All.

临床研究耳科IF 2.9Q1

文献信息

PMID	42808556
原文	在 PubMed 查看原文 ↗
发表日期	2026
作者	Aïna Venkatasamy 查看其他 4 位作者 Anne R J Péporté；Romain Kania；Jean-Yves Tanguy；Benoît Gallix
作者单位	Departement of Radiology, Hôpital Européen Georges Pompidou, APHP, Paris, France. 查看其他 10 条单位信息 In vivo Imaging Research lab, UMR-S970, PARCC, Université Paris Cité, Paris, France.；Department of Radiology Medical Physics, Faculty of Medicine, Medical Center-University of Freiburg, University of Freiburg, Killianstrasse 5a, Freiburg, Germany.；Department of Radiology, Cantonal Hospital Frauenfeld, Frauenfeld, Switzerland.；Department of Otorhinolaryngology Skull Base Center, Assistance Publique Hôpitaux de Paris, France.；Faculté de Médecine, Université Paris Cité, Paris, France.；UMR 1141, Université Paris Cité, Paris, France.；Department of Radiology, Centre Hospitalier Universitaire d'Angers, Angers, France.；Department of Radiology, American Hospital of Paris, Neuilly-sur-Seine, France.；Faculté de médecine, Université de Strasbourg, Strasbourg, France.；Department of Radiology, McGill University, Montreal, Canada.
期刊	Otolaryngology--head and neck surgery : official journal of American Academy of Otolaryngology-Head and Neck Surgery
SCI 分区	Q1
IF	2.9
研究类型	临床研究 · 临床
所属专科	耳科

中文摘要

目的： 因非特异性症状而增加的MRI使用导致小或偶然发现的前庭神经鞘瘤（VS）频繁检出，其中大多数从未需要治疗。本研究旨在开发并验证一种逐步模型，利用早期MRI发现对VS患者进行分层，并减少不必要的长期监测。
研究设计： 回顾性多中心队列研究，并进行外部验证。
地点： 法国（斯特拉斯堡、昂热）和瑞士（弗劳恩费尔德）的三个三级转诊中心。
方法： 纳入年龄≥18岁、散发性VS且在任何干预前有≥3次连续MRI的患者（2000-2020年）。既往治疗、迷路内神经鞘瘤或影像不完整者被排除。分析了训练队列（n=156）和独立验证队列（n=32）。使用逻辑回归和随机森林模型识别治疗预测因素。开发了一种整合基线肿瘤体积、年增长率和症状存在的临床适用逐步算法。
结果： 优化模型在两个队列中均达到100%敏感性，确保未遗漏任何需要治疗的患者。训练队列的特异性为74%，验证队列为48%。在1次随访MRI后，一半队列（分别为74%和48%的患者）可以安全停止监测。较大的基线肿瘤体积、症状存在和较高的增长率是治疗的显著预测因素（P<0.05）。
结论： 分层逐步方法能够安全减少约一半VS患者的长期MRI监测，支持个性化随访和更高效的医疗资源利用。

英文摘要

OBJECTIVE: Increased MRI use for nonspecific symptoms has led to frequent detection of small or incidental vestibular schwannomas (VS), most of which never require treatment. This study aimed to develop and validate a stepwise model to stratify VS patients using early MRI findings and reduce unnecessary long-term surveillance.
STUDY DESIGN: Retrospective multicenter cohort study with external validation.
SETTING: Three tertiary referral centers in France (Strasbourg, Angers) and Switzerland (Frauenfeld).
METHODS: Patients aged ≥18 years with sporadic VS and ≥3 consecutive MRIs before any intervention were included (2000-2020). Prior treatment, intralabyrinthine schwannomas, or incomplete imaging led to exclusion. A training cohort (n = 156) and an independent validation cohort (n = 32) were analyzed. Logistic regression and random forest models were used to identify predictors of treatment. A clinically applicable stepwise algorithm integrating baseline tumor volume, annual growth rate, and symptom presence was developed.
RESULTS: The optimized model achieved 100% sensitivity in both cohorts, ensuring no patient requiring treatment was missed. Specificity was 74% in the training cohort and 48% in the validation cohort. After 1 follow-up MRI, surveillance could be safely discontinued in half of the cohorts (respectively 74% and 48% of patients). Larger baseline tumor volume, symptom presence, and higher growth rate were significant predictors of treatment (P < .05).
CONCLUSION: A stratified, stepwise approach enables safe reduction of long-term MRI surveillance in approximately half of VS patients, supporting personalized follow-up and more efficient healthcare use.