机器人辅助人工耳蜗电极植入的学习曲线及评估
Learning Curve and Evaluation of Robotic Assistance for Cochlear Implant Electrode Insertion.
文献信息
| PMID | 42759040 |
|---|---|
| 原文 | 在 PubMed 查看原文 ↗ |
| 发表日期 | 2026 |
| 作者 | Sarah Draut |
| 作者单位 | Department of Otolaryngology-Head and Neck Surgery, Ludwig-Maximilians-University, Munich, Germany. |
| 期刊 | Otology & neurotology : official publication of the American Otological Society, American Neurotology Society [and] European Academy of Otology and Neurotology |
| SCI 分区 | Q2 |
| IF | 1.9 |
| 研究类型 | 临床研究 · 临床 |
| 所属专科 | 耳科 |
中文摘要
目的： 机器人辅助已被引入以提高人工耳蜗电极植入的精确性并减少耳蜗内创伤。本研究评估与机器人辅助系统相关的学习曲线，并评估早期临床实施期间的客观和主观性能指标。
研究设计： 前瞻性观察性队列研究。
地点： 三级转诊中心。
参与者： 三名具有不同机器人辅助人工耳蜗植入经验水平的外科医生。
干预： 在结构化培训项目后进行机器人辅助人工耳蜗电极植入。
主要结局指标： 机器人使用时间、准备时间、植入时间、总手术时间，以及使用SURG-TLX问卷评估的术中工作负荷。学习曲线通过序贯组间比较、CUSUM分析和非线性指数衰减建模进行评估。模型拟合与线性替代模型进行比较。
结果： 分析了60例连续机器人辅助植入。机器人使用时间从19.1分钟（第1-20例）降至15.0分钟（第41-60例）（P=0.020）。准备时间从12.5分钟降至9.2分钟（P=0.012），最大改善出现在前20例。指数衰减建模显示非线性学习模式，估计机器人使用时间渐近值为16.3分钟（τ=1.65例），基于AIC/BIC模型比较优于线性回归。植入时间和总手术时间在各组之间无显著差异。CUSUM分析提示机器人使用时间在20-30例内达到组水平稳定。SURG-TLX显示第21-40例与第41-60例之间情境压力显著降低（P=0.031）。
结论： 机器人辅助人工耳蜗电极植入显示出明显的学习曲线，主要由机器人设置和操作的调整驱动。手术效率迅速提高，在个体水平早期稳定，并在所有外科医生的汇总水平后期稳定。结构化培训和流程化工作流程支持安全的临床采用。
英文摘要
OBJECTIVE: Robotic assistance has been introduced to improve precision and reduce intracochlear trauma during cochlear implant electrode insertion. This study evaluates the learning curve associated with robotic assistance systems and assesses objective and subjective performance metrics during early clinical implementation.
STUDY DESIGN: Prospective observational cohort study.
SETTING: Tertiary referral center.
PARTICIPANTS: Three surgeons with varying levels of experience performing robotic-assisted cochlear implantation.
INTERVENTIONS: Robotic-assisted cochlear implant electrode insertion after a structured training program.
MAIN OUTCOME MEASURES: Robot usage time, preparation time, insertion time, total operation time, and intraoperative workload assessed using the SURG-TLX questionnaire. Learning curves were evaluated using sequential group comparisons, CUSUM analysis, and nonlinear exponential decay modelling. Model fit was compared with linear alternatives.
RESULTS: Sixty consecutive robotic-assisted insertions were analyzed. Robot usage time decreased from 19.1 minutes (cases 1-20) to 15.0 minutes (cases 41-60) (P=0.020). Preparation time decreased from 12.5 to 9.2 minutes (P=0.012), with the largest improvement in the first 20 cases. Exponential decay modeling demonstrated a nonlinear learning pattern with an estimated asymptotic robot usage time of 16.3 minutes (τ=1.65 procedures), outperforming linear regression based on AIC/BIC model comparison. Insertion time and total operation time did not differ significantly between groups. CUSUM analysis suggested group-level stabilization of robot usage time within 20-30 cases. SURG-TLX showed a significant reduction in situational stress between cases 21-40 and 41-60 (P=0.031).
CONCLUSION: Robot-assisted cochlear implant electrode insertion shows a distinct learning curve primarily driven by adjustment in robotic setup and handling. Procedural efficiency improves rapidly, with early stabilization at the individual level and later stabilization at the aggregated level over all surgeons. Structured training and protocolized workflow support safe clinical adoption.