基于深度学习的磁共振扫描鼻咽癌复发检测器:一项多中心研究
该多中心研究回顾性纳入三个癌症中心2007年4月至2020年7月间5035例治疗后鼻咽癌幸存者的5035次随访MR扫描,开发并评估深度学习模型MODERN用于检测复发。MODERN在内部验证集ROC-AUC为0.88(95%CI 0.86-0.90),外部验证集1为0.88(95%CI 0.86-0.90),外部验证集2为0.85(95%CI 0.82-0.88)。在220次扫描的读者研究中,MODERN单独准确率84.1%,结合MODERN评分后放射科医师准确率提升( competent 84.6%, expert 87.7%)。
要点
- MODERN在三个队列中检测鼻咽癌复发的ROC-AUC为0.85-0.88,外部验证性能稳定。
- MODERN单独准确率84.1%,与高年资放射科医师(85.9%)相当,高于 competent 医师(80.9%)。
- 提供MODERN评分后, competent 医师准确率从80.9%升至84.6%, expert 医师从85.9%升至87.7%。
核心结论
- 研究开发了用于鼻咽癌复发检测的深度学习模型MODERN,性能可靠;人机协作有潜力优化治疗后鼻咽癌患者随访MR扫描的判读流程。
研究设计与样本
- 回顾性多中心研究,从三个癌症中心收集2007年4月至2020年7月间5035例治疗后鼻咽癌幸存者的5035次随访MR扫描,用于开发和评估MODERN模型;另进行220次扫描的读者研究评估两名放射科医师在有/无MODERN时的检测准确率。
诊断标准与关键数值
- MODERN验证队列ROC-AUC:内部0.88(95%CI 0.86-0.90),外部1 0.88(95%CI 0.86-0.90),外部2 0.85(95%CI 0.82-0.88);读者研究中MODERN单独准确率84.1%(95%CI 79.3%-88.9%), competent 医师80.9%(95%CI 75.7%-86.1%,P<0.001 vs MODERN), expert 医师85.9%(95%CI 81.3%-90.5%,P<0.001 vs MODERN); competent 医师+MODERN评分84.6%(95%CI 79.8%-89.3%,P<0.001 vs 无MODERN), expert 医师+MODERN评分87.7%(95%CI 83.4%-92.1%,P<0.001 vs 无MODERN)。
临床应用与阅片要点
- MODERN可辅助放射科医师判读治疗后鼻咽癌随访MR扫描以检测复发,人机协作有潜力优化判读流程。
局限性
- 原文未报告局限性
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVES: Accuracy in the detection of recurrent nasopharyngeal carcinoma (NPC) on follow-up magnetic resonance (MR) scans needs to be improved.
MATERIAL AND METHODS: A total of 5 035 follow-up MR scans from 5 035 survivors with treated NPC between April 2007 and July 2020 were retrospectively collected from three cancer centers for developing and evaluating the deep learning (DL) model MODERN (MR-based Deep learning model for dEtecting Recurrent Nasopharyngeal carcinoma). In a reader study with 220 scans, the accuracy of two radiologists in detecting recurrence on scans with vs without MODERN was evaluated. The performance was measured using the area under the receiver operating characteristic curve (ROC-AUC) and accuracy with a 95% confidence interval (CI).
RESULTS: MODERN exhibited sound performance in the validation cohort (internal: ROC-AUC, 0.88, 95% CI, 0.86-0.90; external 1: ROC-AUC, 0.88, 95% CI, 0.86-0.90; external 2: ROC-AUC, 0.85, 95% CI, 0.82-0.88). In a reader study, MODERN alone achieved reliable accuracy compared to that of radiologists (MODERN: 84.1%, 95% CI, 79.3%-88.9%; competent: 80.9%, 95% CI, 75.7%-86.1%, P < 0.001; expert: 85.9%, 95% CI, 81.3%-90.5%, P < 0.001). The accuracy of radiologists was boosted by the MODERN score (competent with MODERN score: 84.6%, 95% CI, 79.8%-89.3%, P < 0.001; expert with MODERN score: 87.7%, 95% CI, 83.4%-92.1%, P < 0.001).
CONCLUSION: We developed a DL model for recurrence detection with reliable performance. Computer-human collaboration has the potential to refine the workflow in interpreting surveillant MR scans among patients with treated NPC.
引用
- Deng Y, Huang Y, Jing B, et al. Deep learning-based recurrence detector on magnetic resonance scans in nasopharyngeal carcinoma: A multicenter study. Eur J Radiol. 2023. PMID: 37722143. doi:10.1016/j.ejrad.2023.111084
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。