人工智能辅助双参数MRI检测前列腺癌病灶的多中心全交叉多阅片者多病例试验
该多中心全交叉多阅片者多病例试验纳入407例男性患者的双参数MRI,比较10名非专家放射科医生在无AI辅助、有AI辅助及AI单独三种阅片模式下的病灶检测性能。AI辅助较无辅助显著提高病灶水平敏感度(67.3%升至85.5%),病例水平特异度非劣(75.9%升至79.5%),AFROC-AUC更高(86.9%对76.1%)。AI单独也表现出较高诊断准确性(AFROC-AUC 83.1%)。
要点
- AI辅助使病灶水平敏感度从67.3%提升至85.5%,提高18.2%(95%CI 10.7%-25.7%,p<0.001)。
- AI辅助病例水平特异度79.5%对无辅助75.9%,达到非劣效(p<0.001)。
- AI单独病灶水平敏感度88.4%、病例水平特异度77.8%、AFROC-AUC 83.1%。
核心结论
- 与无辅助阅片相比,AI辅助阅片显著提高病灶检测敏感度,且AI单独也表现出较高诊断准确性。
研究设计与样本
- 多中心全交叉多阅片者多病例临床试验,3家机构,10名非专家放射科医生,回顾性收集双参数MRI病例(T2WI、扩散加权成像、表观扩散系数),参考标准由资深放射科医生根据病理报告建立。
关键影像表现/征象
- AI辅助较无辅助显著提高病灶水平敏感度:85.5%(95%CI 81.3%-89.7%)对67.3%(95%CI 58.8%-75.8%),差值18.2%(95%CI 10.7%-25.7%,p<0.001)。
- AI辅助病例水平特异度79.5%(95%CI 74.1%-84.8%)对无辅助75.9%(95%CI 68.7%-83.1%),非劣效(p<0.001)。
- AI辅助AFROC-AUC 86.9%高于无辅助76.1%(p<0.001)。
- AI单独AFROC-AUC 83.1%(95%CI 79.7%-86.6%),病灶水平敏感度88.4%(95%CI 84.0%-92.0%),病例水平特异度77.8%(95%CI 71.5%-83.3%)。
- 亚组分析显示AI辅助对较小病灶和较低前列腺影像报告和数据系统评分的病灶检测改善更明显。
诊断标准与关键数值
- 以资深放射科医生基于病理报告建立的参考标准为金标准。
- 病灶水平敏感度:无辅助67.3%(95%CI 58.8%-75.8%)对AI辅助85.5%(95%CI 81.3%-89.7%),差值18.2%(95%CI 10.7%-25.7%,p<0.001)。
- 病例水平特异度:无辅助75.9%(95%CI 68.7%-83.1%)对AI辅助79.5%(95%CI 74.1%-84.8%),非劣效(p<0.001)。
- AFROC-AUC:无辅助76.1%对AI辅助86.9%(p<0.001)。
- AI单独:AFROC-AUC 83.1%(95%CI 79.7%-86.6%),病灶水平敏感度88.4%(95%CI 84.0%-92.0%),病例水平特异度77.8%(95%CI 71.5%-83.3%)。
临床应用与阅片要点
- 非专家放射科医生在双参数MRI前列腺癌病灶检测中,AI辅助可提高病灶水平敏感度,同时病例水平特异度非劣于无辅助。
- AI单独阅片也表现出较高诊断准确性,但原文未据此给出独立应用的推荐。
- 亚组分析提示AI辅助对较小病灶和较低前列腺影像报告和数据系统评分病灶的检测改善更明显。
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
RATIONALE AND OBJECTIVES: To assess artificial intelligence (AI)'s added value in detecting prostate cancer lesions on MRI by comparing radiologists' performance with and without AI assistance.
MATERIALS AND METHODS: A fully-crossed multi-reader multi-case clinical trial was conducted across three institutions with 10 non-expert radiologists. Biparametric MRI cases comprising T2WI, diffusion-weighted images, and apparent diffusion coefficient were retrospectively collected. Three reading modes were evaluated: AI alone, radiologists alone (unaided), and radiologists with AI (aided). Aided and unaided readings were compared using the Dorfman-Berbaum-Metz method. Reference standards were established by senior radiologists based on pathological reports. Performance was quantified via sensitivity, specificity, and area under the alternative free-response receiver operating characteristic curve (AFROC-AUC).
RESULTS: Among 407 eligible male patients (69.5±9.3years), aided reading significantly improved lesion-level sensitivity from 67.3% (95% confidence intervals [CI]: 58.8%, 75.8%) to 85.5% (95% CI: 81.3%, 89.7%), with a substantial difference of 18.2% (95% CI: 10.7%, 25.7%, p<0.001). Case-level specificity increased from 75.9% (95% CI: 68.7%, 83.1%) to 79.5% (95% CI: 74.1%, 84.8%), demonstrating non-inferiority (p<0.001). AFROC-AUC was also higher for aided than unaided reading (86.9% vs 76.1%, p<0.001). AI alone achieved robust performance (AFROC-AUC=83.1%, 95%CI: 79.7%, 86.6%), with lesion-level sensitivity of 88.4% (95% CI: 84.0%, 92.0%) and case-level specificity of 77.8% (95% CI: 71.5%, 83.3%). Subgroup analysis revealed improved detection for lesions with smaller size and lower prostate imaging reporting and data system scores.
CONCLUSION: AI-aided reading significantly enhances lesion detection compared to unaided reading, while AI alone also demonstrates high diagnostic accuracy.
引用
- Xing Z, Chen J, Pan L, et al. Enhanced Detection of Prostate Cancer Lesions on Biparametric MRI Using Artificial Intelligence: A Multicenter, Fully-crossed, Multi-reader Multi-case Trial. Acad Radiol. 2025. PMID: 40651921. doi:10.1016/j.acra.2025.06.038
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。