双参数MRI人工智能风险分级检测临床有意义前列腺癌的多中心评估
回顾性多中心多厂家研究纳入787例接受前列腺MRI并有病理证实的男性,比较商用AI风险分级系统与PI-RADS对临床有意义前列腺癌(csPCa)的检出效能。AI在患者层面敏感度不劣于PI-RADS且AUC更高,但病灶层面敏感度较低。AI最宜用于活检分流决策支持,尤其是PI-RADS 3病灶。
要点
- AI患者层面敏感度97.6%不劣于PI-RADS的92.6%,可用于高敏感度活检分流。
- PI-RADS 3病灶中AI升级93.1%的csPCa阳性病灶,有助于减少漏诊。
- AI病灶层面敏感度较低,病灶定位仍需放射科医生负责。
核心结论
- AI系统在患者层面敏感度不劣于PI-RADS,但在病灶层面敏感度不劣效性未达到。
- AI风险分级最适用于活检分流决策支持,尤其是对PI-RADS 3病灶的辅助判断。
研究设计与样本
- 回顾性多中心多厂家诊断准确性研究,纳入2014年至2025年间接受前列腺MRI并获得组织病理学证实的男性787例。
- 参考标准为国际泌尿病理学会分级分组≥2级,即临床有意义前列腺癌。
关键影像表现/征象
- 针对PI-RADS 3病灶的策略避免了18.9%的活检,同时维持98.4%的敏感度。
- 在PI-RADS 3病灶中,AI将93.1%的csPCa阳性病灶升级为高风险。
- 在PI-RADS 3病灶中,AI将24.1%的良性病灶判定为低风险。
诊断标准与关键数值
- 患者层面AI敏感度为97.6%,PI-RADS敏感度为92.6%,非劣效性检验p<0.001。
- 病灶层面AI敏感度为78.8%,PI-RADS敏感度为88.8%,非劣效性检验p=0.98。
- 患者层面AUC:AI为0.80,PI-RADS为0.77,差值0.032,95%CI为0.001至0.064,p=0.043。
- 787例患者中380例(48.3%)有临床有意义前列腺癌,中位年龄70岁,四分位距64至75岁。
临床应用与阅片要点
- AI风险分级可作为PI-RADS的补充,用于高敏感度活检分流,但不能替代放射科医生的病灶定位。
- 对PI-RADS 3病灶可结合AI风险分级决定是否需要进一步检查或活检。
- 采用定向PI-RADS 3策略可减少近五分之一活检且保持高敏感度。
局限性
- 病灶层面AI敏感度在预设阈值下低于PI-RADS,非劣效性未达到。
- 研究为回顾性设计。
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: To compare a commercially available artificial intelligence (AI) risk-classification system with PI-RADS for detecting clinically significant prostate cancer (csPCa) on prostate MRI.
MATERIALS AND METHODS: This retrospective multicentre, multivendor diagnostic accuracy study included men who underwent prostate MRI and histopathologic verification between 2014 and 2025. Index tests were DeepHealth Prostate Suite AI risk category and radiologist-assigned PI-RADS; the reference standard was International Society of Urological Pathology Grade Group ≥ 2. Patient-level non-inferiority testing, receiver operating characteristic analysis, lesion-level free-response receiver operating characteristic analysis, and patient-level bootstrap confidence intervals were used. Workflow support analyses modeled biopsy-avoidance strategies and AI-based risk stratification within PI-RADS 3 lesions.
RESULTS: After exclusions, 787 men (median age, 70 years; interquartile range, 64-75) were evaluated; 380 (48.3%) had csPCa. AI sensitivity was non-inferior to PI-RADS at the patient level (97.6% vs 92.6%; p < 0.001) but not at the lesion level (78.8% vs 88.8%; p = 0.98). Patient-level area under the curve was higher for AI than PI-RADS (0.80 vs 0.77; difference, 0.032; 95% confidence interval, 0.001-0.064; p = 0.043). A targeted PI-RADS 3 strategy avoided 18.9% of biopsies while maintaining 98.4% sensitivity. Among PI-RADS 3 lesions, AI upgraded 93.1% of csPCa-positive lesions and assigned a low risk to 24.1% of benign lesions.
CONCLUSION: The AI system was non-inferior for patient-level sensitivity but not lesion-level sensitivity. AI risk classification is best positioned as decision support for biopsy triage, particularly in PI-RADS 3 lesions.
KEY POINTS: Question Can three-tier artificial intelligence risk classification support prostate MRI biopsy triage in men while preserving detection of clinically significant prostate cancer? Findings Artificial intelligence had non-inferior patient-level sensitivity and higher patient-level discrimination than PI-RADS, but lower lesion-level sensitivity at the predefined threshold. Clinical relevance Artificial intelligence risk categories can complement PI-RADS by supporting high-sensitivity biopsy triage, especially for equivocal PI-RADS 3 lesions, while radiologists retain responsibility for lesion mapping.
引用
- Guenzel, Poirot, van Loon, et al. Multicentre evaluation of artificial intelligence risk classification for detection of clinically significant prostate cancer on biparametric MRI. Eur Radiol. 2026. PMID: 42800842. doi:10.1007/s00330-026-12888-8
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。