A multicenter study of artificial intelligence-aided software for detecting visible clinically significant prostate cancer on mpMRI

Insights into Imaging

Original Article

Pengsheng Wu¹  Xiangpeng Wang¹  Zhangli Xing²  Yang Yu²  Xiaoying Wang³  Xiaodong Zhang³  Zhaonan Sun³  Zixuan Kong⁴  Ning Luo⁴  Yuntian Chen⁵  Bin Song⁵  Kexin Wang⁶ 

[1] Beijing Smart Tree Medical Technology Co. Ltd., Beijing, China;Department of Radiology, Fujian Medical University Union Hospital, Fuzhou, Fujian, China;Department of Radiology, Peking University First Hospital, No.8 Xishiku Street, Xicheng District, 100034, Beijing, China;Department of Radiology, The Second Affiliated Hospital of Dalian Medical University, Dalian, Liaoning, China;Department of Radiology, West China Hospital, Sichuan University, Chengdu, Sichuan, China;School of Basic Medical Sciences, Capital Medical University, Beijing, China;

关键词: Prostatic neoplasms;    Deep learning;    Diagnosis;    Computer-assisted;    Magnetic resonance imaging;   

DOI  :  10.1186/s13244-023-01421-w

 received in 2022-11-19, accepted in 2023-04-05,  发布年份 2023

来源: Springer

【 摘 要 】

BackgroundAI-based software may improve the performance of radiologists when detecting clinically significant prostate cancer (csPCa). This study aims to compare the performance of radiologists in detecting MRI-visible csPCa on MRI with and without AI-based software.Materials and methodsIn total, 480 multiparametric MRI (mpMRI) images were retrospectively collected from eleven different MR devices, with 349 csPCa lesions in 180 (37.5%) cases. The csPCa areas were annotated based on pathology. Sixteen radiologists from four hospitals participated in reading. Each radiologist was randomly assigned to 30 cases and diagnosed twice. Half cases were interpreted without AI, and the other half were interpreted with AI. After four weeks, the cases were read again in switched mode. The mean diagnostic performance was compared using sensitivity and specificity on lesion level and patient level. The median reading time and diagnostic confidence were assessed.ResultsOn lesion level, AI-aided improved the sensitivity from 40.1% to 59.0% (18.9% increased; 95% confidence interval (CI) [11.5, 26.1]; p < .001). On patient level, AI-aided improved the specificity from 57.7 to 71.7% (14.0% increase, 95% CI [6.4, 21.4]; p < .001) while preserving the sensitivity (88.3% vs. 93.9%, p = 0.06). AI-aided reduced the median reading time of one case by 56.3% from 423 to 185 s (238-s decrease, 95% CI [219, 260]; p < .001), and the median diagnostic confidence score was increased by 10.3% from 3.9 to 4.3 (0.4-score increase, 95% CI [0.3, 0.5]; p < .001).ConclusionsAI software improves the performance of radiologists by reducing false positive detection of prostate cancer patients and also improving reading times and diagnostic confidence.Clinical relevance statementThis study involves the process of data collection, randomization and crossover reading procedure.Graphical Abstract

【 授权许可】

CC BY   
© The Author(s) 2023

【 图 表 】

41116_2023_36_Article_IEq706.gif

41116_2023_36_Article_IEq501.gif

41116_2023_36_Article_IEq469.gif

41116_2023_36_Article_IEq323.gif