Rosenkrantz Andrew B, Begovic Jovan, Pires Antonio, Won Eugene, Taneja Samir S, Babb James S
Department of Radiology, NYU Langone Health, New York, NY.
Department of Radiology, NYU Langone Health, New York, NY.
Curr Probl Diagn Radiol. 2019 Mar-Apr;48(2):132-141. doi: 10.1067/j.cpradiol.2018.01.003. Epub 2018 Jan 10.
To assess the effect on reader performance of an interactive case-based online tutorial for prostate magnetic resonance imaging (MRI) interpretation using Prostate Imaging and Reporting Data System (PI-RADS).
An educational website was developed incorporating scrollable multiparametric prostate MRI examinations with annotated solutions based on PI-RADS version 2. Three second-year radiology residents evaluated a separate set of 60 prostate MRI examinations both before and after review of the online case material, identifying and scoring dominant lesions. These 60 examinations included 30 benign cases and 30 cases with a dominant lesion demonstrating Gleason score ≥3 + 4 tumor on fusion-targeted biopsy. The readers' pooled performance was compared between the 2 sessions using logistic regression and Wilcoxon signed rank tests.
All readers completed the online material within four-hours. Review of the online material significantly improved sensitivity (from 57.8%-73.3%, P = 0.003) and negative predictive value (from 69.2%-78.2%, P = 0.049), but not specificity (from 70.0%-67.8%, P = 0.692) or positive predictive value (from 59.6%-64.7%, P = 0.389). Reader confidence (1-10 scale; 10 = maximal confidence) also improved significantly (from 5.6 ± 2.7 to 6.3 ± 2.6, P = 0.026). However, accuracy of assigned PI-RADS scores did not improve significantly (from 45.5%-53.3%, P = 0.149).
An online interactive case-based website in prostate MRI interpretation improved novice readers' sensitivity and negative predictive value for tumor detection, as well as readers' confidence. This online material may serve as a resource complementing existing traditional methods of instruction by providing a more flexible educational experience among a larger volume of learners. However, further more targeted educational initiatives regarding the proper application of PI-RADS remain warranted.
评估使用前列腺影像报告和数据系统(PI-RADS)的基于病例的交互式在线教程对前列腺磁共振成像(MRI)解读中读者表现的影响。
开发了一个教育网站,其中包含可滚动的多参数前列腺MRI检查以及基于PI-RADS第2版的带注释解决方案。三名放射科二年级住院医师在查看在线病例资料前后分别评估了另一组60例前列腺MRI检查,识别并对主要病变进行评分。这60例检查包括30例良性病例和30例在融合靶向活检中显示Gleason评分≥3 + 4肿瘤的主要病变病例。使用逻辑回归和Wilcoxon符号秩检验比较了两个阶段读者的综合表现。
所有读者均在四小时内完成了在线资料的学习。查看在线资料后,敏感性(从57.8%提高到73.3%,P = 0.003)和阴性预测值(从69.2%提高到78.2%,P = 0.049)显著提高,但特异性(从70.0%降至67.8%,P = (此处原文有误,应为0.692))或阳性预测值(从59.6%提高到64.7%,P = 0.389)未显著提高。读者信心(1 - 10分制;10分 = 最大信心)也显著提高(从5.6 ± 2.7提高到6.3 ± 2.6,P = 0.026)。然而,指定PI-RADS评分的准确性未显著提高(从45.5%提高到53.3%,P = 0.149)。
前列腺MRI解读的在线交互式病例网站提高了新手读者对肿瘤检测的敏感性和阴性预测值,以及读者信心。该在线资料可作为一种资源,通过为大量学习者提供更灵活的教育体验来补充现有的传统教学方法。然而,关于PI-RADS正确应用的更具针对性的教育举措仍有必要。