Multicenter Multireader Evaluation of an Artificial Intelligence-Based Attention Mapping System for the Detection of Prostate Cancer With Multiparametric MRI.
Multicenter Multireader Evaluation of an Artificial Intelligence-Based Attention Mapping System for the Detection of Prostate Cancer With Multiparametric MRI.
复制标题
DOI：
10.2214/ajr.19.22573
复制
发表时间：
2020-10
期刊：
影响因子：
--
通讯作者：
Turkbey B
中科院分区：
文献类型：
--
作者：
Mehralivand S;Harmon SA;Shih JH;Smith CP;Lay N;Argun B;Bednarova S;Baroni RH;Canda AE;Ercan K;Girometti R;Karaarslan E;Kural AR;Purysko AS;Rais-Bahrami S;Tonso VM;Magi-Galluzzi C;Gordetsky JB;Macarenco RSES;Merino MJ;Gumuskaya B;Saglican Y;Sioletic S;Warren AY;Barrett T;Bittencourt L;Coskun M;Knauss C;Law YM;Malayeri AA;Margolis DJ;Marko J;Yakar D;Wood BJ;Pinto PA;Choyke PL;Summers RM;Turkbey B
本研究的目的是在多中心数据集中评估具有注意力映射的人工智能 (AI) 检测系统与多参数 MRI (mpMRI) 解释在前列腺癌检测中的性能。本研究纳入了来自 5 个机构的 MRI 检查，并由 9 名读者进行了评估。在第一轮中，读者使用前列腺成像报告和数据系统版本 2 评估 mpMRI 研究。4 周后，图像再次与基于人工智能的检测系统输出一起呈现给读者。读者在四个人工智能生成的注意力图框中接受或拒绝病变。框外的其他病变被排除在检测和分类之外。比较了使用仅 mpMRI 和人工智能辅助方法的读者的表现。研究人群包括 152 名病例患者和 84 名对照患者，有 274 个经病理证实的癌症病变。 MRI 中基于病变的 AUC 为 74.9%，AI 中基于病变的 AUC 为 77.5%，无显着差异 (p = 0.095)。 AI 总体检测癌症病变的敏感性高于 mpMRI，但未达到统计学显着性（57.4% vs 53.6%，p = 0.073）。然而，对于移行区病变，AI 的敏感性高于 MRI（61.8% vs 50.8%，p = 0.001）。 AI 的读取时间比 MRI 长（4.66 分钟 vs 4.03 分钟，p < 0.001）。 AI 和 MRI 的读者间一致性中等，无显着差异（58.7% vs 58.5%，p = 0.966）。通过使用人工智能系统，总体灵敏度仅得到最小程度的提高。然而，使用 AI 系统在检测移行区病变方面取得了显着的进步，但平均增加了 40 秒的读取时间。
The purpose of this study was to evaluate in a multicenter dataset the performance of an artificial intelligence (AI) detection system with attention mapping compared with multiparametric MRI (mpMRI) interpretation in the detection of prostate cancer. MRI examinations from five institutions were included in this study and were evaluated by nine readers. In the first round, readers evaluated mpMRI studies using the Prostate Imaging Reporting and Data System version 2. After 4 weeks, images were again presented to readers along with the AI-based detection system output. Readers accepted or rejected lesions within four AI-generated attention map boxes. Additional lesions outside of boxes were excluded from detection and categorization. The performances of readers using the mpMRI-only and AI-assisted approaches were compared. The study population included 152 case patients and 84 control patients with 274 pathologically proven cancer lesions. The lesion-based AUC was 74.9% for MRI and 77.5% for AI with no significant difference (p = 0.095). The sensitivity for overall detection of cancer lesions was higher for AI than for mpMRI but did not reach statistical significance (57.4% vs 53.6%, p = 0.073). However, for transition zone lesions, sensitivity was higher for AI than for MRI (61.8% vs 50.8%, p = 0.001). Reading time was longer for AI than for MRI (4.66 vs 4.03 minutes, p < 0.001). There was moderate interreader agreement for AI and MRI with no significant difference (58.7% vs 58.5%, p = 0.966). Overall sensitivity was only minimally improved by use of the AI system. Significant improvement was achieved, however, in the detection of transition zone lesions with use of the AI system at the cost of a mean of 40 seconds of additional reading time.