Objectives 

The aim of this study was to estimate the prospective utility of a previously retrospectively validated convolutional neural network (CNN) for prostate cancer (PC) detection on prostate magnetic resonance imaging (MRI).

Materials and Methods 

The biparametric (T2-weighted and diffusion-weighted) portion of clinical multiparametric prostate MRI from consecutive men included between November 2019 and September 2020 was fully automatically and individually analyzed by a CNN briefly after image acquisition (pseudoprospective design). Radiology residents performed 2 research Prostate Imaging Reporting and Data System (PI-RADS) assessments of the multiparametric dataset independent from clinical reporting (paraclinical design) before and after review of the CNN results and completed a survey. Presence of clinically significant PC was determined by the presence of an International Society of Urological Pathology grade 2 or higher PC on combined targeted and extended systematic transperineal MRI/transrectal ultrasound fusion biopsy. Sensitivities and specificities on a patient and prostate sextant basis were compared using the McNemar test and compared with the receiver operating characteristic (ROC) curve of CNN. Survey results were summarized as absolute counts and percentages.

Results 

A total of 201 men were included. The CNN achieved an ROC area under the curve of 0.77 on a patient basis. Using PI-RADS ≥3-emulating probability threshold (c3), CNN had a patient-based sensitivity of 81.8% and specificity of 54.8%, not statistically different from the current clinical routine PI-RADS ≥4 assessment at 90.9% and 54.8%, respectively (P = 0.30/P = 1.0). In general, residents achieved similar sensitivity and specificity before and after CNN review. On a prostate sextant basis, clinical assessment possessed the highest ROC area under the curve of 0.82, higher than CNN (AUC = 0.76, P = 0.21) and significantly higher than resident performance before and after CNN review (AUC = 0.76 / 0.76, P ≤ 0.03). The resident survey indicated CNN to be helpful and clinically useful.

Conclusions 

Pseudoprospective paraclinical integration of fully automated CNN-based detection of suspicious lesions on prostate multiparametric MRI was demonstrated and showed good acceptance among residents, whereas no significant improvement in resident performance was found. General CNN performance was preserved despite an observed shift in CNN calibration, identifying the requirement for continuous quality control and recalibration.

中文翻译：

---

放射科住院医师与前列腺癌检测深度学习系统的伪前瞻性临床互动：经验、性能和对间歇性重新校准需求的识别

目标 

本研究的目的是评估先前经过回顾性验证的卷积神经网络 (CNN) 在前列腺磁共振成像 (MRI) 上检测前列腺癌 (PC) 的前瞻性效用。

材料和方法 

2019 年 11 月至 2020 年 9 月期间纳入的连续男性的临床多参数前列腺 MRI 的双参数（T2 加权和扩散加权）部分在图像采集后由 CNN 进行全自动和单独分析（伪前瞻性设计）。放射科住院医师在审查 CNN 结果前后对独立于临床报告（临床旁设计）的多参数数据集进行了 2 项研究前列腺成像报告和数据系统 (PI-RADS) 评估，并完成了一项调查。具有临床意义的 PC 的存在是由国际泌尿病理学会 2 级或更高级别的 PC 联合靶向和扩展系统性经会阴 MRI/经直肠超声融合活检确定的。使用 McNemar 检验比较患者和前列腺六分仪的敏感性和特异性，并与 CNN 的接受者操作特征 (ROC) 曲线进行比较。调查结果总结为绝对数量和百分比。

结果 

总共包括 201 名男性。CNN 在患者基础上实现了 0.77 曲线下的 ROC 面积。使用 PI-RADS ≥3 模拟概率阈值 (c3)，CNN 基于患者的敏感性为 81.8%，特异性为 54.8%，与目前临床常规 PI-RADS ≥4 评估的 90.9% 和 54.8% 没有统计学差异, 分别 ( P = 0.30/ P = 1.0)。总的来说，居民在 CNN 审查前后获得了相似的敏感性和特异性。在前列腺六分仪的基础上，临床评估具有最高的 ROC 曲线下面积 0.82，高于 CNN（AUC = 0.76，P = 0.21），并且显着高于 CNN 审查前后的居民表现（AUC = 0.76 / 0.76，P≤ 0.03)。居民调查表明 CNN 很有帮助并且在临床上有用。

结论 

在前列腺多参数 MRI 上对可疑病变进行基于 CNN 的全自动检测的伪前瞻性准临床整合得到了证明，并在住院医师中表现出良好的接受度，而住院医师的表现没有显着改善。尽管观察到 CNN 校准发生变化，但确定了持续质量控制和重新校准的要求，但 CNN 的一般性能得到了保留。