Detection and PI-RADS classification of focal lesions in prostate MRI: Performance comparison between a deep learning-based algorithm (DLA) and radiologists with various levels of experience

• Published in: European journal of radiology 2021-09, Vol.142, p.109894-109894, Article 109894
• Main Authors: Youn, Seo Yeon, Choi, Moon Hyung, Kim, Dong Hwan, Lee, Young Joon, Huisman, Henkjan, Johnson, Evan, Penzkofer, Tobias, Shabunin, Ivan, Winkel, David Jean, Xing, Pengyi, Szolar, Dieter, Grimm, Robert, von Busch, Heinrich, Son, Yohan, Lou, Bin, Kamen, Ali
• Format: Article
• Language: English
• Subjects: Deep learning; Magnetic resonance imaging; Prostate; Prostate imaging reporting and data system; Prostate neoplasms
• ISSN: 0720-048X; 1872-7727
• DOI: 10.1016/j.ejrad.2021.109894

•Deep learning-based algorithm (DLA) can promisingly assign PI-RADS categories.•PI-RADS categories assigned by the radiologists with different experience level varied.•The sensitivities and specificities of the DLA and expert were similar with PI-RADS ≥ 4.•The performance of DLA was similar to that of clinical reports in clinical practice. To compare the performance of lesion detection and Prostate Imaging-Reporting and Data System (PI-RADS) classification between a deep learning-based algorithm (DLA), clinical reports and radiologists with different levels of experience in prostate MRI. This retrospective study included 121 patients who underwent prebiopsy MRI and prostate biopsy. More than five radiologists (Reader groups 1, 2: residents; Readers 3, 4: less-experienced radiologists; Reader 5: expert) independently reviewed biparametric MRI (bpMRI). The DLA results were obtained using bpMRI. The reference standard was based on pathologic reports. The diagnostic performance of the PI-RADS classification of DLA, clinical reports, and radiologists was analyzed using AUROC. Dichotomous analysis (PI-RADS cutoff value ≥ 3 or 4) was performed, and the sensitivities and specificities were compared using McNemar’s test. Clinically significant cancer [CSC, Gleason score ≥ 7] was confirmed in 43 patients (35.5%). The AUROC of the DLA (0.828) for diagnosing CSC was significantly higher than that of Reader 1 (AUROC, 0.706; p = 0.011), significantly lower than that of Reader 5 (AUROC, 0.914; p = 0.013), and similar to clinical reports and other readers (p = 0.060–0.661). The sensitivity of DLA (76.7%) was comparable to those of all readers and the clinical reports at a PI-RADS cutoff value ≥ 4. The specificity of the DLA (85.9%) was significantly higher than those of clinical reports and Readers 2–3 and comparable to all others at a PI-RADS cutoff value ≥ 4. The DLA showed moderate diagnostic performance at a level between those of residents and an expert in detecting and classifying according to PI-RADS. The performance of DLA was similar to that of clinical reports from various radiologists in clinical practice.