Background Clinically significant prostate cancer (PCa) diagnosis at MRI requires accurate and efficient radiologic interpretation. Although artificial intelligence may assist in this task, lack of transparency has limited clinical translation. Purpose To develop an explainable artificial intelligence (XAI) model for clinically significant PCa diagnosis at biparametric MRI using Prostate Imaging Reporting and Data System (PI-RADS) features for classification justification. Materials and Methods This retrospective study included consecutive patients with histopathologic analysis-proven prostatic lesions who underwent biparametric MRI and biopsy between January 2012 and December 2017. After image annotation by two radiologists, a deep learning model was trained to detect the index lesion; classify PCa, clinically significant PCa (Gleason score ≥ 7), and benign lesions (eg, prostatitis); and justify classifications using PI-RADS features. Lesion- and patient-based performance were assessed using fivefold cross validation and areas under the receiver operating characteristic curve. Clinical feasibility was tested in a multireader study and by using the external PROSTATEx data set. Statistical evaluation of the multireader study included Mann-Whitney U and exact Fisher-Yates test. Results Overall, 1224 men (median age, 67 years; IQR, 62-73 years) had 3260 prostatic lesions (372 lesions with Gleason score of 6; 743 lesions with Gleason score of ≥ 7; 2145 benign lesions). XAI reliably detected clinically significant PCa in internal (area under the receiver operating characteristic curve, 0.89) and external test sets (area under the receiver operating characteristic curve, 0.87) with a sensitivity of 93% (95% CI: 87, 98) and an average of one false-positive finding per patient. Accuracy of the visual and textual explanations of XAI classifications was 80% (1080 of 1352), confirmed by experts. XAI-assisted readings improved the confidence (4.1 vs 3.4 on a five-point Likert scale; P = .007) of nonexperts in assessing PI-RADS 3 lesions, reducing reading time by 58 seconds (P = .009). Conclusion The explainable AI model reliably detected and classified clinically significant prostate cancer and improved the confidence and reading time of nonexperts while providing visual and textual explanations using well-established imaging features. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Chapiro in this issue.

Objective.: To investigate the pathophysiological effect of interleukin 6 (IL-6) on lupus nephritis in MRL-Fas^lprmice.Methods.: We generated IL-6-deficient MRL-Fas^lprmice using a backcross/intercross breeding scheme. Renal pathology was evaluated using immunohistochemistry detection for macrophages, lymphocytes, vascular cell adhesion molecule-1 (VCAM-1), and TUNEL (terminal deoxynucleotide transferase-mediated dUTP nick end-labeling) for apoptotic cells, and renal IgG and C3 deposition by immunofluorescence staining. Expression of inflammatory markers in the spleen was analyzed by quantitative real-time reverse transcription-polymerase chain reaction. Serum cytokine concentrations were detected by FACS analysis.Results.: IL-6 deficiency was highly effective in prolonging survival and ameliorating the clinical, immunological, and histological indicators of murine systemic lupus erythematosus. During the study period of 6 months, MRL-Fas^lprIL-6 −/− mice showed delayed onset of proteinuria and hematuria compared to IL-6-intact control mice. Survival rate was 100% in IL-6-deficient MRL-Fas^lprmice and 25% in the control group at 6 months of age. The absence of IL-6 resulted in significant reduction of infiltrating macrophages in the kidney (p < 0.05), a decrease in renal IgG and C3 deposition, and a reduction of CD4+ and CD8+ lymphocytes. The parenchymal adhesion molecule VCAM-1 was found to be downregulated in kidneys of MRL-Fas^lprIL-6 −/− compared to IL-6-intact mice. We found elevated serum levels of IL-10 and interferon-γ in IL-6-deficient mice, while splenic mRNA showed an overall downregulation of immunoregulatory genes.Conclusion.: IL-6 is a strong promoter of lupus nephritis and may be a promising new therapeutic target in the treatment of human lupus nephritis.

The second version of the PI-RADS™ (Prostate Imaging Reporting and Data System) was introduced in 2015 to standardize the interpretation and reporting of prostate multiparametric magnetic resonance imaging. Recently low cancer detection rates were reported for PI-RADS version 2 category 4 lesions. Therefore the aim of the study was to evaluate the cancer detection rate of PI-RADS version 2 in a large prospective cohort. The study included 704 consecutive men with primary or prior negative biopsies who underwent magnetic resonance imaging/ultrasound fusion guided targeted biopsy and 10-core systematic prostate biopsy between September 2015 and May 2017. All lesions were rated according to PI-RADS version 2 and lesions with PI-RADS version 2 category 3 or greater were biopsied. An ISUP (International Society of Urological Pathology) score of 2 or greater (ie Gleason 3 + 4 or greater) was defined as clinically significant prostate cancer. The overall cancer detection rate of PI-RADS version 2 categories 3, 4 and 5 was 39%, 72% and 91% for all prostate cancer, and 23%, 49% and 77% for all clinically significant prostate cancer, respectively. If only targeted biopsy had been performed, 59 clinically significant tumors (16%) would have been missed. The PI-RADS version 2 score was significantly associated with the presence of prostate cancer (p <0.001), the presence of clinically significant prostate cancer (p <0.001) and the ISUP grade (p <0.001). PI-RADS version 2 is significantly associated with the presence of clinically significant prostate cancer. The cancer detection rate of PI-RADS version 2 category 4 lesions was considerably higher than previously reported. When performing targeted biopsy, the combination with systematic biopsy still provides the highest detection of clinically significant prostate cancer.