Objectives To assess Prostate Imaging Reporting and Data System (PI-RADS)–trained deep learning (DL) algorithm performance and to investigate the effect of data size and prior knowledge on the detection of clinically significant prostate cancer (csPCa) in biopsy-naïve men with a suspicion of PCa. Methods Multi-institution data included 2734 consecutive biopsy-naïve men with elevated PSA levels (≥ 3 ng/mL) that underwent multi-parametric MRI (mpMRI). mpMRI exams were prospectively reported using PI-RADS v2 by expert radiologists. A DL framework was designed and trained on center 1 data (n = 1952) to predict PI-RADS ≥ 4 (n = 1092) lesions from bi-parametric MRI (bpMRI). Experiments included varying the number of cases and the use of automatic zonal segmentation as a DL prior. Independent center 2 cases (n = 296) that included pathology outcome (systematic and MRI targeted biopsy) were used to compute performance for radiologists and DL. The performance of detecting PI-RADS 4–5 and Gleason > 6 lesions was assessed on 782 unseen cases (486 center 1, 296 center 2) using free-response ROC (FROC) and ROC analysis. Results The DL sensitivity for detecting PI-RADS ≥ 4 lesions was 87% (193/223, 95% CI: 82–91) at an average of 1 false positive (FP) per patient, and an AUC of 0.88 (95% CI: 0.84–0.91). The DL sensitivity for the detection of Gleason > 6 lesions was 85% (79/93, 95% CI: 77–83) @ 1 FP compared to 91% (85/93, 95% CI: 84–96) @ 0.3 FP for a consensus panel of expert radiologists. Data size and prior zonal knowledge significantly affected performance (4%, $$p<0.05$$ p < 0.05 ). Conclusion PI-RADS-trained DL can accurately detect and localize Gleason > 6 lesions. DL could reach expert performance using substantially more than 2000 training cases, and DL zonal segmentation. Key Points • AI for prostate MRI analysis depends strongly on data size and prior zonal knowledge. • AI needs substantially more than 2000 training cases to achieve expert performance.

Background Passive paramagnetic markers on magnetic resonance imaging (MRI)-compatible endovascular devices induce susceptibility artifacts, enabling MRI-visibility and real-time MRI-guidance. Optimised visibility is crucial for automatic detection and device tracking but depends on MRI technical parameters and marker characteristics. We assessed marker visibility and automatic detection robustness for varying MRI parameters and marker characteristics in a pulsatile flow phantom. Methods Guidewires with varying iron(II,III) oxide nanoparticle (IONP) concentration markers were imaged using gradient-echo (GRE) and balanced steady-state free precession (bSSFP) sequences at 3 T. Furthermore, echo time (TE), slice thickness (ST) and phase encoding direction (PED) were varied. Artifact width was measured and contrast-to-noise ratios were calculated. Marker visibility and image quality were scored by two MRI interventional radiologists. Additionally, a deep learning model for automatic marker detection was trained and the effects of the parameters on detection performance were evaluated. Two-tailed Wilcoxon signed-rank tests were used (significance level, p < 0.05). Results Medan artifact width (IQR) was larger in bSSFP compared to GRE images (12.7 mm (11.0–15.2) versus 8.4 mm (6.5–11.0)) (p < 0.001) and showed a positive relation with TE and IONP concentration. Switching PED and doubling ST had limited effect on artifact width. Image quality assessment scores were higher for GRE compared to bSSFP images. The deep learning model automatically detected the markers. However, the model performance was reduced after adjusting PED, TE, and IONP concentration. Conclusion Marker visibility was sufficient and a large range of artifact sizes was generated by adjusting TE and IONP concentration. Deep learning-based marker detection was feasible but performance decreased for altered MR parameters. These factors should be considered to optimise device visibility and ensure reliable automatic marker detectability in MRI-guided endovascular interventions.