Background/Objectives: There is a growing need for efficient prostate MRI protocols due to their increasing use in managing prostate cancer (PCa) and potential inclusion in screening. Deep learning reconstruction (DLR) may enhance MR acquisitions and improve image quality compared to conventional acceleration techniques. This systematic review examines DLR approaches to prostate MRI. Methods: A search of PubMed, Web of Science, and Google Scholar identified eligible studies comparing DLR to conventional reconstruction for prostate imaging. A narrative synthesis was performed to summarize the impact of DLR on acquisition time, image quality, and diagnostic performance. Results: Thirty-three studies showed that DLR can reduce acquisition times for T₂w and DWI imaging while maintaining or improving image quality. It did not significantly affect clinical tasks, such as biopsy decisions, and performed comparably to human readers in PI-RADS scoring and the detection of extraprostatic extension. However, AI models trained on conventional data might be less accurate with DLR images. The heterogeneity in image quality metrics among the studies prevented quantitative synthesis. Discussion: DLR has the potential to achieve substantial time savings in prostate MRI while maintaining image quality, which is especially relevant because of increased MRI demands. Future research should address the effect of DLR on clinically relevant downstream tasks, including AI algorithms’ performances and biopsy decisions, and explore task-specific accelerated protocols for screening, image-guided biopsy, and treatment.

Objective: To evaluate and compare cancer detection rates (CDRs) of transition zone (TZ) lesions upgraded from PI-RADSv2.1 score 2 to 3 (“2 + 1”) or from 3 to 4 (“3 + 1”) using DWI and assess their clinical impact. Materials and methods: A systematic literature search was performed in Embase, Medline, and Web of Science for studies assessing TZ lesions with DWI in PI-RADSv2.1, with histology-confirmed grade group ≥ 2 cancer (GG ≥ 2) as the primary outcome. Risk of bias was evaluated using QUADAS-2. Pooled sensitivity, specificity, CDRs, and odds ratios (ORs) were estimated at the lesion level using a bivariate binomial random-effects model. Results: Eight studies with 1535 TZ lesions were included. GG ≥ 2 CDRs for PI-RADS scores of 1, 2, 2 + 1, 3, 3 + 1, 4, and 5 were 2% (95%CI: 0%–12%), 6% (4%–10%), 13% (6%–23%), 19% (15%–25%), 37% (24%–52%), 49% (32%–67%), and 73% (66%–79%), respectively. Scores of 2 + 1 had higher GG ≥ 2 CDRs than 2 (OR 3.37 (1.53–7.44), p = 0.003) but were similar to 3 (OR 0.80 (0.44–1.45), p = 0.46). Scores of 3 + 1 had higher GG ≥ 2 CDRs than 3 (OR 2.67 (1.27–5.59), p = 0.009) but were similar to 4 (OR 0.68 (0.33–1.44), p = 0.32). False-positive rates remained substantial (≥ 2 + 1: 69% (55%–80%); ≥ 3: 54% (46%–62%)). Conclusion: The risk of having significant prostate cancer in “2 + 1” and “3 + 1” TZ lesions, with an upgrading based on DWI images, is appropriately categorized within the PI-RADS v2.1 scoring system, as shown by this meta-analysis. Key Points: Question PI-RADS v2.1 incorporates rules allowing scores of some transition zone (TZ) lesions to be increased. Literature on the clinical impact of these rules is scarce. Findings For TZ lesions upgraded with DWI: “2 + 1” lesions show a cancer detection rate (CDR) of 13%, and “3 + 1” lesions show a CDR of 37%. Clinical relevance Upgraded TZ lesions may impact individualized biopsy-decisions, especially as “3 + 1” lesions harbor significant disease in 2-out-of-5 patients. Still, the high rate of grade group = 1 and benign findings in these sub-categories emphasizes the need for strategies to minimize overdiagnosis. Graphical

Background: Biologic agents used in patients with inflammatory bowel diseases (IBD) may influence the pathophysiology of coexistent metabolic-dysfunction associated steatotic liver disease (MASLD). This study primarily aimed to evaluate the six-month effect of infliximab or vedolizumab vs. no biologics on presumed hepatic steatosis in patients with IBD. Secondary endpoints were their effect on hepatic fibrosis and parameters related to hepatic metabolism. Methods: This prospective, non-randomized, controlled trial assigned adult bio-naïve patients with IBD into three groups: infliximab, vedolizumab, or controls (receiving no biologic). The baseline was the time of the initiation of biologic agents and the endpoint six months later. Hepatic steatosis was evaluated with transabdominal ultrasonography (Hamaguchi score), whereas controlled attenuation parameter (CAP), fatty liver index (FLI), and hepatic steatosis index (HSI) were used as surrogates. Hepatic fibrosis was evaluated with liver stiffness (LS), fibrosis-4 index (FIB-4), and nonalcoholic fatty liver disease (NAFLD) fibrosis score. Results: Sixty-six patients were assigned to infliximab (n = 26), vedolizumab (n = 14), or control (n = 26); At the endpoint, the Hamaguchi score, CAP, FLI, and HSI were not different between groups. LS was not different between groups; however, FIB-4 was increased within all groups, and NAFLD fibrosis score was increased within infliximab and control groups, without significant biologic × time interactions. Conclusions: No positive or adverse effect of infliximab or vedolizumab vs. no biologic agents was shown on presumed hepatic steatosis in patients with IBD, who have not been previously exposed to biologic agents. Although no effect of both biologic agent on LS, a slight but significant increase in FIB-4 and NAFLD fibrosis score warrants further studying.

Radiogenomic modelling faces a significant challenge due to the high-dimensional nature of quantitative radiomic features and limited sample sizes. Feature selection is therefore essential to eliminate irrelevant features and mitigate overfitting. Particle swarm optimisation (PSO) has shown promise for effectively navigating large feature spaces, yet its effectiveness in radiogenomics remains unexplored. This study investigates the value of PSO-based methods, both independently and in combination with other advanced techniques, for MRI-based prediction of human papillomavirus (HPV) status in oropharyngeal squamous cell carcinoma (OPSCC). Baseline contrast-enhanced T1-weighted MR scans from two centres were analysed: 153 patients in an internal cohort (randomly split into 80 % for training and 20 % for testing) and 157 patients in an external validation cohort. Radiomic features were extracted from manually segmented tumours and multiple feature selection methods, including PSO and its ensembles, filter-based methods, wrapper-based approaches, and shrinkage techniques were evaluated. Performance was measured and compared using the area under the receiver operating characteristic curve (AUC). PSO alone had a reasonable predictive power on the internal test set (AUC = 0.76, 95 % CI: 0.57-0.92, p = 0.092). When combined with adaptive LASSO using Shapley values, PSO's performance improved (AUC = 0.81, 95 % CI: 0.61-0.94, p = 0.023). Recursive feature elimination (RFE) selected the most relevant features (AUC = 0.91, 95 % CI: 0.79-1.00, p < 0.001). Despite this, RFE failed to generalise well to the external cohort (AUC = 0.52, 95 % CI: 0.42-0.60, p = 1). Meanwhile, the PSO-adaptive LASSO combination maintained a robust AUC = 0.78 (95 % CI: 0.70-0.85, p < 0.001), indicating superior generalisability. The explainable PSO-adaptive LASSO feature selection method provides generalisable radiogenomic signatures associated with HPV status in OPSCC, outperforming other feature selection approaches. This combination may serve as a robust strategy for developing transferable models in radiogenomics.