Purpose:The amount of fibroglandular tissue (FGT) and background parenchymal enhancement (BPE) are evaluated as part of the BI-RADS reporting standard for the diagnosis of breast cancer in magnetic resonance imaging (MRI). BPE is associated with breast cancer risk and affects the diagnostic accuracy of breast MRI readings but is subject to very high inter-reader variability. Approach:We developed a fully automatic observer-independent method to measure the amount of FGT and BPE based on breast, FGT, and lesion masks, which classifies each into four classes based on thresholds applied to those scores. This retrospective study involved 2840 patients from seven institutions for whom T1-weighted dynamic contrast-enhanced breast MRI sequences with and without fat suppression were acquired using 1.5 or 3T scanners. FGT and BPE assessments for model development and evaluation were performed according to the BI-RADS guidelines by radiologists from the respective institutions and by a radiographer if the assessments were missing. Results:The performance of the models was compared using four-class accuracy and a paired t-test ( α = 0.05 ). Our automatic classification of FGT and BPE in a multi-institutional setting achieved a four-class accuracy of 0.535 for FGT and 0.504 for BPE ( random = 0.25 ). The calibration of the classification thresholds for each site individually resulted in a significant improvement, reaching an accuracy of 0.669 ( p < 0.05 ) for FGT and 0.565 ( p < 0.05 ) for BPE. The area under curve of receiver operating characterstic (ROC-AUC) to classify FGT and BPE into low versus high is 0.883 for FGT and 0.757 for BPE when computed across all sites. The averaged per-site AUCs are 0.926 for FGT and 0.810 for BPE. Conclusion:FGT and BPE classification perform significantly better when classification thresholds are selected per clinical site.

Breast cancer is the most prevalent cancer among women globally, and magnetic resonance imaging (MRI) is increasingly being proposed as a screening tool in addition to its role as a diagnostic imaging modality. This creates a need for automated methods that analyze breast MR images. One of the tasks in BI-RADS-based breast MRI reading is to assess whether implants are present. We trained nnU-Net models using 5-fold crossvalidation for the task of breast implant segmentation and assessed the segmentation quality on a multi-centric dataset consisting of 941 MRI studies, 205 of which contained breast implants. Our models achieved a mean (standard deviation) Dice score of 0.96 (0.03) and an average symmetric surface distance of 0.95 (0.67) mm in cases with breast implants. Of the 736 cases without implants, 575 were correctly predicted to have no implants, which increased to 732 with simple volume filtering. Our findings suggest that breast implant detection and segmentation in MRI can be solved well by deep learning approaches, even with a limited amount of data.

Abstract Background Phase-resolved functional lung magnetic resonance imaging (PREFUL-MRI) enables simultaneous, free-breathing, radiation-free assessment of regional lung perfusion and ventilation. This study aimed to provide preliminary reference data for lung perfusion and ventilation using PREFUL-MRI in healthy adults, and to characterize the physiological dependencies of these metrics on two breathing patterns, sex, and age. Methods In this prospective observational study, 87 healthy adults underwent PREFUL-MRI at 1.5 T during both normal and deep-slow breathing. Perfusion- and ventilation-related metrics were quantified via an automated pipeline. Paired comparisons between breathing states were performed using Wilcoxon signed-rank tests; unpaired comparisons between sexes and age groups (< 45 vs. ≥45 years) used Mann–Whitney U tests, with Holm–Bonferroni and Benjamini–Hochberg false-discovery-rate corrections applied to control for multiple comparisons. Results Mean perfusion (7.7% vs. 6.0%, Holm-Bonferroni adjusted p < 0.001) and ventilation defects (8.6% vs. 5.1%, Holm-Bonferroni adjusted p = 0.010) were decreased, and mean ventilation (15.8% vs. 48.3%, Holm-Bonferroni adjusted p < 0.001) and perfusion defects (1.9% vs. 7.9%, Holm-Bonferroni adjusted p = 0.005) increased during deep breathing compared with normal breathing. Twenty-eight participants had increased lung perfusion while 59 had reduced perfusion during deep breathing relative to normal breathing. During normal breathing, men exhibited higher mean ventilation than women (20.2% vs. 14.2%, Holm-Bonferroni adjusted p = 0.018). During deep breathing, men demonstrated higher total perfusion defect percentage and matched ventilation-perfusion defects than women (FDR adjusted q < 0.05). Total perfusion defect percentage was lower in participants aged ≥ 45 years than in those aged < 45 years (1.8% vs. 2.7%, FDR adjusted q = 0.036). Mean flow-volume loop correlations were similar between breathing patterns, sexes, and age groups after multiple comparison correction ( p > 0.05). Conclusions PREFUL-MRI can captures physiological variations related to breathing pattern, sex, and age in healthy adults. These findings provide a framework for distinguishing normal physiological heterogeneity from pathological change in clinical PREFUL-MRI interpretation. Clinical trial number Not applicable.

PURPOSE:The diagnosis of connective tissue disease-associated interstitial lung diseases (CTD-ILD) is connected to radiation exposure due to periodical CT scans. This study aims to investigate the alternative imaging method, Phase-Resolved Functional Lung (PREFUL), regarding its performance in low-field MRI. A comparison of PREFUL, photon-counting CT (PCCT) and pulmonary function tests (PFT) was performed to identify correlations that could restructure the diagnostics of CTD-ILD. MATERIALS AND METHODS:In this prospective single-center study, free-breathing PREFUL acquisitions of CTD-ILD patients were done after clinically indicated PCCT imaging. The severity and extent of CTD-ILD in PCCT were assessed via the Warrick score and used as a reference. Spearman's correlation coefficient ( r ) was calculated to examine the association between PREFUL, PCCT, and PFT. RESULTS:The data of 31 CTD-ILD patients (64.32±12.36 y, 10 men) were evaluated. Most correlations of PREFUL parameters with PFT were found with the Tiffeneau-Pinelli index (FEV1/FVC). The Warrick score showed excellent inter-rater agreement and correlations ( P <0.05) with the PFT parameters forced vital capacity (FVC) and the diffusing capacity of the lung for carbon monoxide corrected for hemoglobin (DLCOc) [FVC: r =-0.43, DLCOc SB: r =-0.65, DLCOc/VA: r =-0.50]. No correlation was found between PREFUL parameters and PCCT. CONCLUSIONS:The feasibility of PREFUL using low-field MRI was demonstrated in patients with CTD-ILD. Several correlations between PREFUL and PFT parameters were found, indicating that MRI can quantify lung function impairment. Nevertheless, CT remains the gold standard for CTD-ILD assessment and further research in PREFUL is needed.

BACKGROUND:Deep-learning neural network algorithms for detecting prostate cancer in MRI have proliferated in the literature. However, out of 30+ studies published since the PROSTATEx challenge, no studies tested the performance of their algorithm against using true external image data sets (studies came from an outside institution that did not supply any training data to the algorithm) while validating against MR-US fusion biopsy or whole-mount prostatectomy. Using true external data sets paints a much clearer picture of real-world clinical performance of an algorithm. PURPOSE:This work will assess the performance of a published deep learning (DL) neural network algorithm to detect prostate cancer using external studies. The main difference from other studies is the combination of using only MR-US fusion biopsy results as a gold standard; using test data from an institution that did not supply any training data for this version of the algorithm (including studies acquired with an endorectal coil, which were not in the original training set); and comparing the performance of algorithm-generated regions-of-interest (ROIs) versus algorithm heat maps. METHODS:Patients were included in the study if they had a prostate MRI with at least one radiologist-drawn target on MRI and underwent MR-US fusion biopsy where the target was sampled for pathological analysis. Patients were excluded if they had any history of prostate cancer treatment, had previously undergone MR-US fusion biopsy at our institution, were missing MRI acquisitions, had artifacts in image sets, or if the study had been shared for future algorithm development. MR image data was assessed using a DL research prototype (XProstate) from Siemens Healthineers that produced (a) ROIs in suspected cancer areas with a level of suspicion (LoS) score and (b) heat maps with LoS scores across the entire gland. The XProstate prototype had been trained with 2170 studies from eight different academic institutions. Clinical radiologist, XProstate ROI, and XProstate Heat Map scores were assessed with ROC analysis using pathology results from biopsy as a gold standard. RESULTS:202 unique patients were included for assessment of the XProstate research prototype. The ROC curve for the XProstate Heat Map LoS score generated the highest AUC (0.76, 95% CI: 0.70, 0.82) followed by clinical radiologist PI-RADS score (0.73, 95% CI: 0.68, 0.79) and by XProstate ROI LoS score (0.71, 95% CI: 0.65, 0.77). Neither the XProstate Heat Map (AUC difference = 0.03, 95% CI: -0.04, 0.10, p = 0.38) nor the XProstate ROI (AUC difference = -0.02, 95% CI: -0.09, 0.04, p = 0.43) was significantly different from the radiologist PI-RADS score. CONCLUSIONS:The XProstate prototype demonstrated equivalent performance as clinical radiologists when presented with de novo cases that would mirror a real-world clinical deployment. The automatic ROI delineation more closely matched clinical radiologist performance when using a cutoff of PI-RADS 5 for annotating suspicious regions. Overall, the XProstate prototype provided reasonable clinical performance and this study demonstrated the need to assess Deep Learning prototypes with external institutional test data.

Pompe disease is a life-limiting metabolic myopathy characterized by proximal muscle weakness. In late-onset Pompe disease (LOPD), respiratory failure due to respiratory muscle weakness is the leading cause of death. Monitoring relies on spirometry, which is often not feasible in pediatric or severely affected patients, highlighting the need for effort-independent assessments. Phase-resolved functional lung (PREFUL) MRI enables non-invasive, label-free evaluation of lung function during free breathing without active cooperation. In this prospective pilot study, ten LOPD patients and ten age- and sex-matched healthy controls underwent 0.55 T PREFUL MRI. The method proved feasible and detected ventilation-related functional impairments in LOPD patients. Moreover, differences were observed depending on non-invasive ventilation (NIV) status. These findings align with the pathophysiology of respiratory muscle involvement in LOPD. PREFUL MRI may serve as a non-invasive imaging biomarker for respiratory dysfunction and overall disease burden in neuromuscular diseases.

Abstract Objective To assess the performance of a deep learning-based computer-aided detection (DL-CAD) algorithm for prostate lesion detection and classification on biparametric (bp)MRI. Materials and methods This retrospective, single-center study included men undergoing 3-T MRI of the prostate for suspected prostate cancer (PCa) between July and September of 2022. Using the radiology report as the reference standard, detection performance for high-risk lesions (defined as PI-RADS ≥ 3, 4, 5) by the DL-CAD was evaluated per-patient using sensitivity, specificity, PPV, NPV and AUC; and per-lesion using sensitivity and PPV. Kappa statistics was used to assess per-patient detection and per-lesion classification of PI-RADS ≥ 3 lesions. Clinical and imaging factors associated with discordance between DL-CAD and radiology reports were assessed using Mann–Whitney, Chi-square, and Fisher’s exact tests. Results 442 adult males (mean age 65 ± 9 years) were assessed. Per-patient sensitivity, specificity, PPV, and NPV for detection of PI-RADS ≥ 4 and 5 lesions were 65.3%/81.2%/62.7%/82.9% and 82.1%/93.8%/65.7%/97.3%, respectively. Per-patient performance for identifying PI-RADS ≥ 3/4/5 lesions was fair-to-excellent: AUC = 0.67 (0.62–0.71)/0.75 (0.71–0.80)/0.92 (0.89–0.96). For detection of PI-RADS ≥ 4 and 5, per-lesion sensitivity was 60.4% and 78.3%, while PPV was 55.0% and 60.3%. Per-patient agreement between DL-CAD and the reference increased with higher PI-RADS scores (kappa = 0.26 (0.18–0.35)/0.46 (0.37–0.55)/0.68 (0.59–0.78)). Agreement on classification of PI-RADS ≥ 3 lesions was moderate (kappa = 0.56 (0.45–0.68)). Conclusion A pre-trained DL-CAD showed good-to-excellent per-patient performance for the detection of PI-RADS ≥ 4 lesions and moderate performance of PI-RADS ≥ 3 lesion classification. Future prospective studies validating the DL algorithm with histopathologic correlation are warranted. Critical relevance statement A deep learning computer-aided detection (DL-CAD) algorithm showed good-to-excellent per-patient performance for detection of PI-RADS ≥ 4 lesions, moderate performance of PI-RADS ≥ 3 lesion classification and high negative predictive value, which can be applied in the clinic with knowledge of its limitations. Key Points Clinical validation of deep learning computer-aided detection (DL-CAD) models for the detection and classification of prostate lesions on MRI is urgently needed. A pre-trained DL-CAD algorithm showed fair-to-excellent per-patient performance for detection of prostate lesions on biparametric MRI, with moderate performance for PI-RADS ≥ 3 lesion classification. Identification of false negatives and false positives of prostate cancer detection DL-CAD algorithms is important for future improvement and clinical deployment. A DL-CAD-based prostate cancer detection algorithm with high NPV may reduce interpretation time. Graphical Abstract

Early diagnosis and treatment of diabetic nephropathy (DN) is important for improving prognosis of patients, but the noninvasive and reliable diagnostic tool is lacking. We aim to investigate the ability of multiparametric MRI for detecting early DN in high-risk patients with type 2 diabetes. Between June 2023 and October 2024, we prospectively recruited 80 patients(diabetic patients group, n = 42, 53.4 ± 10.0 years old; DN patients group, n = 38, 47.9 ± 10.9 years old) and 16 healthy volunteers(control group, 49.5 ± 9.5 years old). All participants were examined using multiparametric MRI(IVIM, DKI, ASL and T1 mapping). The true diffusion coefficient(D), perfusion fraction(f), mean diffusivity(MD), mean kurtosis(MK), renal blood flow(RBF), T1 values in renal cortex were measured. Renal cortical MRI parameters among 3 groups were compared by one-way analysis of variance. Correlation between estimated glomerular filtration rate(eGFR), 24-hour urine albumin(24 h-UA) and renal cortical MRI parameters was evaluated using Spearman correlation analysis. The diagnostic performances of MRI parameters compared with biochemical indexes for detecting early DN were assessed using receiver operating characteristic curves. Renal cortical MRI parameters demonstrated statistically significant differences among 3 groups(P < 0.050). The eGFR, 24 h-UA significantly correlated with renal cortical MRI parameters(P < 0. 001). The areas under the curve(AUCs) for discriminating diabetic patients group from DN patients group were 0.706, 0.864, 0.732, 0.718, 0.910 and 0.718 for D, f, MD, MK, RBF and T1 values. AUCs of renal cortical f, RBF values were significantly larger than that of eGFR, serum creatinine, 24 h-UA, fasting blood glucose(P < 0.050). Intravoxel incoherent motion diffusion-weighted imaging and arterial spin labeling exhibited considerable promise as noninvasive tools for detecting DN in high-risk patients with type 2 diabetes.

The pathophysiological changes of lung perfusion and ventilation in fibrosing interstitial lung diseases (F-ILD) remain inadequately characterized. This study aimed to analyze lung perfusion and ventilation characteristics in F-ILD patients using phase-resolved functional lung magnetic resonance imaging (PREFUL MRI) as well as their correlation with the severity of F-ILD. This cross-sectional study prospectively included 30 patients diagnosed with F-ILD (19 males, 64.6 ± 9.5 years) and 30 age- and sex-matched normal controls. All participants underwent PREFUL MRI as well as pulmonary function tests. High-resolution CT (HRCT) was performed for the patient cohort. Ventilation and perfusion-related parameters obtained from PREFUL MRI were analyzed and correlated with PFTs and fibrotic lesions identified on HRCT. Compared with normal controls, F-ILD patients showed significant differences in mean perfusion (7.55

Active learning seeks to quickly improve model performance while reducing expert annotation effort, which is especially valuable in medical imaging.We study whether an uncertainty-based active learning scheme scales to a large, heterogeneous data collection and what performance is achievable with a limited annotation budget for nipple segmentation in breast magnetic resonance imaging – a precursor task to report lesion localization that has seen little automation so far.We prospectively evaluated an iterative active learning pipeline on 3,762 examinations from nine institutions / datasets, combining a 3D U-Net with Monte Carlo dropout-based image-wise uncertainty and a simple diversity-aware selection strategy. For almost all institutions, the nipple center-of-gravity error and Dice score improve substantially during the first three iterations, before largely plateauing, indicating diminishing returns. Uncertainty decreases in tandem with performance improvements, making it a practical proxy to guide annotation stopping. Persistent challenges involve absent or ambiguous nipples and institution-specific artifacts. Our results demonstrate that uncertainty-driven active learning can efficiently improve nipple segmentation on multi-center MRI with limited annotations. They highlight the need to detect nipple presence before segmentation and show that even if a model improves on data from most institutions when adding more data to the training, it can still achieve worse results in other institutions.

Medical foundation models improve generalization when training AI models with limited labeled data, but remain confined to a single specialty, such as pathology or radiology, and to either sparse or dense outputs, such as classification or segmentation. Here, we present CoM^3eT (Co-representation Multidimensional Multitask Medical Transformer), a medical vision foundation model that unifies pathology and radiology, sparse and dense predictions, and two- and higher-dimensional inputs by modeling multidimensional context with attention. CoM^3eT outperformed other medical foundation models in an open competition spanning five tomographic, four whole-specimen, and three two-dimensional datasets, covering sparse and dense prediction tasks as well as report generation. When adapted across diverse clinical applications, training fewer than 2.5

Background Image quality affects diagnostic performance; however, current quality metrics rely on subjective reader labels rather than diagnostic outcomes. Purpose To develop and evaluate a diagnostically calibrated artificial intelligence (AI) framework using deep learning (DL) for image quality assessment, linking prostate MRI quality scores to diagnostic performance. Materials and Methods This single-center retrospective study analyzed 12 496 consecutive prostate multiparametric MRI examinations performed between January 2014 and December 2023 at Radboud University Medical Center. A two-step framework was implemented: First, a DL model was trained on the most reliable low-quality (artifact-degraded structures) and high-quality (clear delineated zones) axial T2-weighted images to produce a continuous image quality score. Second, this score was applied in an independent internal test set and calibrated with diagnostic performance measured as (a) area under the receiver operating characteristic curve (AUC) for clinically significant prostate cancer (csPCa) detection by an AI model and (b) accuracy of csPCa detection by radiologists in routine clinical practice. The reference standard was histopathologic findings at biopsy. Diagnostic performance was calculated for quality score 10-percentile thresholds from 10% to 80%. Linear trend between performance and threshold was assessed using a permutation test on the regression slope. Results A total of 1229 T2-weighted series were used for training, and 568 multiparametric MRI examinations for diagnostic testing (1638 male patients; median age, 67 years [IQR, 62-71 years]). The DL model achieved a mean AUC of 0.99 for distinguishing low- versus high-quality images in the training set during fivefold cross-validation. For csPCa detection in the test set, the AI model AUC improved from 0.92 (95% CI: 0.89, 0.94) to 0.99 (95% CI: 0.97, 1.00) (P = .005) across thresholds of DL model image quality score; radiologist accuracy improved from 77% (95% CI: 73, 80) to 85% (95% CI: 78, 91) (P = .01). Conclusion The AI framework provided a standardized method to derive an objective image quality score that correlated with improved AI model diagnostic performance and radiologist accuracy. © RSNA, 2026 Supplemental material is available for this article. See also the editorial by Barrett in this issue.

RATIONALE AND OBJECTIVES:Perineural invasion (PNI) is a critical pathological feature predictive of early recurrence and poor survival in intrahepatic mass-forming cholangiocarcinoma (IMCC), yet it remains undetectable preoperatively. This study aimed to develop a multiparametric MRI-based texture model incorporating diffusion kurtosis imaging (DKI) and gadoxetic acid-enhanced sequences for noninvasive prediction of PNI and stratification of disease-free survival (DFS). MATERIALS AND METHODS:A prospective study was conducted including 99 pathologically confirmed IMCC patients (31 PNI-positive, 68 PNI-negative). All patients underwent 1.5 T MRI with T1-weighted, gadoxetic acid-enhanced (arterial, portal, transitional, hepatobiliary phases), DWI, and DKI sequences. Texture and histogram features were extracted from enhanced phases and parametric maps. A radiomics signature and a clinical-radiomics model were developed using logistic regression. A prognostic index for DFS was constructed using Cox regression. RESULTS:The radiomics signature combining AP_Contrast and DKI_K_Mean predicted PNI with an AUC of 0.766 (optimism-corrected AUC: 0.757). The clinical-radiomics model incorporating the target sign further improved the AUC to 0.837 (optimism-corrected AUC: 0.819), and decision curve analysis confirmed its clinical net benefit. A prognostic index based on microvascular invasion, TP_E.Kurtosis, and DKI_D_Std stratified DFS with an AUC of 0.742, distinguishing high- and low-risk groups with mean DFS of 10.7 vs 21.6 months (log-rank P = .001). CONCLUSION:A multiparametric MRI-based model integrating DKI and gadoxetic acid-enhanced texture features provides a noninvasive tool for preoperative prediction of PNI and DFS stratification in IMCC patients, potentially guiding individualized treatment strategies.

Childhood cancer survivors have a high risk of chronic multi-organ disease that does not plateau over time. To date, there is a lack of sensitive diagnostic techniques that allow early detection of tissue damage before clinical symptoms occur, particularly regarding pulmonary function. Free-breathing phase-resolved functional lung (PREFUL) low-field magnetic resonance imaging (LF-MRI) may enable visualization and quantification of functional and structural lung damage without the need of specific contrast agents. In this single-center, cross-sectional diagnostic study, we performed LF-MRI in a cohort of n = 27 children and adolescents (age range: 5 to 17 years) after treatment for acute lymphoblastic leukemia (ALL; n = 21) and Hodgkin’s disease (HD; n = 6) to determine the frequency of morphologic and functional lung parenchymal changes. Here, we show that despite the absence of clinical symptoms, significant time-dependent pulmonary ventilation and perfusion defects are detected. A negative correlation between the time after the end of therapy and defect-free lung tissue in the cohort of patients treated for ALL (Spearman-coefficient = − 0.69, p = 0.0005) is observed. Our results suggest an increase in pulmonary ventilation and perfusion defects preceding the increase in chronic disease that has already been reported in this patient population. Further research is needed to determine whether the functional abnormalities described in this study are an early morphological correlate of developing organ damage that may become clinically evident over time. PREFUL MRI may be an effective and highly sensitive tool for early detection of these changes in lung function, allowing longitudinal studies for risk stratification and potential future treatment adaptation. Dierl, Hinsen et al. investigate long term pulmonary toxicity in pediatric cancer survivors by the use of Free-breathing phase-resolved functional lung (PREFUL) MRI on a low-field system. Subclinical and time-dependent reduction in pulmonary ventilation and perfusion is revealed. Treatment of children and adolescents with cancer has led to a steady improvement in cure rates. At the same time, our knowledge of the possible long-term effects of conventional chemotherapy-based treatment in children remains limited. In this study, we performed a type of scan called PREFUL MRI in children and adolescents after treatment for acute lymphoblastic leukemia and Hodgkin’s disease that did not require treatment with chemicals called contrast agents or ionizing radiation. Our results show that following treatment there is an increasing and time-dependent dysfunction in the lungs, despite there being no clinical symptoms. Our imaging method might be an effective and highly sensitive tool for early detection of long-term toxicity after cancer treatment. It could also be used to monitor the side effects of new cancer treatments.

Geometric distortion in prostate diffusion-weighted imaging (DWI) can impair lesion localization and reduce the reliability of MRI-based clinical assessment. We propose AutoIQ, an ensemble machine learning framework for automatic quantification and classification of DWI geometric distortion severity. A total of 140 retrospective prostate biparametric MRI examinations were analyzed, including 33 scans with severe distortion requiring repeat acquisition and 107 scans with acceptable distortion based on expert radiologist assessment. AutoIQ combines two complementary distortion quantification strategies: a segmentation-based method measuring prostate boundary mismatch between T2-weighted imaging (T2WI) and DWI, and a registration-based method estimating deformation magnitude after DWI-to-T2WI alignment. The resulting distortion scores were used to train individual classifiers and a logistic-regression ensemble model. Both computational methods significantly differentiated severe from acceptable distortion cases (p < 0.001). On an independent test set, the ensemble model achieved an accuracy of 0.95, F1-score of 0.93, and AUC of 0.98, outperforming individual models. These results suggest that AutoIQ can provide automated, quantitative quality assessment for prostate DWI and may help identify scans that require repeat acquisition.

Providing insights into the feasibility of pulmonary function assessment in real-world cardiovascular magnetic resonance (CMR) practice by applying Phase-REsolved FUnctional Lung imaging (PREFUL). We retrospectively analyzed consecutive patients who underwent PREFUL imaging in addition to routine 1.5T CMR between September 2023 and January 2024. PREFUL was acquired in three coronal slices, with a prototype tool used to derive quantitative perfusion and ventilation defect percentages (QDP and VDP, respectively). Cardiac function was assessed from short-axis cine images. Subgroup analyses included patients with primary pulmonary disease and reduced left ventricular ejection fraction (LVEF). Statistical analyses comprised linear regression, correlation analysis, and Kruskal-Wallis test. The final cohort included N = 172 patients (74 females), median age 60 years (IQR 46–71). PREFUL was feasible in all cases (mean scan time 60 s/slice). Multivariable regression with bootstrap-based backward selection showed associations of QDP with LVEF, pulmonary disease, age, and BMI (all p ≤ 0.005), while VDP was associated with pulmonary disease, age, and male sex (all p < 0.001). QDP correlated negatively with LV stroke volume (ρ (Spearman’s rho) − 0.336, p < 0.001) and cardiac output (ρ − 0.360; p < 0.001) and was higher in patients with LVEF < 50

PURPOSE:To compare prostate cancer lesion detection using conventional and artificial intelligence (AI)-assisted image interpretation at multiparametric MRI (mpMRI). MATERIALS AND METHODS:A retrospective study of 53 consecutive patients who underwent prostate mpMRI and subsequent prostate tissue sampling was performed. Two board-certified radiologists (with 4 and 12 years of experience) blinded to the clinical information interpreted anonymized exams using the PI-RADS v2.1 framework without and with an AI-assistance tool. The AI software tool provided radiologists with gland segmentation and automated lesion detection assigning a probability score for the likelihood of the presence of clinically significant prostate cancer (csPCa). The reference standard for all cases was the prostate pathology from systematic and targeted biopsies. Statistical analyses assessed interrater agreement and compared diagnostic performances with and without AI assistance. RESULTS:Within the entire cohort, 42 patients (79 %) harbored Gleason-positive disease, with 25 patients (47 %) having csPCa. Radiologists' diagnostic performance for csPCa was significantly improved over conventional interpretation with AI assistance (reader A: AUC 0.82 vs. 0.72, p = 0.03; reader B: AUC 0.78 vs. 0.69, p = 0.03). Without AI assistance, 81 % (n = 36; 95 % CI: 0.89-0.91) of the lesions were scored similarly by radiologists for lesion-level characteristics, and with AI assistance, 59 % (26, 0.82-0.89) of the lesions were scored similarly. For reader A, there was a significant difference in PI-RADS scores (p = 0.02) between AI-assisted and non-assisted assessments. Signficant differences were not detected for reader B. CONCLUSION:AI-assisted prostate mMRI interpretation improved radiologist diagnostic performance over conventional interpretation independent of reader experience.

Objective. Lung cancer remains the leading cause of cancer-related mortality worldwide, with Non-Small Cell Lung Cancer (NSCLC) accounting for approximately 85% of all cases. Programmed cell Death Ligand-1 (PD-L1) is a well-established biomarker that guides immunotherapy in advanced-stage NSCLC, currently evaluated via invasive biopsy procedures. This study aims to develop and validate a non-invasive pipeline for stratifying PD-L1 expression using quantitative analysis of IVIM parameter maps—diffusion (D), pseudo-diffusion (D*), perfusion fraction (pf)—and T1-VIBE MRI acquisitions. Approach. MRI data from 43 NSCLC patients were analysed and labelled as PD-L1 positive (≥1%) or negative (<1%) based on immunohistochemistry exam. After pre-processing, 1,171 radiomic features and 512 deep learning features were obtained. Three feature sets (radiomic, deep learning, and fusion) were tested with Logistic Regression, Random Forest, and XGBoost. Four discriminative features were selected using the Mann–Whitney U-test, and model performance was primarily assessed using the area under the receiver operating characteristic curve (AUC). Robustness was ensured through repeated stratified 5-fold cross-validation, bootstrap-derived confidence intervals, and permutation test. Main Results. Logistic Regression generally demonstrated the highest classification performance, with AUC values ranging from 0.78 to 0.92 across all feature sets. Fusion models outperformed or matched the performance of the best standalone radiomics or deep learning model. Among multisequence MRI, the IVIM-D fusion features yielded the best performance with an AUC of 0.92, followed by IVIM-D* radiomic features that showed a similar AUC of 0.91. For IVIM-pf and T1-VIBE derived features, the fusion model yielded the best AUC values of 0.87 and 0.90, respectively. Significance. The obtained results highlight the potential of a combined radiomic-deep learning approach to effectively detect PD-L1 expression from MRI acquisitions, paving the way for a non-invasive PD-L1 evaluation procedure.