Faculty Bibliography

PURPOSE/OBJECTIVE: METHODS:Rep2Rep learning extends the Noise2Noise framework by training a neural network on two repeated MRI acquisitions, using one repetition as input and another as target, without requiring ground-truth data. It incorporates noise-adaptive training, enabling denoising generalization across varying noise-levels and flexible inference with any number of repetitions. Performance was evaluated on both synthetic noisy Brain MRI and 0.55T Prostate MRI data, and compared against supervised learning and Monte Carlo Stein's Unbiased Risk Estimator (MC-SURE). RESULTS:Rep2Rep learning outperforms MC-SURE on both synthetic and 0.55T MRI datasets. On synthetic Brain data, it achieved denoising quality comparable to supervised learning and surpassed MC-SURE, particularly in preserving structural details and reducing residual noise. On the 0.55T Prostate MRI data a reader study showed that Rep2Rep-denoised 2-average images outperformed 8-average noisy images. Rep2Rep demonstrated robustness to noise-level discrepancies between training and inference, supporting its practical implementation. CONCLUSION/CONCLUSIONS:Rep2Rep learning offers an effective self-supervised denoising for low-field MRI by leveraging routinely acquired multi-repetition data. Its noise-adaptivity enables generalization to different SNR regimes without clean reference images. This makes Rep2Rep learning a promising tool for improving image quality and scan efficiency in low-field MRI.

OBJECTIVES/OBJECTIVE:Despite its high negative predictive value (NPV) for clinically significant prostate cancer (csPCa), MRI suffers from a substantial number of false positives, especially for intermediate-risk cases. In this work, we determine whether a deep learning model trained with PI-RADS-guided representation learning can disambiguate the PI-RADS 3 classification, detect csPCa from bi-parametric prostate MR images, and avoid unnecessary benign biopsies. MATERIALS AND METHODS/METHODS:This study included 28,263 MR examinations and radiology reports from 21,938 men imaged for known or suspected prostate cancer between 2015 and 2023 at our institution (21 imaging locations with 34 readers), with 6352 subsequent biopsies. We trained a deep learning model, a representation learner (RL), to learn how radiologists interpret conventionally acquired T2-weighted and diffusion-weighted MR images, using exams in which the radiologists are confident in their risk assessments (PI-RADS 1 and 2 for the absence of csPCa vs. PI-RADS 4 and 5 for the presence of csPCa, n=21,465). We then trained biopsy-decision models to detect csPCa (Gleason score ≥7) using these learned image representations, and compared them to the performance of radiologists, and of models trained on other clinical variables (age, prostate volume, PSA, and PSA density) for treatment-naïve test cohorts consisting of only PI-RADS 3 (n=253, csPCa=103) and all PI-RADS (n=531, csPCa=300) cases. RESULTS:On the 2 test cohorts (PI-RADS-3-only, all-PI-RADS), RL-based biopsy-decision models consistently yielded higher AUCs in detecting csPCa (AUC=0.73 [0.66, 0.79], 0.88 [0.85, 0.91]) compared with radiologists (equivocal, AUC=0.79 [0.75, 0.83]) and the clinical model (AUCs=0.69 [0.62, 0.75], 0.78 [0.74, 0.82]). In the PIRADS-3-only cohort, all of whom would be biopsied using our institution's standard of care, the RL decision model avoided 41% (62/150) of benign biopsies compared with the clinical model (26%, P<0.001), and improved biopsy yield by 10% compared with the PI-RADS ≥3 decision strategy (0.50 vs. 0.40). Furthermore, on the all-PI-RADS cohort, RL decision model avoided 27% of additional benign biopsies (138/231) compared to radiologists (33%, P<0.001) with comparable sensitivity (93% vs. 92%), higher NPV (0.87 vs. 0.77), and biopsy yield (0.75 vs. 0.64). The combination of clinical and RL decision models further avoided benign biopsies (46% in PI-RADS-3-only and 62% in all-PI-RADS) while improving NPV (0.82, 0.88) and biopsy yields (0.52, 0.76) across the 2 test cohorts. CONCLUSIONS:Our PI-RADS-guided deep learning RL model learns summary representations from bi-parametric prostate MR images that can provide additional information to disambiguate intermediate-risk PI-RADS 3 assessments. The resulting RL-based biopsy decision models also outperformed radiologists in avoiding benign biopsies while maintaining comparable sensitivity to csPCa for the all-PI-RADS cohort. Such AI models can easily be integrated into clinical practice to supplement radiologists' reads in general and improve biopsy yield for any equivocal decisions.

OBJECTIVES/OBJECTIVE:Quantitative proton density fat fraction (PDFF) and R2* estimation at lower field strengths, such as 0.55 T, is challenging due to lower signal-to-noise ratio, reduced fat water chemical shift, and increased T2* relaxation times. In this study, we propose a 3D hybrid technique for abdominal imaging at 0.55 T that enables the simultaneous acquisition of T2-weighted and T1-weighted images and quantification of fat fraction and R2* parameters. MATERIALS AND METHODS/METHODS:Numerical simulations were performed to optimize a prototype radial hybrid turbo spin echo gradient echo (TSE-GRE) acquisition scheme for improved PDFF and R2* estimation accuracy. Phantom imaging experiments with and without motion were performed to evaluate the sensitivity of the estimation to external motion. Eleven volunteers were imaged on a prototype 0.55 T system. Data were acquired using the proposed technique under free-breathing conditions, and motion-compensated reconstruction was performed using the respiratory signal from a pilot-tone device. Image contrast and estimation performance were compared with conventional acquisition schemes in vitro and in vivo. RESULTS:Numerical simulations indicated R2* estimation accuracy was more sensitive to the choice of echo time compared with PDFF. Performing motion compensation reduced the mean error in R2* from 24 to 5 s-1 while the mean error in PDFF only reduced from 2.7% to 1.6%. The proposed technique generated T2-weighted images with comparable relative liver-spleen contrast as conventional imaging and there were no significant differences (P>0.05) in the PDFF and R2* values estimated from the hybrid technique compared with conventional multi-echo GRE. Further, the free-breathing acquisition allowed improved slice coverage while overcoming breath-hold limitations of conventional acquisition schemes. CONCLUSIONS:The use of a hybrid TSE-GRE acquisition technique can allow simultaneous morphological and quantitative PDFF and R2* estimation at 0.55 T under free-breathing conditions.

PURPOSE/OBJECTIVE:To improve 0.55T T2-weighted PROPELLER lung MRI by developing a self-supervised framework for joint reconstruction and denoising. METHODS:T2-weighted 0.55T lung MRI datasets from 44 patients with prior COVID-19 infection were used. Each PROPELLER blade was split along the readout direction into two disjoint subsets: one subset for training an unrolled network, and the other for loss calculation. Following the Noise2Noise paradigm, this framework split k-space into two subsets with independent, matched noise but identical underlying signal, enabling joint reconstruction and denoising without external training references. For comparison, coil-wise Marchenko-Pastur Principal Component Analysis (MPPCA) denoising followed by parallel imaging reconstruction was performed. The reconstructed images were evaluated by two experienced chest radiologists. RESULTS:The self-supervised model generated lung images with improved clarity, better delineation of parenchymal and airway structures, and maintained high fidelity in cases with available CT references. In addition, the proposed framework also enabled further reduction of scan time by reconstructing images with adequate diagnostic quality from only half the number of blades. The reader study confirmed that the proposed method outperformed MPPCA across all categories (Wilcoxon signed-rank test, p < 0.001), with moderate inter-reader agreement (weighted Cohen's kappa = 0.55; percentage of exact and within ±1 point agreement = 91%). CONCLUSION/CONCLUSIONS:By leveraging the intrinsic data redundancy in PROPELLER sampling and extending the Noise2Noise concept, the proposed self-supervised framework enabled simultaneous reconstruction and denoising of lung images at 0.55T to address the low-SNR challenge at low-field. It holds great potential for broad use in other low-field MRI applications.

Prostate MRI has transformed lesion detection and risk stratification in prostate cancer, but its impact is constrained by the high cost of the exam, variability in interpretation, and limited scalability. False negatives, false positives, and moderate inter-reader agreement undermine reliability, while long acquisition times restrict throughput. Artificial intelligence (AI) offers potential solutions to address many of the limitations of prostate MRI in the clinical management pathway. Machine learning-based triage can refine patient selection to optimize resources. Deep learning reconstruction enables accelerated acquisition while preserving diagnostic quality, with multiple FDA-cleared products now in clinical use. Ongoing development of automated quality assessment and artifact correction aims to improve reliability by reducing nondiagnostic exams. In image interpretation, AI models for lesion detection and clinically significant prostate cancer prediction achieve performance comparable to radiologists, and the PI-CAI international reader study has provided the strongest evidence to date of non-inferiority at scale. More recent work extends MRI-derived features into prognostic modeling of recurrence, metastasis, and functional outcomes. This review synthesizes progress across five domains-triage, accelerated acquisition and reconstruction, image quality assurance, diagnosis, and prognosis-highlighting the level of evidence, validation status, and barriers to adoption. While acquisition and reconstruction are furthest along, with FDA-cleared tools and prospective evaluations, triage, quality control, and prognosis remain earlier in development. Ensuring equitable performance across populations, incorporating uncertainty estimation, and conducting prospective workflow trials will be essential to move from promising prototypes to routine practice. Ultimately, AI could accelerate the adoption of prostate MRI toward a scalable platform for earlier detection and population-level prostate cancer management. EVIDENCE LEVEL: N/A TECHNICAL EFFICACY: 3.

The development of noninvasive MRI biomarkers as surrogates of histopathological features in kidney tissue requires detailed explorations of contrast. Therefore, we studied kidney diffusion kurtosis imaging (DKI) with a wide array of encodings, including flow compensation, variable directional sampling, and cardiac gating regimes. Twelve healthy volunteers underwent DKI at 5-10 diffusion weightings (b-values) ranging from 0 to 1200 smm^-2 with 12 or 30 directional samplings, bipolar or flow-compensated diffusion gradient waveforms, and at systolic or diastolic cardiac phases. DKI biomarkers, mean diffusivity (MD) and kurtosis (MK), were interrogated using a directionally robust fitting algorithm compared to conventional fits. The combination of flow compensation and cardiac triggering at the diastolic phase in the kidneys reduced flow effects on DKI. In systole, flow-compensated waveforms significantly reduced MD and MK for both cortex and medulla: cortex MD: 3.00 versus 2.55 μm² ms^-1, medulla MD: 2.80 versus 2.39 μm² ms^-1, cortex MK: 0.58 versus 0.45, and medulla MK: 0.60 versus 0.47 (all p < 0.05). Flow suppression alleviated requirements for processing the DKI at higher minimum b-values, as neither MD nor MK significantly differed at the diastolic phase for minimum b-values of 0 versus 200 smm^-2: cortex MD: 2.30 versus 2.28 μm² ms^-1, p = 0.278; medulla MD: 2.29 versus 2.28 μm² ms^-1, p = 0.437; cortex MK: 0.37 versus 0.36, p = 0.308; and medulla MK: 0.40 versus 0.40, p = 0.904. Flow-compensated waveforms mitigate cardiac and respiratory motion-related artifacts at higher diffusion encodings in addition to microcirculation effects. The robust fitting initially developed for brain DKI is highly applicable to the kidneys because it disentangles tissue-specific directional diffusion information from artifacts.

The rapid advancements in artificial intelligence (AI) carry the promise to reshape abdominal imaging by offering transformative solutions to challenges in disease detection, classification, and personalized care. AI applications, particularly those leveraging deep learning and radiomics, have demonstrated remarkable accuracy in detecting a wide range of abdominal conditions, including but not limited to diffuse liver parenchymal disease, focal liver lesions, pancreatic ductal adenocarcinoma (PDAC), renal tumors, and bowel pathologies. These models excel in the automation of tasks such as segmentation, classification, and prognostication across modalities like ultrasound, CT, and MRI, often surpassing traditional diagnostic methods. Despite these advancements, widespread adoption remains limited by challenges such as data heterogeneity, lack of multicenter validation, reliance on retrospective single-center studies, and the "black box" nature of many AI models, which hinder interpretability and clinician trust. The absence of standardized imaging protocols and reference gold standards further complicates integration into clinical workflows. To address these barriers, future directions emphasize collaborative multi-center efforts to generate diverse, standardized datasets, integration of explainable AI frameworks to existing picture archiving and communication systems, and the development of automated, end-to-end pipelines capable of processing multi-source data. Targeted clinical applications, such as early detection of PDAC, improved segmentation of renal tumors, and improved risk stratification in liver diseases, show potential to refine diagnostic accuracy and therapeutic planning. Ethical considerations, such as data privacy, regulatory compliance, and interdisciplinary collaboration, are essential for successful translation into clinical practice. AI's transformative potential in abdominal imaging lies not only in complementing radiologists but also in fostering precision medicine by enabling faster, more accurate, and patient-centered care. Overcoming current limitations through innovation and collaboration will be pivotal in realizing AI's full potential to improve patient outcomes and redefine the landscape of abdominal radiology.

PURPOSE/OBJECTIVE:Pancreatic cystic lesions (PCL) commonly undergo surveillance using MRI with MR cholangiopancreatography (MRCP). Our objective is to compare the performance of a single-shot fat-saturated T2-weighted technique with deep-learning reconstruction (DL HASTE-FS) to a conventional T2-weighted Half fourier Single-shot Turbo spin-Echo (HASTE) sequence and to MRCP for the purpose of PCL detection, characterization, and surveillance. METHODS:In this retrospective study, 91 consecutive patients underwent 3T abdominal MRI with MRCP protocol including DL HASTE-FS and conventional HASTE between 8/2/2023 and 10/3/2023. Three abdominal radiologists rated overall and lesion-specific image quality on a 5-point Likert scale, including pancreatic margin and duct sharpness, and PCL conspicuity. A subset of 70 preselected index PCLs were evaluated for cyst features, confidence of diagnosing side-branch IPMN, and suitability of DL HASTE-FS in replacing MRCP for PCL surveillance. RESULTS:DL HASTE-FS received higher scores for pancreatic duct border sharpness (4.1 vs. 3.9; p = .004), pancreatic duct visibility compared to MRCP (2.0 vs. 1.9; p = .04), cyst conspicuity (4.4 vs. 3.9; p < .001), and sharpness of cyst wall and internal septations (4.3 vs. 3.7; p < .001) compared to conventional HASTE. In contrast, conventional HASTE received higher scores for pancreatic margin sharpness (4.2 vs. 3.8; p < .001) and peripancreatic vessel clarity (4.2 vs. 3.4; p < .001). For the 70 preselected index PCLs, readers visualized more PCLs and had higher confidence in diagnosing SB-IPMN on DL HASTE-FS than on conventional HASTE (3.6 vs. 3.4; p < .001). Finally, DL HASTE-FS was deemed a suitable replacement to MRCP for more cases than conventional HASTE (83% vs. 48%; p < .001). CONCLUSION/CONCLUSIONS:DL HASTE-FS outperforms conventional HASTE for PCL detection and characterization, and is a suitable alternative to 3D MRCP in the context of PCL surveillance, potentially reducing exam time and cost.

BACKGROUND:Access to prostate MRI remains limited due to resource constraints and the need for expert interpretation. PURPOSE/OBJECTIVE:To develop machine learning (ML) models that enable risk-based triage for prostate MRI (ProMT-ML) in the evaluation of prostate cancer. STUDY TYPE/METHODS:Retrospective and prospective. POPULATION/METHODS:A total of 11,879 retrospective MRI scans for suspected prostate cancer from a multi-hospital health system, divided into training (N = 9504) and test (N = 2375) sets. A total of 4551 records for prospective validation. FIELD STRENGTH/SEQUENCE/UNASSIGNED:1.5T and 3T/Turbo-spin echo T2-weighted imaging (T2WI), diffusion-weighted imaging (DWI), and dynamic contrast-enhanced (DCE). ASSESSMENT/RESULTS:Prostate Imaging Reporting and Data System (PI-RADS) scores were retrieved from MRI reports. The Boruta algorithm was used to select final input features from candidate features. Two models were developed using supervised ML to estimate the likelihood of an abnormal MRI, defined as PI-RADS ≥ 3: Model A (with prostate volume) and Model B (without prostate volume). Models were compared to PSA. Prostate biopsy pathology was assessed to evaluate potential clinical impact. STATISTICAL TESTS/METHODS:Area under the receiver operating characteristic curve (AUC) was the primary performance metric. RESULTS:A total of 5580 (46.9%) subjects had a PI-RADS score ≥ 3. After feature selection, Model A included age, PSA, body mass index, and prostate volume, while Model B included age, PSA, body mass index, and systolic blood pressure. Both models A (AUC 0.711) and B (AUC 0.616) significantly outperformed PSA (AUC 0.593). Compared to PSA threshold > 4 ng/mL, Model A demonstrated significantly improved specificity (28.3% vs. 21.9%) and no significant difference in sensitivity (89.0% vs. 86.7%). Among false negatives (Model A: 8.0% (62/776); Model B: 16.8% (130/776)), most (Model A: 87%; Model B: 69%) had benign or clinically insignificant disease on biopsy. On prospective validation, both versions of ProMT-ML significantly outperformed PSA. DATA CONCLUSION/CONCLUSIONS:ProMT-ML provides personalized risk estimates of abnormal prostate MRI and can support triage of this test. TECHNICAL EFFICACY/UNASSIGNED:Stage 4.