Faculty Bibliography

OBJECTIVE:To quantify the operational burden of motion-related repeat MRI sequences across a high-volume health system and evaluate whether repeat acquisitions improve image quality and confidence in brain MRI and magnetic resonance cholangiopancreatography (MRCP). METHODS:This retrospective study analyzed six months of MRI examinations on scanners with sequence-level analytics. Examinations with technologist-labeled repeated sequences were identified. Brain MRI and MRCP baseline-repeat pairs were evaluated by four and two readers, respectively. Readers scored image quality and diagnostic confidence and indicated whether another repeat would be requested. Scores were compared using linear mixed-effects models with patient-level random intercepts. RESULTS:Among 85,349 MRI examinations, 4,072 (4.8%) included at least one repeated sequence. Repeat frequency was highest in pediatrics (9.9%) and lowest in breast MRI (3.5%). In the brain MRI study, 87 baseline-repeat pairs from 77 patients were evaluated. Repeat acquisition significantly improved image quality and diagnostic confidence across all metrics (all p<0.001). 19.5% of repeat sequences did not improve. Improvement declined with greater elapsed examination time. In the MRCP study, 30 baseline-repeat pairs from 27 patients were evaluated. Repeat acquisition did not significantly improve image quality or diagnostic confidence, and both readers would have requested another repeat for half of repeated acquisitions. DISCUSSION/CONCLUSIONS:Motion-driven repeat MRI sequences impose measurable operational burden, but diagnostic yield varies substantially by anatomic region and sequence type. Brain MRI repeats generally improved image quality and diagnostic confidence, whereas MRCP repeats showed limited benefit. These findings support sequence-level tracking to guide targeted motion-mitigation strategies and reduce low-value repeat imaging.

OBJECTIVES/OBJECTIVE:Deep learning (DL)-based reconstruction methods can substantially accelerate MRI acquisition while preserving image quality relative to traditional techniques. However, concerns remain regarding their ability to accurately reconstruct abnormalities at higher acceleration factors. This study quantitatively assesses the impact of higher-order acceleration on image quality in DL-reconstructed brain MRI in normal and abnormal cases, emphasizing the utmost need to render abnormal findings accurately. MATERIALS AND METHODS/METHODS:Raw k-space image data from 5847 brain MRI examinations (fastMRI) were used. A radiologist-annotated subset of 1001 examinations (fastMRI+) served as the test set; the remainder were split into training (n=3715) and validation (n=1131) sets. Images were retrospectively undersampled (acceleration factors 2 to 12) and reconstructed using an end-to-end variational network. Image quality metrics, including normalized mean squared error (NMSE), structural similarity index measure (SSIM), and peak signal-to-noise ratio (PSNR), were computed within bounding-box annotations containing abnormalities and in size-matched normal control regions, and at the slice-level. The region-by-acceleration interaction was tested using linear mixed-effects models with a subject-level random intercept. RESULTS:All metrics demonstrated progressive degradation with increasing acceleration across slice-level and bounding-box-level analyses. At lower acceleration rates, bounding boxes with and without abnormalities showed comparable image quality. However, above 4-fold acceleration, NMSE within abnormal regions exhibited a notably sharper increase with rising acceleration compared to normal regions, while abnormal regions demonstrated a significantly faster decline in PSNR and SSIM. This nonsymmetric information loss was particularly evident for NMSE at higher acceleration rates. The region-by-acceleration interaction was statistically significant for all 3 metrics at both levels (all adjusted P<0.001), with the effect largest and most consistent for NMSE. CONCLUSIONS:Our findings demonstrate a disproportionate, pathology-localized degradation of reconstruction quality as acceleration increases. This emphasizes the importance of rigorous clinical assessment of image quality in DL-MRI reconstruction, particularly at higher undersampling rates, highlighting the need for reconstruction strategies tailored to maintaining diagnostically critical features.

Parallel imaging techniques reduce magnetic resonance imaging (MRI) scan time but image quality degrades as the acceleration factor increases. In clinical practice, conservative acceleration factors are chosen because no mechanism exists to automatically assess the diagnostic quality of undersampled reconstructions. This work introduces a general framework for pixel-wise uncertainty quantification in parallel MRI reconstructions, enabling automatic identification of unreliable regions without access to any ground-truth reference image. Our method integrates conformal quantile regression with image reconstruction methods to estimate statistically rigorous pixelwise uncertainty intervals. We trained and evaluated our model on Cartesian undersampled brain and knee data obtained from the fastMRI dataset using acceleration factors ranging from 2 to 10. An end-to-end Variational Network was used for image reconstruction. Quantitative experiments demonstrate strong agreement between predicted uncertainty maps and true reconstruction error. Using our method, the corresponding Pearson correlation coefficient was higher than 90% at acceleration levels at and above four-fold; whereas it dropped to less than 70% when the uncertainty was computed using a simpler a heuristic notion (magnitude of the residual). Qualitative examples further show the uncertainty maps based on quantile regression capture the magnitude and spatial distribution of reconstruction errors across acceleration factors, with regions of elevated uncertainty aligning with pathologies and artifacts. The proposed framework enables evaluation of reconstruction quality without access to fully-sampled ground-truth reference images. It represents a step toward adaptive MRI acquisition protocols that may be able to dynamically balance scan time and diagnostic reliability.

PURPOSE/OBJECTIVE:The goal of this work is to introduce an automated method to assess the quality of under-sampled MRI reconstructions. THEORY AND METHODS/METHODS:We propose a general framework for pixel-wise uncertainty quantification in accelerated MRI reconstructions, enabling automatic identification of unreliable regions without using ground-truth fully-sampled reference images. Our method integrates conformal quantile regression with learning-based image reconstruction methods to estimate statistically rigorous pixel-wise uncertainty intervals. We trained and evaluated our model on Cartesian undersampled brain and knee data obtained from the fastMRI dataset using acceleration factors ranging from 2 to 10. An end-to-end Variational Network was used for image reconstruction. RESULTS:Quantitative experiments demonstrate strong agreement between predicted uncertainty maps and true reconstruction error. Using our method, the corresponding Pearson correlation coefficient was higher than 90% at acceleration levels at and above four-fold; whereas it dropped to less than 70% when the uncertainty was computed using a simpler heuristic notion (magnitude of the residual). Qualitative examples further show the uncertainty maps based on quantile regression capture the magnitude and spatial distribution of reconstruction errors across acceleration factors, with regions of elevated uncertainty aligning with pathologies and artifacts. CONCLUSION/CONCLUSIONS:The proposed framework enables evaluation of reconstruction quality without access to fully-sampled ground-truth reference images. It represents a step toward adaptive MRI acquisition protocols that may be able to dynamically balance scan time and diagnostic reliability.

Artificial intelligence (AI) has progressed from technical research to routine clinical use, reaching an inflection point where technological capabilities may exceed current regulatory and oversight frameworks. These systems are becoming more complex, progressing from narrow, task-specific algorithms to foundation models and early agentic prototypes. This progression has redistributed risk, responsibility, and clinical judgment, requiring radiologists and health care leaders to understand how policy choices affect patient safety and clinical innovation advancement. This special report provides a roadmap for aligning policy with clinical practice through a practical, lifecycle-based framework centered on patient safety. Translational bialignment is a concept that pairs regulatory science requirements (what AI systems should deliver to clinicians and patients) with implementation science capabilities (what institutions should provide for safe deployment of AI). This framework addresses the full AI lifecycle, from data stewardship and model development to validation, deployment, and monitoring, and articulates shared responsibilities for vendors, institutions, and clinicians grounded in trustworthy AI principles. The analysis focuses on U.S. regulatory frameworks, particularly Food and Drug Administration policies governing medical AI, with relevant highlights from international approaches. Concrete opportunities for radiologists to engage in policy formation, participate in oversight, and collaborate with industry and policymakers are provided to help shape a trustworthy and sustainable AI ecosystem. The alignment of policy, practice, and patient safety will enable medical AI to have a lasting impact on clinical care and public trust. The analysis and recommendations provided represent the authors' perspectives and do not necessarily reflect the official positions of the Radiological Society of North America. Keywords: Artificial Intelligence, Food and Drug Administration, Health Policy, Implementation Science, Large Language Models, Machine Learning, Patient Safety, Regulatory Science, Translational Bialignment © RSNA, 2026.

BACKGROUND AND PURPOSE/OBJECTIVE:Artificial intelligence (AI) models have shown promise in neuroradiology, yet their real-world generalizability remains uncertain, partly due to variability in imaging acquisition and protocols. We aimed to evaluate the impact of data source, scanner manufacturer, scan mode, slice thickness, and the AI models developed by participating teams on AI performance in this secondary analysis of the 2019 American Society of Functional Neuroradiology (ASFNR) AI Competition. MATERIALS AND METHODS/METHODS:We included 1,177 anonymized noncontrast head CT scans from five institutions. Four teams participated, developing models to detect acute ischemic stroke, intracranial hemorrhage, mass effect, and to assess age-appropriate normality. Generalized estimating equations (GEE) were used to evaluate the effects of the aforementioned variables on model performance, and collinearity diagnostics were applied to exclude redundant variables. RESULTS:Due to collinearity with scanner manufacturer, data source was excluded from the model. Across all tasks, the AI model employed significantly influenced performance. Scanner manufacturer was significantly associated with accuracy in detecting intracranial hemorrhage and acute ischemic stroke but not mass effect or age-based normality. Slice thickness significantly associated with detection of intracranial hemorrhage and mass effect, with thinner slices yielding higher accuracy, but showed no effect on ischemic stroke or normality assessments. Scan mode did not significantly influence performance for any task. CONCLUSION/CONCLUSIONS:This secondary analysis demonstrates that imaging acquisition and protocol variability may significantly affect AI model performance. Scanner manufacturer, slice thickness, and the developed AI model were significantly associated with model accuracy, whereas scan mode had no significant impact. Among these, the developed AI model consistently proved most influential, reflecting the importance of training data, model architecture, and preprocessing methods.

BACKGROUND:Foundation models have shown remarkable potential in medical imaging by leveraging extensive pretraining on general datasets to enable fine-tuning for specific tasks. This is thought to be particularly beneficial for tasks where annotated data is scarce. A key underlying assumption, however, is that these models can learn from small amounts of training data more efficiently than existing state-of-the-art models. PURPOSE/OBJECTIVE:This study aims to characterize the performance of two major foundation segmentation models (SAM and MedSAM) when fine-tuned to segment neuroanatomic structures across a spectrum of dataset sizes, compared to a standard fully-supervised UNet model. METHODS:This study used 1,113 T1-weighted 3D MRIs from the Human Connectome Project's Young Adult cohort with corresponding Freesurfer-generated, manually-refined segmentations of 93 gray and white matter regions. The dataset was divided into 891 (80%) training MRIs, 111 (10%) validation MRIs, and 111 (10%) testing MRIs. SAM and MedSAM models were first fine-tuned and compared against a standard UNet model using Dice score to establish the baseline performance using all training 3D volumes. Subsequently, MedSAM and UNet models were fine-tuned across a varying number of training volumes to assess performance with diminishing dataset size, down to a single MRI, as well as no MRIs (zero-shot) for the MedSAM and SAM models. RESULTS:Using the entire training set, UNet outperformed MedSAM and SAM across most regions, with median Dice scores of 0.88 versus 0.82 and 0.84, respectively (p < 0.001). With diminishing dataset size, UNet continued to perform as well as or better than MedSAM in the three studied regions, down to even a single 3D volume. In the zero-shot setting, SAM and MedSAM showed some ability to segment with overall median Dice scores of 0.66 and 0.59, respectively. CONCLUSIONS:SAM and MedSAM did not outperform a standard UNet model in segmentation tasks, even in extremely limited training data settings, contrary to the foundation model hypothesis, suggesting that foundation models do not necessarily yield superior fine-tuned performance compared to standard segmentation models in the low data setting. Instead, the potential benefit of foundation models will depend on the characteristics of the task at hand and the behavior and capacity of the specific foundation model in question. Thus, it will be essential to benchmark against standard supervised deep learning methods for each distinct application to demonstrate the added value of using a foundation model.

Mild traumatic brain injury (mTBI) is a prevalent yet often overlooked public health concern due to the absence of detectable abnormalities on CT or conventional MRI scans. Approximately 18.3%-31.3% of mTBI patients experience persistent symptoms 3-6 months post-injury, despite normal imaging results, making diagnosis and treatment challenging. In recent years, advanced neuroimaging modalities have emerged with the potential to reveal subtle physiological and structural brain changes that are invisible to traditional imaging. Diffusion MRI (dMRI), for instance, is particularly valuable for detecting white matter injury; perfusion MRI assesses alterations in cerebral blood flow; sodium MRI (²³Na MRI) provides insights into ionic homeostasis; and functional MRI (fMRI) detects disruptions in functional brain network connectivity. In this review, we first explore the underlying mechanisms of mTBI and then summarize current evidence supporting the use of advanced MRI techniques to detect injury signatures associated with these mechanisms. Finally, we highlight populations at heightened risk for repeated injuries-underscoring the urgent need for more sensitive diagnostic tools that can identify injury early, guide return-to-activity decisions, and prevent cumulative brain damage. EVIDENCE LEVEL: N/A. TECHNICAL EFFICACY: Stage 3.

BACKGROUND AND PURPOSE/OBJECTIVE:Mild traumatic brain injury (MTBI) is a common public health concern with potential long-term consequences, yet its underlying pathophysiology remains poorly understood. Clinical heterogeneity of individuals having diverse extent and array of symptoms has impeded the identification of reliable imaging biomarkers. Traditional group-level analyses may obscure biologically meaningful subtypes. This study uses latent class analysis (LCA) to classify MTBI subjects into symptom-defined subgroups and examines corresponding WM microstructural alterations using advanced diffusion MRI. MATERIALS AND METHODS/METHODS:Sixty-one MTBI patients within one month of injury completed the Rivermead Post-Concussion Symptoms Questionnaire (RPQ). LCA was used to identify symptom-based subgroups. Of these, 54 MTBI patients underwent multi-shell diffusion MRI and were compared with 31 controls. WM changes were assessed across subgroups using ROI-based diffusion analyses. RESULTS:LCA identified three distinct MTBI subgroups: those with minimal to no symptoms (31.5%), the cognitively symptomatic (38.9%), and the more globally symptomatic (29.6%). The three groups were associated with different patterns of diffusion MRI differences compared with controls. The cognitively symptomatic subgroup showed predominantly central WM differences, the globally symptomatic subgroup exhibited more peripheral differences with right-hemisphere predominance and sparing the corpus callosum, marked by reduced fractional anisotropy and kurtosis and elevated diffusivities, the less symptomatic subgroup demonstrated focal differences in the callosal genu, with increased fractional anisotropy and kurtosis and decreased diffusivity measures. CONCLUSIONS:MTBI comprises biologically distinct phenotypes with subgroup-specific WM signatures on diffusion MRI. Even individuals with minimal to no symptoms show WM differences compared with controls, underscoring the limitations of symptom reporting alone. Integrating symptom-based classification with advanced diffusion MRI may improve diagnostic precision to help risk stratification and provide insight into mechanisms of injury. ABBREVIATIONS/BACKGROUND:LCA = latent class analysis; MTBI = mild traumatic brain injury; RPQ = Rivermead post-concussion symptoms questionnaire.

BACKGROUND AND PURPOSE/OBJECTIVE:Na) MRI provides unique information about ionic homeostasis in the brain. However, in vivo quantification of regional brain sodium is highly challenging due to low SNR and limited spatial resolution. Here, we employ our novel anatomically guided reconstruction (AGR) method to overcome these challenges and enable precise quantification of regional brain total sodium concentration (TSC). MATERIALS AND METHODS/METHODS:< 0.05. RESULTS:. CONCLUSIONS:The AGR helps sodium quantification in healthy human brains by reducing the partial volume effect and variance of TSC in non-cortical brain regions. Our normative values of TSC in the brain regions set the stage to better understand derangements of sodium metabolism and homeostasis in neurological disease. ABBREVIATIONS/BACKGROUND:= sodium-potassium pump; PVC= partial volume correction; PVE= partial volume effect; TSC= total sodium concentration; VH= vitreous humor.