As the integration of artificial intelligence (AI) into radiology workflows continues to evolve, establishing standardized processes for the evaluation and deployment of AI models is crucial to ensure success. This paper outlines the creation of a Radiology AI Council at a large academic center and subsequent development of framework in the form of a rubric to formalize the evaluation of radiology AI models and onboard them into clinical workflows. The rubric aims to address the challenges faced during the deployment of AI models, such as real-world model performance, workflow implementation, resource allocation, return on investment (ROI), and impact to the broader health system. Using this comprehensive rubric, the council aims to ensure that the process for selecting AI models is both standardized and transparent. This paper outlines the steps taken to establish this rubric, its components, and initial results from evaluation of 13 models over an 8-month period. We emphasize the importance of holistic model evaluation beyond performance metrics, and transparency and objectivity in AI model evaluation with the goal of improving the efficacy and safety of AI models in radiology.

Human epidermal growth factor receptor 2 (HER2) is a crucial determinant of breast cancer prognosis and treatment options. The study aimed to establish an MRI-based habitat model to quantify intratumoral heterogeneity (ITH) and evaluate its potential in predicting HER2 expression status. Data from 340 patients with pathologically confirmed invasive breast cancer were retrospectively analyzed. Two tasks were designed for this study: Task 1 distinguished between HER2-positive and HER2-negative breast cancer. Task 2 distinguished between HER2-low and HER2-zero breast cancer. We developed the ITH, deep learning (DL), and radiomics signatures based on the features extracted from dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI). Clinical independent predictors were determined by multivariable logistic regression. Finally, a combined model was constructed by integrating the clinical independent predictors, ITH signature, and DL signature. The area under the receiver operating characteristic curve (AUC) served as the standard for assessing the performance of models. In task 1, the ITH signature performed well in the training set (AUC = 0.855) and the validation set (AUC = 0.842). In task 2, the AUCs of the ITH signature were 0.844 and 0.840, respectively, which still showed good prediction performance. In the validation sets of both tasks, the combined model exhibited the best prediction performance, with AUCs of 0.912 and 0.917 respectively, making it the optimal model. A combined model integrating clinical independent predictors, ITH signature, and DL signature can predict HER2 expression status preoperatively and noninvasively.

The 28th Annual Scientific Sessions of the Society for Cardiovascular Magnetic Resonance (SCMR) took place from January 29 to February 1, 2025, in Washington, D.C. SCMR 2025 brought together a diverse group of 1714 cardiologists, radiologists, scientists, and technologists from more than 80 countries to discuss emerging trends and the latest developments in cardiovascular magnetic resonance (CMR). The conference centered on the theme "Leading the Way to Accessible, Sustainable, and Efficient CMR," highlighting innovations aimed at making CMR more clinically efficient, widely accessible, and environmentally sustainable. The program featured 728 abstracts and case presentations with an acceptance rate of 86% (728/849), including Early Career Award abstracts, oral abstracts, oral cases and rapid-fire sessions, covering a broad range of CMR topics. It also offered engaging invited lectures across eight main parallel tracks and included four plenary sessions, two gold medalists, and one keynote speaker, with a total of 826 faculty participating. Focused sessions on accessibility, efficiency, and sustainability provided a platform for discussing current challenges and exploring future directions, while the newly introduced CMR Innovations Track showcased innovative session formats and fostered greater collaboration between researchers, clinicians, and industry. For the first time, SCMR 2025 also offered the opportunity for attendees to obtain CMR Level 1 Training Verification, integrated into the program. Additionally, expert case reading sessions and hands-on interactive workshops allowed participants to engage with real-world clinical scenarios and deepen their understanding through practical experience. Key highlights included plenary sessions on a variety of important topics, such as expanding boundaries, health equity, women's cardiovascular disease and a patient-clinician testimonial that emphasized the profound value of patient-centered research and collaboration. The scientific sessions covered a wide range of topics, from clinical applications in cardiomyopathies, congenital heart disease, and vascular imaging to women's heart health and environmental sustainability. Technical topics included novel reconstruction, motion correction, quantitative CMR, contrast agents, novel field strengths, and artificial intelligence applications, among many others. This paper summarizes the key themes and discussions from SCMR 2025, highlighting the collaborative efforts that are driving the future of CMR and underscoring the Society's unwavering commitment to research, education, and clinical excellence.

Bayesian neural networks (BNNs) treat neural network weights as random variables, which aim to provide posterior uncertainty estimates and avoid overfitting by performing inference on the posterior weights. However, the selection of appropriate prior distributions remains a challenging task, and BNNs may suffer from catastrophic inflated variance or poor predictive performance when poor choices are made for the priors. Existing BNN designs apply different priors to weights, while the behaviours of these priors make it difficult to sufficiently shrink noisy signals or they are prone to overshrinking important signals in the weights. To alleviate this problem, we propose a novel R2D2-Net, which imposes the R^2-induced Dirichlet Decomposition (R2D2) prior to the BNN weights. The R2D2-Net can effectively shrink irrelevant coefficients towards zero, while preventing key features from over-shrinkage. To approximate the posterior distribution of weights more accurately, we further propose a variational Gibbs inference algorithm that combines the Gibbs updating procedure and gradient-based optimization. This strategy enhances stability and consistency in estimation when the variational objective involving the shrinkage parameters is non-convex. We also analyze the evidence lower bound (ELBO) and the posterior concentration rates from a theoretical perspective. Experiments on both natural and medical image classification and uncertainty estimation tasks demonstrate satisfactory performance of our method.

Substantial gaps exist in the neuroprognostication of cardiac arrest patients who remain comatose after the restoration of spontaneous circulation. Most studies focus on predicting survival, a measure confounded by the withdrawal of life-sustaining treatment decisions. Severe cerebral edema (SCE) may serve as an objective proximal imaging-based surrogate of neurologic injury. We retrospectively analyzed data from 288 patients to automate SCE detection with machine learning (ML) and to test the hypothesis that the quantitative values produced by these algorithms (ML_SCE) can improve predictions of neurologic outcomes. Ground-truth SCE (GT_SCE) classification was based on radiology reports. The model attained a cross-validated testing accuracy of 87% [95% CI: 84%, 89%] for detecting SCE. Attention maps explaining SCE classification focused on cisternal regions (p<0.05). Multivariable analyses showed that older age (p<0.001), non-shockable initial cardiac rhythm (p=0.004), and greater ML_SCE values (p<0.001) were significant predictors of poor neurologic outcomes, with GT_SCE (p=0.064) as a non-significant covariate. Our results support the feasibility of automated SCE detection. Future prospective studies with standardized neurologic assessments are needed to substantiate the utility of quantitative ML_SCE values to improve neuroprognostication.

Magnetic resonance (MR) tagging is an imaging technique for noninvasively tracking tissue motion in vivo by creating a visible pattern of magnetization saturation (tags) that deforms with the tissue. Due to longitudinal relaxation and progression to steady-state, the tags and tissue brightnesses change over time, which makes tracking with optical flow methods error-prone. Although Fourier methods can alleviate these problems, they are also sensitive to brightness changes as well as spectral spreading due to motion. To address these problems, we introduce the brightness-invariant tracking estimation (BRITE) technique for tagged MRI. BRITE disentangles the anatomy from the tag pattern in the observed tagged image sequence and simultaneously estimates the Lagrangian motion. The inherent ill-posedness of this problem is addressed by leveraging the expressive power of denoising diffusion probabilistic models to represent the probabilistic distribution of the underlying anatomy and the flexibility of physics-informed neural networks to estimate biologically-plausible motion. A set of tagged MR images of a gel phantom was acquired with various tag periods and imaging flip angles to demonstrate the impact of brightness variations and to validate our method. The results show that BRITE achieves more accurate motion and strain estimates as compared to other state of the art methods, while also being resistant to tag fading.

Tumor-infiltrating lymphocytes (TILs) are capable of recognizing tumor antigens, impacting tumor prognosis, predicting the efficacy of neoadjuvant therapies, contributing to the development of new cell-based immunotherapies, studying the tumor immune microenvironment, and identifying novel biomarkers. Traditional methods for evaluating TILs primarily rely on histopathological examination using standard hematoxylin and eosin staining or immunohistochemical staining, with manual cell counting under a microscope. These methods are time-consuming and subject to significant observer variability and error. Recently, artificial intelligence (AI) has rapidly advanced in the field of medical imaging, particularly with deep learning algorithms based on convolutional neural networks. AI has shown promise as a powerful tool for the quantitative evaluation of tumor biomarkers. The advent of AI offers new opportunities for the automated and standardized assessment of TILs. This review provides an overview of the advancements in the application of AI for assessing TILs from multiple perspectives. It specifically focuses on AI-driven approaches for identifying TILs in tumor tissue images, automating TILs quantification, recognizing TILs subpopulations, and analyzing the spatial distribution patterns of TILs. The review aims to elucidate the prognostic value of TILs in various cancers, as well as their predictive capacity for responses to immunotherapy and neoadjuvant therapy. Furthermore, the review explores the integration of AI with other emerging technologies, such as single-cell sequencing, multiplex immunofluorescence, spatial transcriptomics, and multimodal approaches, to enhance the comprehensive study of TILs and further elucidate their clinical utility in tumor treatment and prognosis.

Stroke is among the top three causes of death worldwide, and accurate identification of stroke lesion boundaries is critical for diagnosis and treatment. Supervised deep learning methods have emerged as the leading solution for stroke lesion segmentation but require large, diverse, and annotated datasets. The ISLES'24 challenge addresses this need by providing longitudinal stroke imaging data, including CT scans taken on arrival to the hospital and follow-up MRI taken 2-9 days from initial arrival, with annotations derived from follow-up MRI. Importantly, models submitted to the ISLES'24 challenge are evaluated using only CT inputs, requiring prediction of lesion progression that may not be visible in CT scans for segmentation. Our winning solution shows that a carefully designed preprocessing pipeline including deep-learning-based skull stripping and custom intensity windowing is beneficial for accurate segmentation. Combined with a standard large residual nnU-Net architecture for segmentation, this approach achieves a mean test Dice of 28.5 with a standard deviation of 21.27.

Medical image segmentation, particularly in multi-domain scenarios, requires precise preservation of anatomical structures across diverse representations. While deep learning has advanced this field, existing models often struggle with accurate boundary representation, variability in organ morphology, and information loss during downsampling, limiting their accuracy and robustness. To address these challenges, we propose the Context Enhancement Network (CENet), a novel segmentation framework featuring two key innovations. First, the Dual Selective Enhancement Block (DSEB) integrated into skip connections enhances boundary details and improves the detection of smaller organs in a context-aware manner. Second, the Context Feature Attention Module (CFAM) in the decoder employs a multi-scale design to maintain spatial integrity, reduce feature redundancy, and mitigate overly enhanced representations. Extensive evaluations on both radiology and dermoscopic datasets demonstrate that CENet outperforms state-of-the-art (SOTA) methods in multi-organ segmentation and boundary detail preservation, offering a robust and accurate solution for complex medical image analysis tasks. The code is publicly available at https://github.com/xmindflow/cenet.

This paper presents a novel method for monocular patient-to-image intraoperative registration, specifically designed to operate without any external hardware tracking equipment or fiducial point markers. Leveraging a synthetic microscopy surgical scene dataset with a wide range of transformations, our approach directly maps preoperative CT scans to 2D intraoperative surgical frames through a lightweight neural network for real-time cochlear implant surgery guidance via a zero-shot learning approach. Unlike traditional methods, our framework seamlessly integrates with monocular surgical microscopes, making it highly practical for clinical use without additional hardware dependencies and requirements. Our method estimates camera poses, which include a rotation matrix and a translation vector, by learning from the synthetic dataset, enabling accurate and efficient intraoperative registration. The proposed framework was evaluated on nine clinical cases using a patient-specific and cross-patient validation strategy. Our results suggest that our approach achieves clinically relevant accuracy in predicting 6D camera poses for registering 3D preoperative CT scans to 2D surgical scenes with an angular error within 10 degrees in most cases, while also addressing limitations of traditional methods, such as reliance on external tracking systems or fiducial markers.