Deformable registration is essential in medical image analysis, especially for handling various multi- and mono-modal registration tasks in neuroimaging. Existing studies lack exploration of brain MR-CT registration, and face challenges in both accuracy and efficiency improvements of learning-based methods. To enlarge the practice of multi-modal registration in brain, we present SR-Reg, a new benchmark dataset comprising 180 volumetric paired MR-CT images and annotated anatomical regions. Building on this foundation, we introduce MambaMorph, a novel deformable registration network based on an efficient state space model Mamba for global feature learning, with a fine-grained feature extractor for low-level embedding. Experimental results demonstrate that MambaMorph surpasses advanced ConvNet-based and Transformer-based networks across several multi- and mono-modal tasks, showcasing impressive enhancements of efficacy and efficiency. Code and dataset are available at https://github.com/mileswyn/MambaMorph.

To compare clinical ratings and signal-to-noise ratio (SNR) measures of a commercially available Deep Learning-based MRI reconstruction method (T2_(DR)) against conventional T2- turbo spin echo brain MRI (T2_(CN)). 100 consecutive patients with various neurological conditions underwent both T2_(DR) and T2_(CN) on a Siemens Vida 3 T scanner with a 64-channel head coil in the same examination. Acquisition times were 3.33 min for T2_(CN) and 1.04 min for T2_(DR). Four neuroradiologists evaluated overall image quality (OIQ), diagnostic safety (DS), and image artifacts (IA), blinded to the acquisition mode. SNR and SNR_eff (adjusted for acquisition time) were calculated for air, grey- and white matter, and cerebrospinal fluid. The mean patient age was 43.6 years (SD 20.3), with 54 females. The distribution of non-diagnostic ratings did not differ significantly between T2_(CN) and T2_(DR) (IA p = 0.108; OIQ: p = 0.700 and DS: p = 0.652). However, when considering the full spectrum of ratings, significant differences favouring T2_(CN) emerged in OIQ (p = 0.003) and IA (p < 0.001). T2_(CN) had higher SNR (157.9, SD 123.4) than T2_(DR) (112.8, SD 82.7), p < 0.001, but T2_(DR) demonstrated superior SNR_eff (14.1, SD 10.3) compared to T2_(CN) (10.8, SD 8.5), p < 0.001. Our results suggest that while T2_(DR) may be clinically applicable for a diagnostic setting, it does not fully match the quality of high-standard conventional T2_(CN), MRI acquisitions.

Vascular cognitive impairment (VCI) is an umbrella term for diseases associated with cognitive decline induced by substantive brain damage following pathological changes in the cerebrovascular system. The primary clinical manifestations include behavioral abnormalities and diminished learning and memory cognitive functions. If the location and extent of brain injury are not identified early and therapeutic interventions are not promptly administered, it may lead to irreversible cognitive impairment. Therefore, the early diagnosis of VCI is crucial for its prevention and treatment. Prior to the onset of cognitive impairment in VCI, magnetic resonance imaging (MRI) radiomics can be utilized for early assessment and diagnosis, thereby guiding clinicians in providing precise treatment for patients, which holds significant potential for development. This article reviews the classification of VCI, the concept of radiomics, the application of MRI radiomics in VCI, and the limitations of radiomics in the context of advancements in its application within the central nervous system. CRITICAL RELEVANCE STATEMENT: This article explores how MRI radiomics can be used to detect VCI early, enhancing clinical radiology practice by offering a reliable method for prediction, diagnosis, and identification, which also promotes standardization in research and integration of disciplines. KEY POINTS: MRI radiomics can predict VCI early. MRI radiomics can diagnose VCI. MRI radiomics distinguishes VCI from Alzheimer's disease.

Brain lesion segmentation is challenging in medical image analysis, aiming to delineate lesion regions precisely. Deep learning (DL) techniques have recently demonstrated promising results across various computer vision tasks, including semantic segmentation, object detection, and image classification. This paper offers an overview of recent DL algorithms for brain tumor and stroke segmentation, drawing on literature from 2021 to 2024. It highlights the strengths, limitations, current research challenges, and unexplored areas in imaging-based brain lesion classification based on insights from over 250 recent review papers. Techniques addressing difficulties like class imbalance and multi-modalities are presented. Optimization methods for improving performance regarding computational and structural complexity and processing speed are discussed. These include lightweight neural networks, multilayer architectures, and computationally efficient, highly accurate network designs. The paper also reviews generic and latest frameworks of different brain lesion detection techniques and highlights publicly available benchmark datasets and their issues. Furthermore, open research areas, application prospects, and future directions for DL-based brain lesion classification are discussed. Future directions include integrating neural architecture search methods with domain knowledge, predicting patient survival levels, and learning to separate brain lesions using patient statistics. To ensure patient privacy, future research is anticipated to explore privacy-preserving learning frameworks. Overall, the presented suggestions serve as a guideline for researchers and system designers involved in brain lesion detection and stroke segmentation tasks.

Meningioma consistency critically impacts surgical planning, as soft tumors are easier to resect than hard tumors. Current assessments of tumor consistency using MRI are subjective and lack quantitative accuracy. Integrating deep learning and radiomics could enhance the predictive accuracy of meningioma consistency. A retrospective study analyzed 204 meningioma patients from two centers: the Second Affiliated Hospital of Guangzhou Medical University and the Southern Theater Command Hospital PLA. Three models-a radiomics model (Rad_Model), a deep learning model (DL_Model), and a combined model (DLR_Model)-were developed. Model performance was evaluated using AUC, accuracy, sensitivity, specificity, and precision. The DLR_Model outperformed other models across all cohorts. In the training set, it achieved AUC 0.957, accuracy of 0.908, and precision of 0.965. In the external test cohort, it maintained superior performance with an AUC of 0.854, accuracy of 0.778, and precision of 0.893, surpassing both the Rad_Model (AUC = 0.768) and DL_Model (AUC = 0.720). Combining radiomics and deep learning features improved predictive performance and robustness. Our study introduced and evaluated a deep learning radiomics model (DLR-Model) to accurately predict the consistency of meningiomas, which has the potential to improve preoperative assessments and surgical planning.

Reconstructive flap surgery aims to restore the substance and function losses associated with tumor resection. Automatic flap segmentation could allow quantification of flap volume and correlations with functional outcomes after surgery or post-operative RT (poRT). Flaps being ectopic tissues of various components (fat, skin, fascia, muscle, bone) of various volume, shape and texture, the anatomical modifications, inflammation and edema of the postoperative bed make the segmentation task challenging. We built a artificial intelligence-enabled automatic soft-tissue flap segmentation method from CT scans of Head and Neck Cancer (HNC) patients. Ground-truth flap segmentation masks were delineated by two experts on postoperative CT scans of 148 HNC patients undergoing poRT. All CTs and flaps (free or pedicled, soft tissue only or bone) were kept, including those with artefacts, to ensure generalizability. A deep-learning nnUNetv2 framework was built using Hounsfield Units (HU) windowing to mimic radiological assessment. A transformer-based 2D "Segment Anything Model" (MedSAM) was also built and fine-tuned to medical CTs. Models were compared with the Dice Similarity Coefficient (DSC) and Hausdorff Distance 95th percentile (HD95) metrics. Flaps were in the oral cavity (N = 102), oropharynx (N = 26) or larynx/hypopharynx (N = 20). There were free flaps (N = 137), pedicled flaps (N = 11), of soft tissue flap-only (N = 92), reconstructed bone (N = 42), or bone resected without reconstruction (N = 40). The nnUNet-windowing model outperformed the nnUNetv2 and MedSam models. It achieved mean DSCs of 0.69 and HD95 of 25.6 mm using 5-fold cross-validation. Segmentation performed better in the absence of artifacts, and rare situations such as pedicled flaps, laryngeal primaries and resected bone without bone reconstruction (p < 0.01). Automatic flap segmentation demonstrates clinical performances that allow to quantify spontaneous and radiation-induced volume shrinkage of flaps. Free flaps achieved excellent performances; rare situations will be addressed by fine-tuning the network.

Brain tumor causes life-threatening consequences due to which its timely detection and accurate classification are critical for determining appropriate treatment plans while focusing on the improved patient outcomes. However, conventional approaches of brain tumor diagnosis, such as Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) scans, are often labor-intensive, prone to human error, and completely reliable on expertise of radiologists.Thus, the integration of advanced techniques such as Machine Learning (ML) and Deep Learning (DL) has brought revolution in the healthcare sector due to their supporting features or properties having ability to analyze medical images in recent years, demonstrating great potential for achieving accurate and improved outcomes but also resulted in a few drawbacks due to their black-box nature. As understanding reasoning behind their predictions is still a great challenge for the healthcare professionals and raised a great concern about their trustworthiness, interpretability and transparency in clinical settings. Thus, an advanced algorithm of explainable artificial intelligence (XAI) has been synergized with hybrid model comprising of DenseNet201 network for extracting the most important features based on the input Magnetic resonance imaging (MRI) data following supervised algorithm, support vector machine (SVM) to distinguish distinct types of brain scans. To overcome this, an explainable hybrid framework has been proposed that integrates DenseNet201 for deep feature extraction with a Support Vector Machine (SVM) classifier for robust binary classification. A region-adaptive preprocessing pipeline is used to enhance tumor visibility and feature clarity. To address the need for interpretability, multiple XAI techniques-Grad-CAM, Integrated Gradients (IG), and Layer-wise Relevance Propagation (LRP) have been incorporated. Our comparative evaluation shows that LRP achieves the highest performance across all explainability metrics, with 98.64% accuracy, 0.74 F1-score, and 0.78 IoU. The proposed model provides transparent and highly accurate diagnostic predictions, offering a reliable clinical decision support tool. It achieves 0.9801 accuracy, 0.9223 sensitivity, 0.9909 specificity, 0.9154 precision, and 0.9360 F1-score, demonstrating strong potential for real-world brain tumor diagnosis and personalized treatment strategies.

Distinguishing gray matter (GM) from white matter (WM) is essential for CT of the brain. The recently established photon-counting detector CT (PCD-CT) technology employs a novel detection technique that might allow more precise measurement of tissue attenuation for an improved delineation of attenuation values (Hounsfield units - HU) and improved image quality in comparison with energy-integrating detector CT (EID-CT). To investigate this, we compared HU, GM vs. WM contrast, and image noise using automated deep learning-based brain segmentations. We retrospectively included patients who received either PCD-CT or EID-CT and did not display a cerebral pathology. A deep learning-based segmentation of the GM and WM was used to extract HU. From this, the gray-to-white ratio and contrast-to-noise ratio were calculated. We included 329 patients with EID-CT (mean age 59.8 ± 20.2 years) and 180 with PCD-CT (mean age 64.7 ± 16.5 years). GM and WM showed significantly lower HU in PCD-CT (GM: 40.4 ± 2.2 HU; WM: 33.4 ± 1.5 HU) compared to EID-CT (GM: 45.1 ± 1.6 HU; WM: 37.4 ± 1.6 HU, p < .001). Standard deviations of HU were also lower in PCD-CT (GM and WM both p < .001) and contrast-tonoise ratio was significantly higher in PCD-CT compared to EID-CT (p < .001). Gray-to-white matter ratios were not significantly different across both modalities (p > .99). In an age-matched subset (n = 157 patients from both cohorts), all findings were replicated. This comprehensive comparison of HU in cerebral gray and white matter revealed substantially reduced image noise and an average offset with lower HU in PCD-CT while the ratio between GM and WM remained constant. The potential need to adapt windowing presets based on this finding should be investigated in future studies. CNR = Contrast-to-Noise Ratio; CTDIvol = Volume Computed Tomography Dose Index; EID = Energy-Integrating Detector; GWR = Gray-to-White Matter Ratio; HU = Hounsfield Units; PCD = Photon-Counting Detector; ROI = Region of Interest; VMI = Virtual Monoenergetic Images.

Chemical exchange saturation transfer (CEST) MRI is a powerful molecular imaging technique for detecting metabolites through proton exchange. While CEST MRI provides high sensitivity, its clinical application is hindered by prolonged scan time due to the need for imaging across numerous frequency offsets for parameter estimation. Since scan time is directly proportional to the number of frequency offsets, identifying and selecting the most informative frequency can significantly reduce acquisition time. We propose a novel deep learning-based framework that integrates frequency selection and parameter estimation to accelerate CEST MRI. Our method leverages channel pruning via batch normalization to identify the most informative frequency offsets while simultaneously training the network for accurate parametric map prediction. Using data from six healthy volunteers, channel pruning selects 13 informative frequency offsets out of 53 without compromising map quality. Images from selected frequency offsets were reconstructed using the MR Multitasking method, which employs a low-rank tensor model to enable under-sampling of k-space lines for each frequency offset, further reducing scan time. Predicted parametric maps of amide proton transfer (APT), nuclear overhauser effect (NOE), and magnetization transfer (MT) based on these selected frequencies were comparable in quality to maps generated using all frequency offsets, achieving superior performance compared to Fisher information-based selection methods from our previous work. This integrated approach has the potential to reduce the whole-brain CEST MRI scan time from the original 5:30 min to under 1:30 min without compromising map quality. By leveraging deep learning for frequency selection and parametric map prediction, the proposed framework demonstrates its potential for efficient and practical clinical implementation. Future studies will focus on extending this method to patient populations and addressing challenges such as B₀ inhomogeneity and abnormal signal variation in diseased tissues.

Effective treatment for brain tumors relies on accurate detection because this is a crucial health condition. Medical imaging plays a pivotal role in improving tumor detection and diagnosis in the early stage. This study presents two approaches to the tumor detection problem focusing on the healthcare domain. A combination of image processing, vision transformer (ViT), and machine learning algorithms is the first approach that focuses on analyzing medical images. The second approach is the parallel model integration technique, where we first integrate two pre-trained deep learning models, ResNet101, and Xception, followed by applying local interpretable model-agnostic explanations (LIME) to explain the model. The results obtained an accuracy of 98.17% for the combination of vision transformer, random forest and contrast-limited adaptive histogram equalization and 99. 67% for the parallel model integration (ResNet101 and Xception). Based on these results, this paper proposed the deep learning approach-parallel model integration technique as the most effective method. Future work aims to extend the model to multi-class classification for tumor type detection and improve model generalization for broader applicability.