MRI Images Research Articles

This study aimed to determine the diagnostic yield of CT-guided bone biopsies in pediatric patients, the outcome of non-diagnostic CT biopsy results, and to establish factors associated with non-diagnostic biopsy results. This is a retrospective study of consecutive pediatric patients ≤ 21 years who underwent CT-guided bone biopsies in three tertiary referral hospitals from December 2011 to March 2022. Clinical information, pre-biopsy CT and MRI images, procedural details, pathological results, and follow-up were assessed. Fisher's exact test was used to compare categorical variables. Mann-Whitney U-test and unpaired t-test were used to compare non-parametric and parametric variables, respectively. Statistical significance was set a p < 0.05. A total of 138 patients (mean age 13.9 ± 4.5 years; 95 (60%) male patients) with 157 CT-guided bone biopsies were studied, which yielded 38.2% (60/157) non-diagnostic, 23.6% (37/157) benign, and 38.2% (60/157) malignant results. Most non-diagnostic lesions (88.3% [53/60]) were subsequently determined to be benign. Factors associated with non-diagnostic biopsy results were cystic lesions (p = 0.003) incidental lesions (p = 0.03), fewer (p = 0.02) and shorter (p = 0.01) tissue cores, non-aggressive radiological features, including narrow zone of transition (p < 0.001), sclerotic margin (p < 0.001), no cortical destruction (p < 0.001), no periosteal reaction (p < 0.001), or no extra-osseous soft tissue mass (p < 0.001). About one-third of CT-guided bone biopsies in pediatric patients yielded non-diagnostic results, though most were ultimately confirmed to be benign. In children and adolescents with suspected primary bone tumors, CT-guided bone biopsy with non-diagnostic histopathological results strongly favors benignity in lesions with non-aggressive imaging features and should align management towards a more conservative approach. Question Limited data exist on the prevalence and outcome of non-diagnostic CT-guided bone biopsy in pediatric patients to guide clinical management. Findings About one-third of CT-guided bone biopsies in pediatric patients were non-diagnostic, though most were ultimately benign. Lesions with non-aggressive features were associated with non-diagnostic results. Clinical relevance In the setting of multidisciplinary care for patients with suspected primary bone tumors, non-diagnostic CT-guided bone biopsy-particularly in lesions with non-aggressive imaging features-strongly favors benignity, which should steer management towards a more conservative approach.

Breast cancer imaging without gadolinium-based contrast agent: A review of current applications and future trends.

Brain problems lead to the loss of physical functions like speech and movement. Thus, early brain tumour diagnosis is fundamental for improving the survival of patients. Existing traditional methods follow deep neural structural design where the selection of relevant characteristics descriptors and classifiers is a main challenge. Therefore, the deep learning-based recognition of various abnormalities in the brain has been suggested. Initially, the required brain image is taken from the public dataset. The image data are then passed to the segmentation process, in which the adaptive refinement network (ARN) performs the segmentation as it is robust to outliers and can manage the intricate structure of tumours. Further, enhance the segmentation process by implementing the fitness-based flamingo search algorithm (FFSA), which optimizes the parameters in the segmentation model by efficiently exploring the search area and converging on the most favourable solutions. The resultant segmented images are sent to an ensemble convolutional neural network (CNN) with Bayesian learning (ECNN-BL) for classification. By combining several systems, ensembles can overcome overfitting issues, which lead to better generalization to new data and improved accuracy and robustness. Here, the ensemble CNN is the combination of the visual geometry group-16 (VGG16), residual neural network (Resnet), and Xception that performs effective classification. The superiority of the developed model is determined by taking a similar analysis with existing approaches. From the findings, the designed system is dependable and efficient in identifying brain diseases using the MRI images.

Ovarian cancer (OC) and endometrial cancer (EC) are highly heterogeneous gynecological malignancies with distinct molecular subtypes, therapeutic responses, and clinical outcomes. Traditional biopsy-based profiling often fails to capture the spatial and temporal complexity of these tumors. Radiogenomics, integrating imaging features with genomic and molecular data, has emerged as a promising approach to non-invasively analyze tumor heterogeneity. The purpose of this abstract is to critically examine and synthesize existing research on the application of radiogenomics in OC and EC, focusing on its ability to correlate imaging phenotypes with molecular biomarkers. This narrative review aims to demonstrate how radiogenomics can enhance tumor characterization, support biomarker prediction, and inform prognosis and therapeutic decision-making with non-invasive methods.This narrative review critically synthesizes current literature on radiogenomics applications in OC and EC. Studies using CT, MRI, and PET imaging were evaluated for their ability to correlate imaging phenotypes with molecular biomarkers, gene expression profiles, and clinical outcomes. The analysis emphasizes the role of radiogenomics in enhancing tumor characterization, predicting biomarker status, forecasting treatment response and prognosis.Radiogenomics has successfully identified associations between imaging features and key molecular alterations, such as BRCA mutations, homologous recombination deficiency (HRD), and immune-related biomarkers in OC, as well as POLE mutations, microsatellite instability (MSI), and tumor mutational burden (TMB) in EC. Predictive models incorporating radiomic features have demonstrated notable performance in estimating prognosis, treatment response, and recurrence risk across both cancer types.Radiogenomics has a strong potential to enhance personalized cancer care by analyzing tumor heterogeneity. However, clinical application requires methodological standardization, prospective validation, and integration into precision oncology workflows. · Radiogenomics enables non-invasive assessment of spatial and molecular heterogeneity in OC and EC.. · Integration of imaging and genomic data improves prediction of biomarkers, therapy response, and survival.. · Future clinical applications depend on methodological standardization, prospective validation, and integration into precision oncology workflows.. · Dolciami M, Celli V, Panico C et al. Unraveling Tumor Heterogeneity in Gynecological Cancer Using a Radiogenomics Approach Rofo 2025; DOI 10.1055/a-2698-8545.

Intervertebral disc degeneration (IDD) is a major cause of chronic lower back pain and associated spinal disorders. Animal models play a crucial role in elucidating the pathophysiological mechanisms underlying IDD and in the development of potential treatments. Various techniques have been used to induce IDD, including annulus fibrosus puncture, nucleus pulposus aspiration (NPA), and chemical injections. However, few studies have explored the use of NPA combined with puncture needle burning (PNB) or anhydrous ethanol injection (AEI) to induce IDD. We compared the efficacy and consistency of three induction methods - NPA, NPA + PNB, and NPA + AEI - in rabbits. The extent of degeneration was assessed using MRI, X-ray, and histological analyses. Twenty-four male New Zealand white rabbits (weighing 3.5-4.0 kg) were randomly allocated to three groups (n = 8 per group). Degeneration was induced in the L2/3, L3/4, L4/5, and L5/6 intervertebral discs using one of the three methods: NPA, NPA + PNB, or NPA + AEI. The L6/7 discs served as the internal control. Four weeks post-procedure, the degree of disc degeneration was evaluated via MRI and X-ray imaging. Histological assessment was performed using hematoxylin and eosin and safranin O/fast green staining. The severity of degeneration was quantified using the Masuda histological scoring system. All three methods successfully induced significant IDD, as confirmed by imaging and histological analyses. The NPA + PNB group had the most severe and uniform degeneration, consistently corresponding to Pfirrmann grade IV. The NPA + AEI group had a comparable severity of degeneration but with more inter-segmental variability. In contrast, the NPA-only group had milder degeneration, predominantly within Pfirrmann grades II-III. All experimental groups had significant reductions in terms of intervertebral disc height and nucleus pulposus area, with the most pronounced reductions observed in the NPA + PNB group. MRI indices and histological scores consistently indicated that the NPA + PNB method produced the most severe and reproducible degeneration. Histological staining revealed decreased cellularity, fissures in the annulus fibrosus, and collapse of the cartilage endplate in all intervention groups, with the NPA + PNB group exhibiting the most extensive degeneration. These findings confirm that NPA, NPA + PNB, and NPA + AEI are effective techniques for inducing IDD in a rabbit model. Among these, the NPA + PNB method represents one of the best options for achieving a reliable, reproducible, and minimally invasive model, producing consistent and severe degeneration with a high degree of standardization. This technique presents a valuable tool for future research investigating the pathogenesis of IDD and evaluating the efficacy of potential therapeutic strategies.

In Multiple Sclerosis, inflammation and neurodegeneration disrupt structural and functional brain networks. While the association between structural connectivity and disability is rather clear, functional connectivity changes are not yet characterised as a physiological response to the disease, as functionally meaningful adaptation or as a deceptive response. We explored the topology of brain networks of 65 Multiple Sclerosis patients over up to seven years in comparison to 59 controls. Connectomes based on probabilistic tractography from diffusion weighted imaging and resting-state MRI, were analysed with graph theory. The hub disruption index estimated connectivity perturbation in relation to the network hierarchy. In controls, we observed a transient increase in functional hub connectivity in the 5th and 6th age decade as a response to a subtle diffuse loss of structural connectivity, before structural and functional connectomes show a pronounced loss of hub connectivity. In Multiple Sclerosis, structural hub disruption was present from the disease onset while the transient upregulation of functional hub connectivity in the middle age was lacking. Patients seem to transition directly into an exhausted hub connectivity configuration. However, we observed the transient functional reorganisation of hubs in the first years after disease onset. Multiple Sclerosis patients present a probable physiological response to structural connectivity loss very early in the disease, potentially leading to an accelerated hub overload with accelerated neurodegeneration. The onset of chronic progression in the 5th age decade might be partially driven by the absence of the physiological increased hub connectivity observed in healthy individuals.

This study addresses the common challenges in medical image segmentation and recognition, including boundary ambiguity, scale variation, and the difficulty of modeling long-range dependencies, by proposing a unified framework based on a hierarchical attention mechanism. The framework consists of a local detail attention module, a global context attention module, and a cross-scale consistency constraint module, which collectively enable adaptive weighting and collaborative optimization across different feature levels, thereby achieving a balance between detail preservation and global modeling. The framework was systematically validated on multiple public datasets, and the results demonstrated that the proposed method achieved Dice, IoU, Precision, Recall, and F1 scores of 0.886, 0.781, 0.898, 0.875, and 0.886, respectively, on the combined dataset, outperforming traditional models such as U-Net, Mask R-CNN, DeepLabV3+, SegNet, and TransUNet. On the BraTS dataset, the proposed method achieved a Dice score of 0.922, Precision of 0.930, and Recall of 0.915, exhibiting superior boundary modeling capability in complex brain MRI images. On the LIDC-IDRI dataset, the Dice score and Recall were improved from 0.751 and 0.732 to 0.822 and 0.807, respectively, effectively reducing the missed detection rate of small nodules compared to traditional convolutional models. On the ISIC dermoscopy dataset, the proposed framework achieved a Dice score of 0.914 and a Precision of 0.922, significantly improving the accuracy of skin lesion recognition. The ablation study further revealed that local detail attention significantly enhanced boundary and texture modeling, global context attention strengthened long-range dependency capture, and cross-scale consistency constraints ensured the stability and coherence of prediction results. From a medical economics perspective, the proposed framework has the potential to reduce diagnostic costs and improve healthcare efficiency by enabling faster and more accurate image-based clinical decision-making. In summary, the hierarchical attention mechanism presented in this work not only provides an innovative breakthrough in mathematical modeling but also demonstrates outstanding performance and generalization ability in experiments, offering new perspectives and technical pathways for intelligent segmentation and recognition in medical imaging.

In this study, we propose multiple deep learning models for classifying gray matter MRI images of healthy individuals, prodromal Parkinson’s disease (PD) subjects, and diagnosed PD patients. The two proposed models extend conventional deep learning architectures—MedicalNet3D and 3D ResNet18—by performing feature extraction separately for each class and inputting these features into distinct multilayer perceptron (MLP) classifiers constructed via fine-tuning. To mitigate overfitting problem and improve generalizability, we introduce a training method based on group-wise feature fusion, in which subject IDs are separated to avoid data leakage during training. Through comparative experiments using the PPMI database, the effectiveness of the proposed approach was validated.

To compare two identically sized wireless resonator designs, one with strongly coupled units and the other with decoupled units, for their ability to enhance receive performance in MRI when used with local receive arrays. Both wireless resonator designs were fabricated and experimentally evaluated for detuning efficiency, SNR improvements, and parallel imaging performance (g-factor) at 1.5 T. They were used alongside a 12-channel head receive array, with the standard body coil serving as the RF transmitter. Experimental data showed that the wireless resonator with decoupled units consistently outperformed that with coupled units, with up to threefold improvement in SNR and a reduction of maximum/average g-factor from 4.6/1.8 to 3.1/1.3. Notably, compared to the original receive array (maximum/average: 3.9/1.7), the decoupled design further improved the g-factor, highlighting superior performance in accelerated imaging. Wireless resonators with decoupled units offer significant advantages in improving MRI image quality and parallel imaging performance over their coupled counterparts. Their ease of detuning and pronounced gains in SNR and g-factor make them a compelling choice for wireless resonator designs.

Background: The long-term survival of heart transplantation (HTx) recipients is influenced by a range of cardiovascular, immunological, and procedural factors. Accurately predicting post-HTx outcomes remains a major clinical challenge, especially when relying solely on noninvasive methods. Objective: To test the hypothesis that structural and functional indices derived from multi-parametric cardiac MRI-derived can be used to predict cardiovascular events in HTx recipients. Materials and methods: With the approval of institutional review board (IRB), 170 HTx recipients (106 males, age: 47.8 ± 16 years, Range: 19 – 79 years) were recruited for a comprehensive multi-parametric cardiac MRI scan. MRI images were processed to derive global cardiac function and volumes, and myocardial T2 values and T1 values. Pre- and post-Gadolinium T1 was used to calculate extra-cellular volume (ECV) fraction. Cardiovascular events were defined as a composite of any emergency visit, hospitalization or death due to graft failure or reception, myocardial infarction, HF and other events that cannot rule out a cardiovascular origin of complications. Identification of predictors of adverse outcomes at long-term follow-up was based on a Cox proportional hazards model (CPH). Statistical analysis was performed by using SPSS (version 22.0). Results: MRI images were eligilbe for quantitative analysis. See figure 1. The patients were followed for 6 to 4504 days (Median = 2616 days) after multi-parametric cardiac MRI. In total, 140 cardiovascular events occurred (6 to 3294 days, Median = 627 days). The CPH model fits the data (p &lt; 0.001). After the adjustment of traditional cardiovascular risk factors and demographic data, multiple MRI-derived indices were identified as significant predictors of survival time (time between baseline cardiac MRI and adverse event), including left ventricular (LV) end-diastolic volume (LVEDV) (p &lt; 0.001), LV end-systolic volume (LVESV) (p &lt; 0.001), LV stroke volume (LVSV) (p = 0.005), right ventricular (RV) stroke volume (RVSV) (p &lt; 0.001), RV cardiac output (RVCO) ( p = 0.03), myocardial ECV (p &lt; 0.001) and T2 value (p = 0.008). See figure 2. Conclusions: Multi-parametric indices of cardiac tissue (T2, ECV) and function (LVEDV, LVESV, LVSV, RVSV, RVCO) can independently predict adverse clinical outcomes in HTx recipients at long term follow-up (median &gt; 7 years). MRI may offer new imaging biomarkers for early identification of risks for post-HTx complication.

Background: With the advent of stereotactic radioablation for cardiac arrhythmias, accurate non-contact mapping tools are increasingly important. Magnetocardiography (MCG) is promising but historically limited by supercooled sensors, extensive shielding, and long recording durations. A novel magnetic sensor system developed by TDK Corporation may overcome these prior limitations, but rigorous validation against established methods has not been performed. Specifically, no prior study has directly validated this novel MCG system through a three-way comparison with electrocardiographic imaging (ECGi) and a gold-standard pacing location for atrial arrhythmia localization. Hypothesis: We hypothesized that the novel MCG sensor system would perform comparably to ECGi in accurately localizing atrial activation origins, enabling the creation of reliable atrial activation maps. Methods: Six swine (42.2 ± 5.3 kg) underwent placement of pacing wires in the right atrium to simulate focal atrial arrhythmias. Anatomical MRI scans precisely defined the gold-standard pacing lead position (Fig. A) and ECG electrode locations via fiducials. Atrium anatomy was segmented from MRI images and smoothed to create accurate anatomical models. Simultaneous MCG and ECGi signals were recorded during controlled atrial pacing. Latency maps were generated from denoised, beat-averaged signals (Fig. B). Localization accuracy between MCG and ECGi was compared using a paired Wilcoxon signed-rank test. Results: Approximately 1000 P-waves per animal were analyzed. Median absolute localization error was 24.1 mm (IQR 18.6–30.2 mm) for ECGi and 31.0 mm (IQR 23.2–37.3 mm) for MCG (p=ns; Fig. C). Although localization error was numerically higher for MCG, differences were not statistically significant given the limited sample size. Conclusions: Our preliminary results demonstrate the feasibility of using a novel, solid-state MCG sensor system for non-invasive atrial arrhythmia localization. The difference in localization accuracy between ECGi and MCG was not statistically significant in this initial animal cohort. This first-of-its-kind multimodal validation suggests that novel MCG technology may serve as a viable complementary mapping modality, warranting further validation in larger studies.

Deep O-SegNet based Stroke Lesion Segmentation and HX-ShuffleNet with Optimization for Stroke Classification Using MRI Image

Background: Transthyretin amyloid cardiomyopathy (ATTR-CM) is a life-threatening condition, disease progression is currently measured using a combination of clinical, laboratory, MRI Imaging and ECG investigations multiple times a year, causing a significant disease burden to the patients. Simplifying investigations offers the potential to reducing disease burden, improve access and reduce health inequalities. Our novel mathematical method, Symmetric Projection Attractor Reconstruction (SPAR) analyses the complete morphology and variability of waveforms, This method has been shown to non-invasively assess vascular aging using PPG signals. SPAR amplifies small morphology changes on ECG which may be missed if focus resides on single points such as PQRST intervals or amplitudes, producing SPAR attractor images which can be readily interpreted by clinicians. Hypothesis: We hypothesised that SPAR analysis of ECG can sensitively detect aberrant changes on ECGs overtime to provide a new method to differentiate patients between stable and progression of their heart failure and to distinguish a threshold for significant progression. Methods: Longitudinal ECG and medical histories were collected and anonymised for ATTR-CM patients. ECG Waveform analysis was performed using a bespoke MATLAB software tool to transform ECG time-series into corresponding attractors. Results: Initial evaluation of a pilot analysis of a sub-set of the available data revealed attractors were able to see a stable attractor for patients who remained stable over 40 months. In contrast, patients who deteriorated, showed large attractor feature changes (fig1). Quantifying delta changes from the last ECG produces a cumulative graph indication a threshold for clinically concerning changes is possible(fig2). Conclusion: This pilot study illustrates that SPAR attractors highlight changes on the ECG that were consistent with deterioration in the patient’s clinical condition. Additional analysis for delta changes between attractor time points showed a disease trajectory. Further evaluation of the whole dataset to create a simplified diagnostic where patients could provide an ECG at home allowing more frequent clinical monitoring remotely negating the need for patients to undergo complex imagining and laboratory testing. Offering remote monitoring of patients to only bring deteriorating patients to clinic, reducing unecessary investigations and travel to health centres.

Introduction: Drug Reaction with Eosinophilia and Systemic Symptoms (DRESS) syndrome is a severe hypersensitivity reaction characterized by rash, fevers, eosinophilia and multi-organ involvement. Cardiac involvement in the form of eosinophilic myocarditis is particularly rare, with a mortality rate of 52% within 72 hours of admission. Case Presentation: A 34-year-old woman presented epigastric pain and rash following 10-day course of trimethoprim-sulfamethoxazole, metronidazole, and fluconazole for a vaginal infection. Initial evaluation revealed fever of 102°F, eosinophilia (0.7 × 10 9 /L), and elevated liver enzymes (AST 234 U/L, ALT 396 U/L, ALP 365 U/L). Skin biopsy was consistent with a drug-induced hypersensitivity reaction, and despite treatment with 1 mg/kg of prednisone, she developed chest pain, dyspnea, hypotension and elevated troponins (&gt;5000 ng/L). Electrocardiogram revealed diffuse ST-segment elevations, and echocardiograms demonstrated a significant reduction in left ventricular ejection fraction (55% to 25%) with global hypokinesis. Cardiac MRI (Image A/B) confirmed severe myocarditis with endomyocardial biopsy (Image C) revealing eosinophilic infiltration consistent with eosinophilic myocarditis. The patient progressed to cardiogenic shock requiring VA ECMO. Given the lack of improvement with medical management the decision was made to proceed with an orthotopic heart transplant. Post-transplant, she was managed with a regimen of immunosuppressants, and recovery was closely monitored with serial cardiac biopsies and imaging studies. Discussion: Cardiac involvement in DRESS syndrome is exceptionally rare but carries a high mortality rate when complicated by myocarditis. Early recognition and prompt management are crucial for improving outcomes. While systemic corticosteroids are the mainstay of treatment for DRESS, their efficacy in preventing myocarditis may be limited. Other immunomodulating therapies have been attempted with variable success. Our patient’s case highlights the limitations of current therapeutic options for severe DRESS-associated myocarditis. Despite aggressive medical management, progression to cardiogenic shock necessitated mechanical circulatory support and ultimately heart transplantation. Clinicians should maintain a high index of suspicion for cardiac involvement in patients with DRESS who present with hemodynamic instability or elevated cardiac biomarkers, even in the absence of classic cardiac symptoms.

Background: Hemorrhagic transformation (HT) is a common outcome of acute ischemic stroke (AIS), especially following thrombolytic or endovascular reperfusion therapy. Early detection of HT may guide therapeutic decisions and reduce risk. With the advancement of artificial intelligence in neuroimaging, several studies have investigated machine learning (ML) and radiomics models for predicting HT using imaging and clinical data. Objective: The purpose of this meta-analysis was to analyze the diagnostic efficacy of ML-based radiomic models for predicting HT after AIS and their generalizability via external validation. Methods: We conducted a systematic review and meta-analysis of works published until May 2025 using PubMed, EMBASE, Scopus, and IEEE Xplore. The inclusion criteria were studies that used ML-based radiomic or deep learning models with CT, MRI, or multimodal imaging to predict HT in AIS patients. A bivariate random-effects model was used to examine the pooled sensitivity, specificity, diagnostic odds ratio (DOR), and area under the summary receiver operating characteristic curve (AUC-SROC). The risk of bias was assessed using the QUADAS-2 method. Subgroup analyses were performed based on the imaging modality and algorithm type. An external validation cohort (n=1,150) was used to assess the generalizability of top-performing models. Results: 18 studies (n = 3,945 patients; ~455 with HT) met the inclusion criteria. The pooled sensitivity and specificity for machine learning-based radiomics models were 0.81 (95% CI: 0.78-0.84) and 0.84 (95% CI: 0.80-0.88), respectively, with a diagnostic odds ratio of 22.5 (95% CI: 15.0-33.8). The overall AUC-SROC value was 0.88 (95% CI: 0.85–0.91). MRI-based models outperformed CT-only models (AUC: 0.85; p = 0.01). yielded an AUC of 0.87, sensitivity of 0.83, and specificity of 0.82, indicating generalizability. There was no significant publication bias, and heterogeneity was large (I square = 46%). Conclusion: ML-based radiomics models that use multimodal neuroimaging show high prediction for hemorrhagic transition after AIS, with consistent results across modalities and external datasets. MRI-based models provide marginally higher diagnosis accuracy.

Background: Early identification of subclinical atherosclerosis permits timely preventive intervention to reduce cardiovascular events. Although carotid intima-media thickness (CIMT) and coronary artery calcium scoring (CACS) are established, novel molecular and hybrid imaging approaches may offer superior lesion characterization and risk stratification. Methods: Following PRISMA guidelines, we searched PubMed, Scopus, Embase, and Web of Science for studies published January 2010–December 2024 evaluating imaging modalities for subclinical atherosclerosis in asymptomatic or high-risk adults. Two reviewers independently screened titles/abstracts and full texts against predefined inclusion (original research or reviews reporting diagnostic performance or clinical utility of CIMT, CACS, MRI, PET, molecular, or hybrid imaging) and exclusion (symptomatic cohorts, animal studies, case reports, editorials) criteria, resolving discrepancies by consensus. Data were extracted via a standardized form—capturing study design, population, imaging technique details, plaque/molecular endpoints, and diagnostic metrics—and study quality was appraised using QUADAS-2. Given the heterogeneity of modalities and outcomes, findings were synthesized narratively by imaging modality. Results: Nine studies met inclusion criteria. CIMT and CACS demonstrated moderate sensitivity and specificity for detecting early atherosclerosis. Advanced MRI and PET-based methods provided enhanced plaque composition and inflammatory assessment. Molecular imaging probes and hybrid PET/MRI further improved detection of high-risk lesions but were limited by cost and accessibility. Conclusions: CIMT and CACS remain valuable for population screening, but emerging MRI, PET, molecular, and hybrid modalities offer deeper insights into plaque vulnerability. Standardization of imaging protocols and cost-effectiveness analyses are urgently needed to facilitate integration of these advanced techniques into routine clinical practice.

Aim Vertebral osteomyelitis is a severe complication of disseminated coccidioidomycosis. Coccidioidal vertebral osteomyelitis (CVO) and pyogenic(bacterial) vertebral osteomyelitis (PVO) can present similarly; given the difference in treatment, accurate diagnosis is key. We compared CT and MRI imaging findings between patients with CVO and PVO to determine if any difference exists. Method A multicenter hospital system located in the Southwestern United States was queried for patients diagnosed with disseminated coccidioidomycosis between 2013 and 2022, resulting in 370 records. These charts were manually assessed, and 42 patients were found to have vertebral dissemination (CVO). A similar query was performed in the same medical system for patients diagnosed for PVO, which resulted in 7977 patients. Eighty cases were then obtained, to give a final 2:1 ratio between PVO and CVO cases (matching done based on year of diagnosis and race). Each patient's imaging reports, CT and MRI, were then evaluated for the presence of specific radiological findings shown to be indicative of vertebral osteomyelitis. Results In our PVO cohort, 50 received CT scans and 68 received MRIs. For patients with CVO, 31 received CT scans and 34 received MRIs. Based on findings from either method of imaging, abscess was more common among CVO patients (81% vs 53%, p=0.003), as was bone destruction (83% vs. 65%, p=0.036); however, discitis was more commonly found in PVO patients (58% vs 39% p=0.046). Conclusion Imaging presents a potential data point for clinicians attempting to determine the etiology of a patient's vertebral osteomyelitis. Spinal CT with contrast showed efficacy in the detection of abscess and bone destruction and is a reasonable imaging study compared to MRI for resource-limited settings in Coccidioides endemic regions.

The early and accurate identification of brain abnormalities plays a vital role in improving patient outcomes and treatment planning [1], [2]. This project focuses on developing an intelligent medical image analysis system capable of detecting and classifying brain tumors automatically from MRI data [3], [4]. The proposed approach utilizes advanced three-dimensional convolutional neural network (3D-CNN) architectures that effectively capture spatial and contextual information from volumetric MRI images [5]–[7]. The system undergoes preprocessing steps such as skull stripping, normalization, and data augmentation to enhance input quality and model robustness [8], [9]. Through deep feature extraction and layer-wise learning, the model distinguishes between tumor and non-tumor regions with high precision [10], [11]. Experimental results demonstrate that the proposed deep learning framework outperforms conventional 2D models by leveraging 3D spatial relationships within the MRI scans [12]–[15]. This automated solution significantly reduces diagnostic time, assists radiologists in clinical decision-making, and contributes to improved brain healthcare through intelligent image-based diagnosis [16]–[19]. Furthermore, the integration of explainable AI techniques provides interpretability and transparency, which are crucial for clinical trust and real-world applicability [20], [25].

Background: Infective endocarditis (IE) is a life-threatening infection of the endocardial surface of the heart. Streptococcus agalactiae is a rare but aggressive cause of IE, typically associated with large vegetations and high embolic potential. Diagnosis can be challenging due to its rapid progression and non-specific symptoms. Case Description: A 64-year-old male with a history of type 2 diabetes mellitus and alcohol use disorder presented with one day of confusion. He was admitted to the ICU for encephalopathy, sepsis, and atrial fibrillation with rapid ventricular response. Initial encephalopathy workup ruled out toxic, structural, and metabolic causes. Lumbar puncture ruled out meningitis. Blood cultures grew Streptococcus agalactiae , and appropriate IV antibiotics were initiated. Given possible IE per Duke’s criteria, a transthoracic echocardiogram (TTE) was obtained but was unremarkable. The patient was stabilized and transferred to the general medicine ward. He remained encephalopathic with persistent leukocytosis, raising concern for occult untreated infection. A detailed physical examination revealed midline thoracic and lumbar spine tenderness and a swollen right wrist and hand. MRI imaging confirmed T8-T9 and L3-L4 discitis-osteomyelitis with spinal epidural abscesses, as well as right wrist septic arthritis. Given concern for septic emboli being the source of multifocal infection, further evaluation with transesophageal echocardiogram (TEE) was performed; this revealed a 1.4 x 1.6 cm vegetation on the posterior mitral leaflet, without significant regurgitation or stenosis. The patient underwent incision and drainage of spine and wrist, which resolved leukocytosis and encephalopathy. He was discharged in stable condition on IV antibiotics, pending outpatient evaluation for mitral valve replacement. Discussion: This case highlights a unique clinical presentation, caused by a rare infectious etiology of IE. Although the initial TTE was negative, the patient’s persistent leukocytosis and encephalopathy on appropriate antibiotic therapy prompted reevaluation, with recognition of subtle physical exam findings leading to more targeted imaging. This case emphasizes that a negative TTE does not exclude IE, having a sensitivity of 50-60%, thus warranting further exploration with TEE when clinically suspicious. Ultimately the fundamentals of medicine —ongoing reassessment and a thorough physical exam— were the key drivers in overcoming diagnostic uncertainty.

Alzheimer’s disease gradually erodes brain function, stringently disrupting memory andreasoning, expressly among older adults. Identifying the condition in its preliminary stages is decisive fortimely support and potentially more operative care.This study investigated the application of deep learningmodels for the automated detection of AD from MRI images. Three Convolutional Neural Network (CNN)architectures are utilized specifically—VGG16, Xception, and ResNet50. The models are evaluated in bothbinary classification and multi-class classification. Standard evaluation metrics are used to assess modelperformance. For binary classification, ResNet50 had the highest accuracy (97.96%), followed by VGG16 (97.10%) and Xception (95.93%). In multi-class classification, ResNet50 additionally led (95.39%), slightly ahead of VGG16 (94.92%) and Xception (94.93%).These results underscore the strong potential of ResNet50, in particular, for clinical application, demonstrating reliable generalization to previously unseen MRI images. The study highlights the potential of deep learning models to enhance early detection of Alzheimer’s disease by supporting clinical diagnosis, improving accuracy, and enabling timely interventions. Automated MRI analysis may also reduce costs and expand access to quality screening, especially in resource-limited settings reinforcing the growing case for integrating AI into medical imaging workflows.