Routine MRI Research Articles

Deep Resolve Boost applied to accelerated acquisition (DRB-ACC) offers the potential to reduce MRI acquisition time and improve image quality. However, studies on the potential impact on artifacts, anatomical delineation, and the depiction of imaging findings are scarce. This study aims to fill this gap and evaluates the clinical performance of DRB-ACC in 2D MRI sequences. In this prospective observational study, 256 paired 2D sequences (T2/TIRM, and FLAIR) were acquired from 200 patients undergoing routine neuroradiological MRI on a 3T scanner. For each examination, both standard and DRB-ACC (predominantly high strength) were acquired in the same session using identical scanner settings, except for acceleration factor; same surface coil, sequence orientation and slice thickness. Image quality, anatomical structure delineation, artifact presence, and lesion conspicuity were independently assessed by two readers using standardized scoring. DRB-ACC sequences demonstrated good or fair image quality in 94.5% of cases, with improved or unchanged quality compared to standard acquisition in 95.7% of sequences. However, anatomical delineation was inferior in key regions such as the hippocampus, brainstem, and cerebellum. Artifacts were more pronounced in 27.3% and newly introduced in 84.8% of the accelerated DRB sequences, commonly affecting the brainstem and deep gray matter. Lesion depiction was equivalent to standard images in 91.6% of cases, with limited instances of improved (7.2%) or degraded (1.2%) delineation. DRB-ACC enables acceleration of 2D MRI while maintaining image quality and lesion visibility. At the same time, the presence of artifacts and reduced delineation of certain anatomical structures underscores the need for caution in the interpretation of image findings and selective use in clinical routine, particularly in clinical scenarios requiring high anatomical detail such as for epilepsy screening or in cases with suspected brainstem pathology. ACC＝ accelerated acquisition; DRB＝ Deep Resolve Boost.

Introduction: Quantifying adiposity, a key biomarker of metabolic health, typically requires imaging that involves radiation, high costs, and manual effort. We developed an AI framework to segment external white adipose tissue (EWAT) from routine non-contrast MRI, offering a radiation-free, low-effort alternative. Hypothesis: We hypothesized that combining classical image processing with deep learning would enable accurate, robust EWAT segmentation from routine T1/T2-weighted MRI, without specialized sequences or manual labeling. Methods: In 105 Type 1 diabetes patients, T1/T2-weighted axial abdominal MRI scans at the aortic bifurcation were used to develop three segmentation approaches: Region Growing with automatic seed selection, iterative pixel aggregation and adaptive thresholds; UNet CNN trained on 52 masks from region-growing results, with Dice&Binary Cross-Entropy loss; and, Fine-Tuned UNet, optimized on 48 complex cases using extensive augmentations (flips, crops, brightness shifts, Gaussian noise) to enhance robustness and generalizability. Key challenges like artifacts, low fat volume, and anatomical overlap were addressed via local adjustments and hyperparameter tuning. Three independent clinicians scored segmentation quality (0–3) for anatomical alignment (Accuracy), circumference capture (Completeness), target area segmentation (Coverage), and boundary continuity (Smoothness). Results: Table 1 summarizes the mean clinical evaluation scores across all patients and metrics. The UNet and Fine-Tuned UNet consistently outperformed Region Growing in all four metrics, with mean accuracy of 2.81 and 2.80, respectively, versus 2.16 for Region Growing. Figure 1 shows example segmentations for each method. Figures 2 and 3 visualize model performance, highlighting mean scores in complex cases and the percentage of perfect (3/3) segmentations. The Fine-Tuned UNet had the highest mean accuracy (2.80) in challenging images, while the base UNet had the most perfect scores overall (73.3%). Conclusion: This unsupervised AI framework enables accurate, radiation-free EWAT segmentation from routine MRI. All methods, including deep learning, were trained without manual labeling, using region-growing outputs as pseudo ground truth. Clinical evaluations confirmed that the UNets achieved superior accuracy, completeness, coverage, and smoothness, particularly in complex cases. This scalable, cost-effective approach supports broader validation in cardiometabolic populations.

Background: Congenital scoliosis (CS) results from vertebral malformations and is often associated with intraspinal anomalies such as syringomyelia, diastematomyelia, and Chiari malformation. MRI is the gold standard for their detection, yet regional data from Pakistan are limited. Objective: To determine the spectrum and frequency of intraspinal anomalies detected by MRI in patients with congenital scoliosis managed at a tertiary care center over a 10-year period. Methodology: A retrospective observational study was conducted at the Department of Spine and Orthopedic Surgery, Hayatabad Medical Complex, Peshawar, from January 2013 to December 2023. A total of 122 patients with congenital scoliosis who underwent MRI were included. Data regarding demographics, scoliosis curve characteristics, Cobb’s angle, and MRI findings were retrieved from medical records. Intraspinal anomalies such as syringomyelia, diastematomyelia, tethered cord, and Arnold Chiari malformation were identified. Statistical analyses included descriptive measures and multivariate logistic regression to explore associations between scoliosis severity and anomaly type, with significance set at p<0.05. Results: The mean age of patients was 14.66 ± 7.44 years (range 4–54), with a slight female predominance (54.9%). Right-sided curves were more common (87.7%). Cobb’s angle ranged from 30° to 110°, with a mean of 65.43° ± 14.78. MRI detected intraspinal anomalies in 20.5% of patients: syrinx (14.8%), diastematomyelia (4.1%), and Arnold Chiari malformation (0.8%). Severity of scoliosis (higher Cobb’s angle) was significantly associated with syrinx and diastematomyelia. Surgical interventions included Growing Rod Fixation (32.8%), Posterior Vertebral Column Resection (23.8%), and Posterior Instrumented Spine Fusion (19.7%), with favorable postoperative neurological outcomes in the majority of patients with anomalies. Conclusion: One-fifth of CS patients had intraspinal anomalies, most commonly syrinx. Routine MRI is crucial for early detection and safe surgical planning. Keywords: Congenital scoliosis, Intraspinal anomalies, MRI, spinal dysraphism, Syringomyelia.

Accuracy of detecting critical findings using abbreviated brain MRI scan protocols as a prerequisite for AI-driven on-the-fly scan protocol adaptation.

Cranial vault reconstruction (CVR) is a frequent treatment in patients with craniosynostosis. Complications are infrequent, including dural tears, CSF leaks, venous sinus injuries, and hemorrhages. Prior research demonstrated the limited utility of routine post-operative CT scans in uncomplicated craniosynostosis surgery, showing only one of 469 postop CTs after CVR resulted in a change in management. In this study, we investigate the utility of postop MRI after CVR. One hundred eleven consecutive routine MRIs for 107 patients on postoperative day 1 after open CVR spanning September 2018 to February 2023 were reviewed. Average age at surgery was 2 years 5 months. 25.2% of patients (n=28) had a prior craniosynostosis procedure. 71.1% of patients (n=79) had multi-suture synostosis, and 29.7% of patients (n=33) had syndromic synostosis, with Apert (12%, n=14) and Crouzon (10%, n=11) most prevalent. 71.1% of patients (n=79) underwent fronto-orbital advancement (FOA) and 23.4% of patients underwent posterior CVR (n=26). Only 2 postoperative MRIs changed management. One found an incidental cervical neoplastic lesion prompting follow-up with oncology. Another found asymptomatic and scattered foci of T2 prolongation, which resolved on a subsequent scan at 2 months postop and never produced clinical changes. Eighty-two percent of radiology impressions included post-operative extra-axial fluid collections, whereas 41% contained pneumocephalus, neither of which required intervention. Overall, 111 consecutive routine post-op day one MRIs yielded no changes in inpatient management related to post-surgical imaging findings. There are benefits of postop imaging (family and care team reassurance, education, new baseline), but routine postop MRI has limited utility for acute management.

This study aimed to develop and validate a radiomics fusion model based on CT and MRI for distinguishing between spinal osteosarcoma and chondrosarcoma, and to compare the performance of models derived from different imaging modalities. A retrospective analysis was conducted on 63 patients with histologically confirmed spinal osteosarcoma (n=20) and chondrosarcoma (n=43). Radiomics features were extracted from CT and MRI (T1-weighted, T2-weighted, and T2-weighted fat-suppressed) sequences, followed by feature selection using univariate logistic regression and LASSO. Eight machine learning models were utilized to construct radiomics models, based on CT, MR, both CT and MR, and clinical information combined with CT and MR. Models were evaluated via five-fold cross-validation and compared against radiologists' interpretations using the area under the receiver operating characteristic curve (AUC), accuracy, sensitivity, specificity, F1 score, and Matthews correlation coefficient. The MRI-based radiomics model using linear discriminant analysis achieved the highest diagnostic performance (AUC=0.963, sensitivity=95.3%, specificity=80.0%), significantly outperforming both CT-based models (AUC=0.700) and radiologists' diagnosis (p<0.001). The CTMR and clinico-CTMR models did not show significant improvement over the MR model. The MR model demonstrated excellent calibration and clinical utility, with substantial net benefit across threshold probabilities. The superior performance of the MRI-based model highlighted the value of MRI radiomics in tumor differentiation. This clinically practical tool may support preoperative diagnosis using routine MRI, potentially facilitating more timely treatment decisions. In conclusion, the MRI-based radiomics model enabled accurate preoperative discrimination between spinal osteosarcoma and chondrosarcoma.

Developing a deep-learning model for automated multi-tissue, multi-condition knee MRI analysis and assessing its clinical potential. This retrospective dual-center study included 3121 MRI studies from 3018 adults, who underwent routine knee MRI examinations at a radiologic practice (2012-2019). Twenty-three conditions across cartilage, menisci, bone marrow, ligaments, and other soft tissues were manually labeled. A 3D slice transformer network was trained for binary classification and evaluated in terms of the area under the receiver operating characteristic curve (AUC), sensitivity, and specificity using a five-fold cross-validation and an external test set of 448 MRI studies (429 adults) from a university hospital (2022-2023). To assess differences in diagnostic performance, two inexperienced and two experienced radiology residents read 50 external test studies with and without model assistance. Paired t-tests were used for statistical analysis. Averaged over cross-validation tests, the model's AUC was at least 0.85 for 8 conditions and at least 0.75 for 18 conditions. Generalization on the external test set was robust, with a mean absolute AUC difference of 0.05 ± 0.03 per condition. Model assistance improved accuracy and sensitivity for inexperienced residents, increased inter-reader agreement for both groups, and increased sensitivity and shortened reading times by 10% (p = 0.045) for experienced residents. Specificity decreased slightly when conditions with low model performance (AUC < 0.75) were included. Our deep-learning model performed well across diverse knee conditions and effectively assisted radiology residents. Future work should focus on more fine-grained predictions for subtle or rare conditions to enable comprehensive joint assessment in clinical practice. Question Increasing MRI utilization adds pressure on radiologists, necessitating comprehensive AI models for image analysis to manage this growing demand efficiently. Findings Our AI model enhanced diagnostic performance and efficiency of resident radiologists when reading knee MRI studies, demonstrating robust results across diverse conditions and two datasets. Clinical relevance Model assistance increases the sensitivity of radiologists, helping to identify pathologies that were overlooked without AI assistance. Reduced reading times suggest potential alleviation of radiologists' workload.

BackgroundHER2-positive breast cancer is leading to aggressive tumor growth and a higher risk of metastasis, particularly to the central nervous system (CNS). Routine brain imaging for asymptomatic HER2-positive patients is debated, with no current consensus; Given the severe clinical implications of brain metastases, further research is needed to determine the cost-effectiveness and clinical utility of routine imaging for high-risk patients to improve outcomes and inform targeted screening protocols.MethodsThis retrospective, monocentric study at the General Hospital of Vienna (AKH Wien) examined female HER2-positive breast cancer patients at first diagnosis to assess brain metastasis from January 2019 to February 2024. The study included patients with asymptomatic confirmed HER2 positive breast cancer. Data were collected through comprehensive medical records and brain imaging with MRI.ResultsAmong 110 female patients meeting the inclusion criteria, 4 (3.6%) were diagnosed with brain metastases. Ki67 showed a marginal association with brain metastasis (p = 0.054), and tumor grade was a significant predictor, with intermediate differentiated tumors (G2 vs. G3) associated with a higher risk of brain metastases (p = 0.041) and brain metases are correlating with the axillary lymphnode status and the tumor sizeAlso, the absence of positive Östrogen and Progesteron receptors is a predictor in upcoming brain metastases (p < 0.001). Other factors like age were not significantly associated.ConclusionThis study found limited benefit in routine MRI for detecting asymptomatic brain metastases in HER2-positive breast cancer, given the low prevalence (3.6%). A targeted imaging approach for high-risk patients, like those with the absence of Hormon receptors and higher stage tumors, may be effective.

BackgroundPatients with type 2 diabetes mellitus (T2DM) exhibit a heightened susceptibility to developing dementia, especially those who already present with mild cognitive impairment (MCI). Nevertheless, the fundamental etiology remains elusive, underscoring the pressing need for an objective and precise diagnostic approach in clinical settings. This study investigates the utilization of machine learning algorithms in conjunction with high-resolution sagittal T1-weighted structural imaging to facilitate automated diagnosis of T2DM patients with MCI, differentiating them from both T2DM patients without MCI and healthy controls (HCs).MethodsThirty patients with T2DM and MCI, thirty T2DM patients without MCI, and thirty matched healthy controls (HCs) were enrolled to identify independent biomarkers and develop a diagnostic model for early cognitive impairment in T2DM. Whole-brain structural features-including cortical surface area, volume, thickness, curvature index, folding index, Gaussian curvature, mean curvature, thickness standard deviation, nuclear volume, hippocampal volume, and white matter volume-were extracted from the images of brains using automated segmentation methods. The minimum redundancy maximum relevance (MRMR) method was employed to filter out irrelevant and redundant features, reducing the dimensionality of the dataset. Subsequently, the least absolute shrinkage and selection operator (LASSO) algorithm was applied for further feature selection, ensuring the retention of only the most diagnostic features. The Random Forest (RF) classifier, a powerful machine learning model within the realm of artificial intelligence, was meticulously trained utilizing a curated feature set that had undergone rigorous refinement. To ensure the robust diagnostic performance and generalizability of the established random forest model, a 5-fold cross-validation was employed, providing a dependable estimation of the model’s effectiveness.ResultsThe FreeSurfer software automatically segmented the cerebral imaging data into up to 70 regions. For model establishment, a comprehensive set of 705 structural features, a series of neuropsychological tests, and standard laboratory tests were utilized. Ultimately, 8 features were selected through two feature selection strategies aimed at refining the optimal features. These included bilateral brainstem volume, left hippocampus volume, left transverse temporal gyrus volume, bilateral posterior corpus callosum volume, left medial orbitofrontal cortex Gaussian curvature, glycosylated hemoglobin, blood sugar levels, and the Digit Span Test (DST) backward score. The Random Forest (RF) model, based on the combined features, exhibited the excellent performance, with mean AUCs of 0.959 (95% CI, 0.940–0.997, mean specificity = 94.2%, mean sensitivity = 88.3%, mean accuracy = 88.3% and mean precision = 88.3%) for the training dataset and mean AUCs of 0.887 (95% CI, 0.746–0.992, mean specificity = 85.0%, mean sensitivity = 70.0%, mean accuracy = 70.0% and mean precision = 69.6%) for the testing dataset, based on 5-fold cross-validation.ConclusionThe RF model, leveraging a combination of features, demonstrates high accuracy in diagnosing T2DM with MCI. The exclusion of T2DM patients with complications may limit generalizability to the broader T2DM population, potentially inflating performance estimates. Among these features, 8 optimal indicators comprising 5 structural features, 1 neuropsychological test feature, and 2 standard laboratory test features emerge as the potential independent biomarkers for detecting early-stage cognitive impairment in T2DM patients. These features hold significant importance in understanding the pathophysiological mechanisms of T2DM-related cognitive impairment. Our fully automated model is capable of swiftly processing MRI data, enabling precise and objective differentiation of T2DM with MCI. This model significantly enhances diagnostic efficiency and holds considerable value in clinical practice, facilitating early diagnosis of T2DM with MCI.

Retrospective study. To investigate the correlation between sagittal cervical canal diameter on lateral whole spine radiographs and the incidence of syringomyelia in scoliosis patients. Neurosurgical evaluation and management are imperative in scoliosis patients with syringomyelia, as there is an increased risk of neurological deterioration during deformity correction surgery. Nevertheless, routine MRI screening is not indicated in scoliosis patients with a typical curve and normal neurology. Twenty-nine scoliosis patients with syringomyelia were grouped in Gp1, and 1029 adolescent idiopathic scoliosis (AIS) patients were grouped in Gp2. Propensity score matching (PSM) with one-to-one nearest neighbour matching (match tolerance 0.05) was performed. Twenty-eight pairs of patients were matched. Sagittal cervical canal diameter and the ratio between cervical canal/vertebral body were measured. Receiver-operating-characteristics (ROC) curves for both parameters were drawn to predict the incidence of syringomyelia. Sagittal canal diameter of C5 to C7 in Gp1 was significantly increased compared with Gp2 (C5: 17.78±3.61 vs. 16.32±1.18; P=0.047), (C6: 18.67±3.30 vs. 16.68±1.25; P=0.004), (C7: 18.99±3.24 vs. 17.06±1.17; P=0.004). The ratio between canal diameter/vertebral body in Gp1 was significantly increased compared with Gp2. (C5: 1.42±0.38 vs. 1.23±0.15; P=0.015), (C6: 1.47±0.35 vs. 1.22±0.14; P=0.001), (C7: 1.46±0.29 vs. 1.20±0.13; P<0.001). ROC analysis of C7 canal diameter showed acceptable discrimination of AUC 0.663 in predicting syringomyelia (P=0.036). The cutoff value of 17.65mm had a sensitivity of 60.7% and a specificity of 71.4%. C7 canal ratio showed excellent discrimination of AUC 0.794 in predicting syringomyelia (P<0.001). The cutoff value of 1.22 had a sensitivity of 78.6% and a specificity of 71.4%. Sagittal cervical canal diameter measured in radiograph can be utilized as an effective screening tool in the diagnostic evaluation of syringomyelia. We propose C7 canal diameter of 17.65mm or C7 ratio of 1.22 as an indication for MRI scan to rule out syringomyelia in scoliosis patients.

Abstract BACKGROUND PET using the radiolabeled amino acid O-(2-[18F]-fluoroethyl)-L-tyrosine ([18F]-FET) has considerable clinical value for follow-up evaluation of central nervous system tumors in children and adolescents. As medical procedures must be justified socio-economically, we determined cost-effectiveness of [18F]-FET PET for identification of treatment-related changes. PATIENTS AND METHODS We analyzed clinical data from two different studies that assessed the value of [18F]-FET PET to differentiate between brain and spinal tumor relapse and treatment-related changes in children and adolescents. Cost calculation was based on the German statutory health insurance system perspective. Due to subtle differences in the diagnostic approach of the studies, two separate clinical scenarios including 80 patients with 105 lesions were considered: Decision tree model 1 determined cost-effectiveness of simultaneous [18F]-FET PET and MRI in comparison to MRI alone to identify treatment-related changes. Decision tree model 2 determined cost-effectiveness of [18F]-FET PET alone to identify treatment-related changes when prior routine follow-up MRI findings were suspicious for tumor relapse. Deterministic and probabilistic sensitivity analyses tested the robustness of the results. RESULTS Model 1 revealed that the rate of identified treatment-related changes increased by 52% when adding [18F]-FET PET to MRI, resulting in costs of €3,314.51 for each additional correctly identified lesion with treatment-related changes by [18F]-FET PET that MRI would have misclassified. Model 2 revealed that [18F]-FET PET correctly identified treatment-related changes in 90% of lesions when prior MRI findings were suspicious for tumor relapse, resulting in costs of €1,740.37 for each lesion. CONCLUSION Integrating [18F]-FET PET in the follow-up of in children and adolescents with brain and spinal tumor may help improving patient care at acceptable costs.

MRI screening is increasingly used in elite cricket to support the early detection of lumbar bone stress injuries (LBSIs). However, its impact on injury outcomes and career trajectories remains unclear. This study evaluated short-term and long-term outcomes from LBSI in young male fast bowlers, comparing injuries detected through MRI screening with those detected clinically. This retrospective cohort study analysed 15 years (2009-2024) of injury surveillance data from the England and Wales Cricket Board. LBSIs in male fast bowlers (aged 16-24) were classified as screening-detected (via routine MRI) or clinically-detected (imaging-confirmed following symptom onset). Outcomes were evaluated across short-term (diagnosis to return-to-play (RTP)) and long-term (5-year post-RTP) timeframes. 197 LBSIs (44 screening-detected, 153 clinically-detected) were detected in 142 bowlers (mean age at injury: 19.4±2.1 years). Screening-detected injuries were more likely to be stress reactions (p=0.002), while complete fractures were exclusively clinically-detected. Screening-detected injuries had shorter RTP times (p<0.001) and were managed conservatively with 100% RTP success. Bowlers with screening-detected injuries had greater in-season availability (p<0.001) and returned to preinjury playing standards post-RTP. Over 5 years, 51%-67% of bowlers remained reinjury-free, with comparable probabilities between detection methods (p≥0.536); 5-year playing standard trajectories showed improvement in both groups. MRI screening may facilitate earlier detection of LBSIs compared with symptom-based pathways, enabling conservative management, shorter RTP times and greater availability, while allowing for continued career progression. However, the observed risk of reinjury highlights the need for improved preventive strategies.

Preliminary study of multiple diffusion MRI in defining brain microstructural changes in hypertensive individuals.

Cognitive decline in older adults, particularly during the preclinical stages of Alzheimer's disease (AD), presents a critical opportunity for early detection and intervention. While T1-weighted MRI is widely used in AD research, its capacity to identify early vulnerability and monitor longitudinal progression remains incompletely characterized. We analyzed longitudinal T1-weighted MRI data from 224 cognitively unimpaired older adults followed for up to 12years. Participants were stratified by clinical outcome into converters to mild cognitive impairment (HC-converters, n = 112) and stable controls (HC-stable, n = 112). Groups were matched at baseline for age (mean ~ 74-75years), education (~ 16.4years), and cognitive scores (MMSE ≈ 29; CDR-SB ≈ 0.04). Four MRI-derived biomarkers were examined: brain-predicted age difference (brain-PAD), mean cortical thickness, AD-cortical signature, and hippocampal volume. Brain-PAD showed the strongest baseline association with future conversion (β = 1.25, t = 3.52, p = 0.0009) and highest classification accuracy (AUC = 0.66; sensitivity = 62%, and specificity = 67%). Longitudinal mixed-effects models focusing on the group × time interaction revealed a significant positive slope in brain-PAD for converters (β = 0.0079, p = 0.003) and a non-significant trend in stable controls (β = 0.0047, p = 0.075), indicating incipient divergence in brain aging trajectories during the preclinical window. Hippocampal volume and AD-cortical signature declined similarly in both groups. The mean cortical thickness had limited discriminative or dynamic utility. These findings support brain-PAD, derived from routine T1-weighted MRI using machine learning, as a sensitive, performance-independent biomarker for early risk stratification and monitoring of cognitive aging trajectories.

Fully automated measurement of aortic pulse wave velocity from routine cardiac MRI studies.

Automated vertebral bone quality score measurement on lumbar MRI using deep learning: Development and validation of an AI algorithm.

Temporomandibular disorders (TMDs) involve functional and structural disturbances of the temporomandibular joint (TMJ), often requiring early imaging for effective management. This study compared the diagnostic performance of three MRI sequences-PDFSE, MEDIC, and 3D DESS-in evaluating early intra-articular changes in 108 TMJs of symptomatic patients. Key parameters assessed included disc morphology, joint effusion, bone marrow status, and image quality. PD-FSE showed the highest signal intensity ratios for the mandibular condyle, disc, and lateral pterygoid muscle, indicating superior contrast. MEDIC provided the best image quality and demonstrated the highest sensitivity for detecting bone marrow edema. Although PD-FSE detected moderate joint effusions most frequently, the difference was not statistically significant. Inter-observer agreement across all sequences was excellent. Overall, PD-FSE offered superior contrast, while MEDIC enhanced detection of subtle bone and fluid changes. A combined approach using both sequences may improve early diagnosis and clinical decision-making in TMD management.

Background and ObjectivesPreterm birth is associated with altered white matter development and long-term neurodevelopmental impairments. Skin-to-skin care has known benefits for physiologic regulation and bonding in preterm infants, but impacts on early brain structure remain unclear. The aim of this study was to describe the association between in-hospital skin-to-skin care and white matter microstructure in very preterm infants, focusing on frontolimbic tracts involved in stress regulation and socioemotional development.MethodsThe design was a single-center retrospective observational analysis of clinical data from the electronic medical records and diffusion MRI scans. Participants were infants born at <32 weeks gestational age (GA) who received a routine predischarge MRI. Skin-to-skin care was quantified as duration per instance and daily exposure rate (in minutes) before the MRI was obtained. Diffusion MRI assessed mean diffusivity (MD) and fractional anisotropy (FA) in the cingulum, anterior thalamic radiations (ATRs), and uncinate fasciculus. Hierarchical regression models evaluated associations between skin-to-skin care and white matter metrics, adjusting for GA, medical acuity, postmenstrual age at scan, and MRI coil type.ResultsA total of 88 preterm infants (mean GA 29 weeks; 49% female) were included. Skin-to-skin care duration per instance was positively associated with MD in the cingulum (B = 0.002, 95% CI 0.0004–0.003, ΔR2 = 0.080) and ATRs (B = 0.002, 95% CI 0.0003–0.003, ΔR2 = 0.057). Skin-to-skin care daily exposure rate was also positively associated with ATR MD (B = 0.038, 95% CI 0.001–0.076, ΔR2 = 0.046). Both skin-to-skin metrics were negatively associated with ATR FA (duration: B = −0.0005, 95% CI −0.001 to −0.0001, ΔR2 = 0.046; rate: B = −0.016, 95% CI −0.028 to −0.004, ΔR2 = 0.075). No significant associations were found for the uncinate fasciculus. Findings remained significant after adjusting for socioeconomic status and visitation frequency and after excluding infants with white matter injury.DiscussionSkin-to-skin care was associated with neonatal white matter microstructure in specific frontolimbic tracts. Limitations include the retrospective design and single-center setting. Future studies should consider how early caregiving experiences, such as skin-to-skin care, may influence brain development in preterm infants.

We present a deep learning-based approach for accurate bone segmentation directly from routine T1-weighted MRI scans, with the goal of enabling MRI-only virtual surgical planning in head and neck oncology. Current workflows rely on CT for bone modeling and MRI for tumor delineation, introducing challenges related to image registration, radiation exposure, and resource use. To address this, we trained a deep neural network using CT-based segmentations of the mandible, cranium, and inferior alveolar nerve as ground truth. A dataset of 100 patients with paired CT and MRI scans was collected. MRI scans were resampled to the voxel size of CT, and corresponding CT segmentations were rigidly aligned to MRI. The model was trained on 80 cases and evaluated on 20 cases using Dice similarity coefficient, Intersection over Union (IoU), precision, and recall. The network achieved a mean Dice of 0.86 (SD ± 0.03), IoU of 0.76 (SD ± 0.05), and both precision and recall of 0.86 (SD ± 0.05). Surface deviation analysis between CT- and MRI-derived bone models showed a median deviation of 0.21 mm (IQR 0.05) for the mandible and 0.30 mm (IQR 0.05) for the cranium. These results demonstrate that accurate CT-like bone models can be derived from standard MRI, supporting the feasibility of MRI-only surgical planning.

Prior studies have reported that imaging evaluation of osseous morphology in femoroacetabular impingement (FAI) is best performed with CT, which exposes patients to ionizing radiation. In recent years, a number of studies have evaluated whether various novel MRI protocols, which do not expose patients to ionizing radiation, can effectively assess osseous morphology in patients with FAI. Our institution incorporated in- and out-of-phase sequences into a routine MRI protocol to better assess acetabular version; however, it is unknown how in- and out-of-phase MRI compares with CT imaging in FAI evaluation. (1) How reliably do acetabular version measurements taken from in- and out-of-phase MRI agree with acetabular version measurements taken from CT imaging? (2) How similar are hip morphometric measurements taken from routine MRI sequences as compared with hip morphometric measurements taken from hip-specific CT? We conducted a retrospective electronic medical record review of the patients of two attending sports medicine orthopaedic surgeons from May 2014 to May 2024 who were evaluated for symptomatic hips. It is the general practice of these surgeons to obtain both hip-specific CT scans and in- and out-of-phase MRI for patients with suspected FAI. Patients were included if they had a diagnosis of FAI, were older than 12 years of age, underwent hip-specific morphometric CT scanning and in- and out-of-phase MRI of the affected side, and had imaging interpretation performed by fellowship-trained musculoskeletal radiologists at our institution. Hip morphometric measurements were retrospectively recorded from prospectively interpreted radiology reports. Our initial chart review yielded 178 patients (188 hips) with a diagnosis of FAI who underwent both CT and MRI imaging. After the application of inclusion and exclusion criteria, 30 patients (33 hips) lacked an in- and out-of-phase MRI, 11 patients (11 hips) had the imaging performed on contralateral hips, and 42 patients (44 hips) lacked complete morphometric measurements, yielding 95 patients (100 hips) who were included in our study. Our study population comprised 72% (68 of 95) females with a mean ± SD age of 29 ± 9 years and BMI of 25.3 ± 4.7 kg/m2. Of those included, 56 patients had their measurements confirmed by our institution's fellowship-trained musculoskeletal radiologists to assess for intrarater and interrater reliability. The assessed morphometric measurements included: midcoronal angle, midsagittal angle, alpha angle, femoral neck angle, and femoral neck version at the 1, 2, and 3 o'clock positions. These measurements were statistically compared with intraclass correlation coefficients (ICCs) to assess intermodality measurement agreement and thus determine the reliability between in- and out-of-phase MRI and CT. Each morphometric measurement also underwent t-tests to determine the similarity in measurements between in- and out-of-phase MRI and hip-specific CT sequences. Acetabular version measurements taken from the unique sequences of in- and out-of-phase MRI demonstrated ICCs of 0.62, 0.67, and 0.80 at 1, 2, and 3 o'clock, respectively. Other morphometric measurements with standard MRI sequencing demonstrated ICCs ranging from a low of 0.41 (poor) for femoral neck angle to a high of 0.73 (moderate) for femoral version. Higher ICCs demonstrate moderate to good agreement between imaging modalities for acetabular version measurements with unique axial sequences of in- and out-of-phase MRI and CT scans. ICC values comparing measurements from routine MRI protocol and CT scans demonstrate poor to moderate agreement in morphometric measurements between imaging modalities. This indicates a reliable agreement in morphometric measurements between in- and out-of-phase MRI and CT and less reliability in agreement for measurements made with routine MRI. The in- and out-of-phase MRI protocol had moderate to good reliability in correlation to CT for evaluating acetabular version in patients with FAI. Hip osseous and soft tissue evaluation may be effectively achieved using in- and out-of-phase MRI axial sequences in addition to routine hip MRI sequences. Surgeons who perform hip arthroscopy may consider being more selective in using CT for evaluating FAI. Future research may incorporate additional MRI sequences to better evaluate FAI hip morphometric measurements. Level IV, diagnostic study.