Tomography Magnetic Resonance Imaging Research Articles

Scan/rescan reliability of magnetic resonance imaging (MRI)

Abstract Background Magnetic resonance imaging (MRI) is increasingly used to estimate the geometric dimensions of lower lumbar vertebrae. While MRI-based measurements have demonstrated good reliability with interclass correlation coefficients (ICCs) of 0.80 or higher, many evaluations focus solely on the comparison of identical MRI images. This approach primarily reflects analyst dexterity and does not assess the reliability of the entire process, including imaging and image selection. Objective To evaluate the inter- and intra-rater reliability of the entire process of using MRI to measure biomechanically relevant lumbar spinal characteristics, incorporating imaging, image selection, and analysis. Methods: A dataset of 144 low-back MRI scans was analyzed. Reliability assessments were performed under different conditions: (1) identical scans rated by the same analyst at different times (intra-rater reliability) and (2) distinct scans of the same subject obtained by different MRI operators and analyzed by different analysts (inter-rater reliability). Mean absolute differences in measurements were calculated, and sources of variability, such as breathing artifacts, were noted. Results Larger discrepancies were observed when comparing distinct scans analyzed by different MRI operators and analysts. In the “worst-case” scenario, where both the MRI operator and analyst differed, a 4.05% mean absolute difference was noted for anterior endplate measurements. This was higher than the 2.76% difference observed when analysts re-rated their own scans after one month. Despite these discrepancies, the variability in measurements was relatively low and primarily attributed to factors like breathing artifacts. Conclusion The process of using MRI to derive biomechanical measures, particularly for bony structures, demonstrates robust reliability. Variability in measurements is minimal even under challenging conditions, supporting the use of MRI for biomechanical assessments.

Early Detection of Alzheimer's Disease from MR Images Using Fine-Tuning Neighborhood Component Analysis and Convolutional Neural Networks

Abstract This study develops an automatic algorithm for detecting Alzheimer's disease (AD) using magnetic resonance imaging (MRI) through deep learning and feature selection techniques. It utilizes a dataset of 6400 MRI images from Kaggle, categorized into four classes. Initially, the study employs pretrained CNN architectures—DenseNet-201, MobileNet-v2, ResNet-18, ResNet-50, ResNet-101, and ShuffleNet—for classification using five fold cross-validation, with DenseNet-201 achieving the highest accuracy of 82.11%. Due to the dataset's size and imbalance, as well as the long training times, the study aims to create a more efficient algorithm. The CNNs are used as deep feature extractors from AD images, and the extracted features are reduced using a new fine-tuning neighborhood component analysis (FTNCA) algorithm, which minimizes loss and determines the optimal tolerance value. The essential features are then classified using various machine learning algorithms, including artificial neural network (ANN), K-nearest neighbor (KNN), Naïve Bayes, and support vector machine (SVM). Experimental results reveal that reducing the feature set from 2048 to 344 allows the ResNet-50-FTNCA-KNN model to achieve 100% accuracy, significantly enhancing AD detection. This approach will aid in the early diagnosis and treatment of AD patients.

Prenatal Diagnosis of Caroli's Disease by Ultrasound and MRI Imaging.

To present the imaging features of Caroli's disease (CD) on prenatal ultrasound and magnetic resonance imaging (MRI). This was a retrospective case series of prenatally diagnosed CD between 2017 and 2024. Clinical data from these cases were collected and reviewed. Five fetuses with CD were included, three of which had a definite combination of ARPKD and suspected in the other 2. Prenatal ultrasonography revealed multiple intrahepatic bile duct dilatations in four fetuses, each of which displayed the "horn comb" sign in a cross-section of the liver. All five fetuses had abnormal kidney ultrasounds: three showed enlarged and hyperechogenic kidneys and two showed hyperechogenic kidneys. The MRI scans of all fetuses showed a "central dot" (C-DOT) sign in the liver. By MRI, three fetuses had enlarged kidneys, one slightly had hyperintensity kidneys, and one had no significant kidney abnormalities. Pregnancy termination was chosen in all cases. CD may be identified by fetal ultrasound through the characteristic arrangement of intrahepatic dilated bile ducts ("horn comb" sign). Fetal MRI is advantageous for detecting the C-DOT sign, which confirms the diagnosis of CD. In our experience, these findings tend to become apparent in the late second to early third trimester of pregnancy.

Evaluation of sectional anatomic, micro-computed tomographic, and magnetic resonance imaging features of the thorax in Syrian hamsters (Mesocricetus auratus).

Clinically, the rodent thorax is important because of the variety of problems that may affect the heart, lungs, and other thoracic structures. Syrian hamsters are the most common pet and experimental hamster species. Sectional imaging of small mammals is widely increasing in use for clinical and research settings; however, no reports on the thoracic sectional imaging anatomy in this species have been made. Therefore, this study aimed at evaluating micro-computed tomography (MCT), magnetic resonance image (MRI), and gross sections of the Syrian hamster thorax. Eight healthy adult male and female Syrian hamsters were examined. Anesthetized hamsters were evaluated by MCT and 3 Tesla MRI. After imaging, the frozen slab sections were photographed, and identified anatomic structures were matched with structures on corresponding MCT and/or MRI images. Clinically relevant anatomic features of the Syrian hamster thorax that were identified on transverse and sagittal anatomic sections could be discerned on the corresponding MCT and MRI scans. The three matched images were exhibited the following features: (1) the narrow thoracic cavity and small lung space, (2) the heart positioning ventral and medial to lung, (3) heart ventricles locating towards the left side, and (4) parallel cardiac position relative to the sternum. The obtained results were similar to those have only been reported in rabbits, regardless of heart orientation and lung lobe covering the heart. The results of this study support the use of MCT and MRI as diagnostic tools in Syrian hamsters and provide reference values for the clinically relevant anatomic structures of thorax.

Segmenting Small Blood Vessels in MRA/MRI: A Machine Learning Approach with Bilateral Filtering Integration

The adoption of artificial intelligence (AI) in various sectors, including healthcare, has gained significant popularity due to its potential to improve services. In the medical field, misdiagnosis has been a major problem, leading to increased mortality rates. Accurate diagnosis is crucial for effective treatment and management of diseases. This research aims to develop a machine-learning model for segmenting small blood vessels in magnetic resonance angiography (MRA) and magnetic resonance imaging (MRI) datasets using bilateral filtering. The research identifies the limitations of existing machine learning models in blood vessel segmentation, particularly the loss of important edge information due to convolutions that blur images. To address this issue, a non-linear bilateral filter is introduced to enhance the segmentation of blood vessels in MRI images. The proposed framework aims to improve the accuracy of the segmentation algorithm by reducing image blurring and noise through bilateral filtering. The objectives of this research include training and testing a machine-learning prototype using bilateral filtering, exploring the weaknesses of existing models in blood vessel segmentation, and developing a machine-learning model specifically designed for segmenting small blood vessels using bilateral filtering. Various studies have proposed machine learning algorithms, such as convolutional neural networks, for blood vessel segmentation. The review emphasizes the importance of bilateral filtering in improving classification accuracy by reducing image blurring. In conclusion, this research aims to contribute to the field of medical image analysis by developing a framework that utilizes bilateral filtering to enhance the segmentation of small blood vessels in MRA and MRI datasets. The proposed machine learning model has the potential to improve the accuracy of blood vessel segmentation, enabling more accurate diagnoses and reducing misdiagnosis-related mortality rates.

The correlation histopathological and conventional/advanced MRI techniques in glial tumors.

We aimed to elucidate the histopathological pre-diagnosis of cranial gliomas with magnetic resonance imaging (MRI) techniques in gliomas. A total of 82 glioma patients were enrolled to our study. Pre-operative conventional MRI images (non-contrast T1/T2/flair/contrast-enhanced T1) and advanced MRI images (DAG and ADC mapping, MRI spectroscopy and perfusion MRI [PMRI]) were analyzed. Conventional MRI alone is useful in radiological pre-evaluation in low-grade glioma in 54.8% and 86.3% in high-grade glioma. Additional advanced MRI techniques were beneficial in comparing low-grade gliomas in 98% and 83.9% in high-grade glioma. On ROC analysis, ADC cutoff value 0.905 mm2/s (p = 0.001), rCBV cutoff value 1.77 (p = 0.001), Cho/NAA cut-off value 2.20 (p = 0.001), and Cho/Cr cutoff value 2.01 (p = 0.001) were achieved. Significant results were obtained when ADC, Cho/NAA, and Cho/Cr were analyzed into four histopathologically grade groups besides (p = 0.001). NAA/Cr values were not significant in pathological grading. rCBV measurements were statistically significant between Grades I and IV and between II and IV. Using additional advanced MRI techniques such as PMRI, magnetic resonance spectroscopy, and DWI with conventional MRI could enhance the accuracy of histopathological grading in cranial glioma.

Medical Image Enhancement using Histogram Processing and Feature Extraction of Cancer Classification

Magnetic Resonance Imaging (MRI) is a critical tool in medical diagnostics, particularly for cancer detection and classification. However, the quality of MRI images often suffers from noise, low contrast, and intensity inhomogeneity, which can hinder accurate diagnosis. Image enhancement techniques, such as histogram processing, play a crucial role in improving image quality by enhancing contrast and emphasizing important fea- tures. This paper explores the application of histogram processing for MRI image enhancement, focusing on contrast enhancement and noise reduction techniques. The enhanced images are subsequently processed for feature extraction, which is crucial for the accurate classification of cancerous tissues. Key features such as texture, shape, and intensity are extracted to distinguish between malignant and be- nign tumors. The classification process utilizes machine learning algorithms, which are trained on the extracted features to achieve high accuracy in cancer classification. The proposed method demonstrates significant improvement in image quality, leading to better feature extraction and more accurate cancer diagnosis. This approach ultimately aids in early detection and improved treatment planning for patients. Index Terms—Image enhancement, Histogram processing, Seg- mentation, Feature extraction, SVM classifier.

MRI imaging characteristics of brainstem and midbrain aqueduct in patients with iNPH

Idiopathic normal pressure hydrocephalus (iNPH) is frequently difficult to diagnose due to the absence of specific symptoms, yet early detection and surgical intervention are essential for preventing sequela such as irreversible dementia. This study explores the specific magnetic resonance imaging (MRI) features of the brainstem and mesencephalic aqueduct in patients with iNPH. Head MRI data of 50 iNPH patients and 30 healthy matched controls were compared for mesencephalic aqueduct length, diameter, and angle, structural features of the brainstem at the sagittal plane, brainstem component volume ratios, angle between the brainstem and spinal cord, and the area and morphology of the pontine cisterns. Compared to healthy individuals, iNPH patients exhibited significant dilation of the mesencephalic aqueduct diameter, a reduced aqueduct angle, and a decrease in the sagittal plane area of the brainstem, with the most pronounced reduction in the midbrain region. Notably, the CSF spaces surrounding the brainstem were dilated, resulting in the prepontine cistern presenting a “hammer-like” shape on the sagittal plane. The prevalence of this hammer shape was positively correlated with prepontine cistern area in patients with iNPH. These unique imaging characteristics may facilitate the clinical recognition of iNPH for early diagnosis and treatment.

Influence of gender and bruxism on the masseter muscle: A population-based magnetic resonance imaging study

This study aimed to (1) examine the cross-sectional area (CSA) of the masseter muscle in relation to bruxism and (2) investigate if there are gender-related differences in the masseter muscle in a population-based magnetic resonance imaging (MRI) study. The study included 720 subjects aged 30-89 years (391 women and 329 men) from the Study of Health in Pomerania, a cross-sectional population-based study assessing the prevalence and incidence of common population-relevant diseases and their risk factors in Northeastern Germany. The participants underwent both a whole-body MRI and a full oral examination. The CSAs of the masseter muscles on both sides were measured from MRI images. The presence or absence of awake and/or sleep grinding and clenching, unilateral chewing, and other parafunctional activities were obtained from the dental interview. Linear and ordinal logistic regression models were used to examine the associations between the CSAs of the masseter, bruxism, and gender. The frequency of bruxism and reports of muscle or joint pain were significantly higher in women. The analysis revealed that a larger masseter CSA was significantly associated with bruxism only in men. The larger CSA was manifested only in the higher frequency bruxers. Bruxism had heterogeneous effects on the masseter muscle between genders. Although a higher prevalence of bruxism was reported by women, the larger CSA of the masseter muscle was significantly higher in bruxing men but not in women. This study emphasizes the need for a gender-specific approach when evaluating the clinical implications of bruxism on the masseter muscle.

Alzheimer's disease classification by supervised and intelligent techniques

Background Significant advancements in neuroimaging have emerged over the past decade, notably through positron emission tomography (PET) and magnetic resonance imaging (MRI) for diagnosing Alzheimer's disease (AD) and its precursor, mild cognitive impairment (MCI). Combining imaging modalities with machine learning (ML) techniques enhances diagnostic accuracy. Objective To develop predictive models using pre-treatment brain imaging data to distinguish between normal controls (NC), MCI, and AD stages, improving diagnostic precision. Methods We utilized the Alzheimer's Disease Neuroimaging Initiative database, processing 3D MRI, PET Florbetaben, and PET Flortaucipir images. Techniques included convolutional neural networks (CNN), fuzzy logic, and multi-layer perceptron (MLP). Feature extraction involved amyloid-β volume, tau protein levels, and empty space volumes. Results The fuzzy logic approach achieved a classification accuracy of 99.1%, outperforming CNN (90.67%) and MLP (94%). Integration of multimodal data significantly enhanced performance compared to single-modality approaches. Conclusions Our study demonstrates that integrating advanced ML techniques with multimodal neuroimaging can effectively classify AD stages. These findings address critical gaps in early detection and provide a foundation for future clinical applications.

Multi-scale multimodal deep learning framework for Alzheimer's disease diagnosis

Multimodal neuroimaging data, including magnetic resonance imaging (MRI) and positron emission tomography (PET), provides complementary information about the brain that can aid in Alzheimer's disease (AD) diagnosis. However, most existing deep learning methods still rely on patch-based extraction from neuroimaging data, which typically yields suboptimal performance due to its isolation from the subsequent network and does not effectively capture the varying scales of structural changes in the cerebrum. Moreover, these methods often simply concatenate multimodal data, ignoring the interactions between them that can highlight discriminative regions and thereby improve the diagnosis of AD. To tackle these issues, we develop a multimodal and multi-scale deep learning model that effectively leverages the interaction between the multimodal and multiscale of the neuroimaging data. First, we employ a convolutional neural network to embed each scale of the multimodal images. Second, we propose multimodal scale fusion mechanisms that utilize both multi-head self-attention and multi-head cross-attention, which capture global relations among the embedded features and weigh each modality's contribution to another, and hence enhancing feature extraction and interaction between each scale of MRI and PET images. Third, we introduce a cross-modality fusion module that includes a multi-head cross-attention to fuse MRI and PET data at different scales and promote global features from the previous attention layers. Finally, all the features from every scale are fused to discriminate between the different stages of AD. We evaluated our proposed method on the ADNI dataset, and the results show that our model achieves better performance than the state-of-the-art methods.

Prospective Evaluation of Accelerated Brain MRI Using Deep Learning-Based Reconstruction: Simultaneous Application to 2D Spin-Echo and 3D Gradient-Echo Sequences.

To prospectively evaluate the effect of accelerated deep learning-based reconstruction (Accel-DL) on improving brain magnetic resonance imaging (MRI) quality and reducing scan time compared to that in conventional MRI. This study included 150 participants (51 male; mean age 57.3 ± 16.2 years). Each group of 50 participants was scanned using one of three 3T scanners from three different vendors. Conventional and Accel-DL MRI images were obtained from each participant and compared using 2D T1- and T2-weighted and 3D gradient-echo sequences. Accel-DL acquisition was achieved using optimized scan parameters to reduce the scan time, with the acquired images reconstructed using U-Net-based software to transform low-quality, undersampled k-space data into high-quality images. The scan times of Accel-DL and conventional MRI methods were compared. Four neuroradiologists assessed the overall image quality, structural delineation, and artifacts using Likert scale (5- and 3-point scales). Inter-reader agreement was assessed using Fleiss' kappa coefficient. Signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) were calculated, and volumetric quantification of regional structures and white matter hyperintensities (WMHs) was performed. Accel-DL showed a mean scan time reduction of 39.4% (range, 24.2%-51.3%). Accel-DL improved overall image quality (3.78 ± 0.71 vs. 3.36 ± 0.61, P < 0.001), structure delineation (2.47 ± 0.61 vs. 2.35 ± 0.62, P < 0.001), and artifacts (3.73 ± 0.72 vs. 3.71 ± 0.69, P = 0.016). Inter-reader agreement was fair to substantial (κ = 0.34-0.50). SNR and CNR increased in Accel-DL (82.0 ± 23.1 vs. 31.4 ± 10.8, P = 0.02; 12.4 ± 4.1 vs. 4.4 ± 11.2, P = 0.02). Bland-Altman plots revealed no significant differences in the volumetric measurements of 98.2% of the relevant regions, except in the deep gray matter, including the thalamus. Five of the six lesion categories showed no significant differences in WMH segmentation, except for leukocortical lesions (r = 0.64 ± 0.29). Accel-DL substantially reduced the scan time and improved the quality of brain MRI in both spin-echo and gradient-echo sequences without compromising volumetry, including lesion quantification.

Medial patellofemoral ligament reconstruction normalizes patellar kinematics but fails to predict cartilage contact area: A prospective 3D MRI study

AbstractIntroductionThe medial patellofemoral ligament (MPFL) is the main patellar stabilizer in low knee flexion degrees (0–30°). Isolated MPFL reconstruction (MPFLr) is therefore considered the gold standard of surgical procedures for low flexion patellofemoral instabilities (PFIs). Despite excellent clinical results, little is known about the effect of MPFLr on kinematic parameters (KPs) of the patellofemoral joint in vivo. This study investigates the effect of MPFLr on KP of patellofemoral articulation, using a three‐dimensional (3D) in vivo magnetic resonance imaging (MRI) analysis at different flexion and loading positions, and analyzes the correlation of these parameters with the patellofemoral cartilage contact area (CCA).MethodsIn this prospective, matched‐pair cohort study of 30 individuals, 15 patients with low flexion PFI and 15 knee‐healthy individuals were included. Patients were analyzed pre and post‐operatively after MPFLr. MRI images were obtained at 0°, 15° and 30° with and without muscle activation, using a custom‐designed pneumatic loading device. Patellar shift, tilt and rotation were determined in 3D bone and cartilage models of each individual, guaranteeing the highest reliability. Subsequently, the KPs were correlated with patellofemoral CCA.ResultsPatients with low flexion PFI had a leg geometry of 0.5 ± 2.6° valgus and a TTTG of 11.4 ± 4.4 mm. Eleven patients had moderate (Type A/B) and 2 had severe (Type C/D) trochlear dysplasia. Without muscle activation, patients showed significantly increased patellar shift (0–30°; p0° = 0.011, p15° = 0.004 and p30° = 0.015) and tilt (15°; p15° = 0.041). Muscle activation did not compensate for maltracking in these patients, but even increased tilt and shift further in extension (p0° = 0.002 and p0° = 0.001). MPFLr statistically reduced patellofemoral tilt from 0° to 30° flexion during passive flexion and tended to approach the values of knee‐healthy individuals (pext = 0.008, p15° = 0.006 and p30° = 0.003). Post‐operatively, muscle activation led to comparable tilt and shift as in healthy individuals. Tilt, shift and rotation did not correlate with CCA neither in healthy individuals nor in pre‐ or post‐operative patients.ConclusionIsolated MPFLr can normalize patellar shift and tilt in patients with low flexion instability. Considering the influence of muscle activation, passive stabilization through MPFLr seems to be the basic precondition for physiologically active patella stabilization. The investigated KPs as easy‐to‐measure parameters in clinical practice cannot be used to assume normalized CCA for low flexion degrees. Therefore, methodologically demanding methods are still required to calculate the patellofemoral CCA.Level of EvidenceLevel II.

Correlation Between MRI Findings, ODI, and VAS Score in Lumbar Disc Herniation

Introduction. Magnetic resonance imaging (MRI) is the preferred method for diagnosing the causes of lumboischialgia, as it offers the highest sensitivity and specificity compared to other imaging techniques. In clinical practice, there is often a notable discrepancy between patients’ clinical symptoms and the radiological findings. While there are various clinical tests for lumboischialgia, the Oswestry Disability Index (ODI) and the Visual Analogue Score (VAS) are the most commonly used and reliable. This article aims to explore the correlation between patients’ subjective pain experiences and their level of disability due to lumboischialgia and disc herniation as detected by MRI. Methods. In this prospective clinical study, a total of 100 patients of both genders, aged 18 to 65 years, were included. These patients were referred for magnetic resonance imaging of the lumbar spine due to complaints of lumboischialgia. MRI of the lumbar spine was performed, and the extent of degenerative changes was evaluated. Participants completed a questionnaire regarding their subjective pain experience and functional status, and the Oswestry Disability Index (ODI) and Visual Analogue Score (VAS) were calculated. Results. The statistical analysis revealed a significant correlation between the severity of disc herniation (classification) and the intensity of spinal pain (p = 0.010), as well as with the disability index (p = 0.003). Conclusion. A statistically significant relationship was confirmed between the levels of pain and disability and the degree of intervertebral disc herniation observed on MRI images of the lumbar spine. Keywords: lumbal disk herniation, lumboischialgia, pain.

A study on the correlation between MRI-assessed breast volume asymmetry and patient satisfaction after immediate breast reconstruction in breast cancer patients

Immediate breast reconstruction provides breast cancer patients with a valuable opportunity to restore breast shape. However, post-reconstruction breast asymmetry remains a common issue that affects patient satisfaction. This study aims to quantify breast asymmetry after surgery using magnetic resonance imaging (MRI) and assess its impact on both breast satisfaction and overall outcome satisfaction, offering scientific evidence to guide improvements in preoperative evaluation. We retrospectively collected MRI images and demographic characteristics from 98 patients who underwent unilateral immediate breast reconstruction. All patients underwent breast MRI scans post-reconstruction, and the MRI images were converted into 3D models using 3D Slicer software. The implant volumes were measured and compared with the actual implant volumes inserted during surgery to assess the accuracy of the MRI measurements. Breast symmetry after surgery was quantified using MRI-measured breast volumes, and patient satisfaction was analyzed through BREAST-Q questionnaires to explore the relationship between breast symmetry, post-reconstruction breast satisfaction, and overall outcome satisfaction. The implant volumes measured in the 3D reconstructed MRI images exhibited high concordance with the volumes recorded during surgery (r = 0.980, p < 0.001). Breast asymmetry values were significantly higher in the high BMI group compared to the low BMI group (P = 0.013), while the asymmetry ratio was significantly lower in the high weight group compared to the low weight group (P = 0.006). The asymmetry ratio showed a weak correlation with breast satisfaction (r = − 0.341, P = 0.002) and outcome satisfaction (r = − 0.371, P = 0.004). When grouped by low asymmetry (asymmetry > 10%), there was no significant difference in breast satisfaction (P = 0.054) or outcome satisfaction scores (P = 0.23) between the groups. However, in cases of high asymmetry (asymmetry ratio > 40%), breast satisfaction (P = 0.003) and outcome satisfaction scores (P = 0.005) were significantly lower in the asymmetry group compared to the symmetry group. Breast volume asymmetry after immediate breast reconstruction is associated with patient satisfaction, with a significant decrease in satisfaction when asymmetry exceeds 40%. This suggests that preoperative evaluation should focus on breast symmetry to improve post-reconstruction patient satisfaction and surgical outcomes.

Automatic MRI-TRUS Fusion Technique for Transperineal Biopsy Guidance: From Preoperative Planning to Intraoperative Navigation.

Targeted and systematic transperineal biopsy of lesions guided by magnetic resonance imaging (MRI) and transrectal ultrasonography (TRUS) fusion technique may optimize the biopsy procedure and enhance the detection of prostate cancer. We described the transperineal biopsy guided by an automatic MRI-TRUS fusion technique, and evaluated the accuracy and feasibility of this method in a prospective single-center study. The proposed method focuses on automating the delineation of prostate contours in both the MRI and TRUS images, the registration and fusion of MRI and TRUS images, the generation and visualiztion of the systematic biopsy cores in their corresponding locations within the 2D and the 3D views, as well as the computation and visualiztion of needle trajectories from preoperative planning to intraoperative navigation. A total of 76 patients with clinically suspected prostate cancer underwent systematic (SBx) and targeted (TBx) biopsies, all performed by a single urologist with more than 10 years of experience. The detection rates of prostate cancer (PCa) and clinically significant prostate cancer (csPCa) were recorded. We also measured preoperative registration time, duration of the overall surgical procedure, and postoperative complication rates within the first week following the surgery. Descriptive analyses were presented in this study. PCa was identified in 73.7% (56/76) of the subjects, while csPCa was identified in 61.8% (47/76). The preoperative registration time was 5.0 min (IQR: 4.4-6.0), while the overall surgery duration was 24.8 min (IQR: 23.2-27.2). Postoperatively, 12 patients experienced immediate hematuria, and one patient reported dysuria 1 day following surgery. The automatic MRI-TRUS fusion technique for transperineal biopsy is feasible and safe, with preoperative planning to intraoperative navigation it offering convenient and efficient preoperative preparation and surgical procedure.

The Role of Diffusion-Weighted Magnetic Resonance Imaging in Differentiating Benign and Malignant Breast Lesions: A Comprehensive Review

Background: Breast magnetic resonance imaging (MRI) has emerged as a vital tool for detecting and characterizing breast lesions, particularly in differentiating benign from malignant tumors. This review explores the role of diffusion-weighted imaging (DWI) in breast MRI, highlighting its advantages over conventional imaging techniques. Methods: A systematic literature search was conducted across multiple databases, including PubMed and Scopus, to gather studies assessing DWI's sensitivity and specificity in breast cancer detection. Results: The findings indicate that DWI provides high diagnostic accuracy, with a pooled sensitivity of 86.5% and specificity of 83.5%, comparable to dynamic contrast-enhanced MRI (DCE-MRI). Notably, utilizing apparent diffusion coefficient (ADC) thresholds can minimize unnecessary biopsies without compromising sensitivity. However, challenges remain in achieving standardization in DWI protocols, particularly regarding b-value selection and imaging techniques, which impact ADC accuracy. Advanced methodologies such as intravoxel incoherent motion (IVIM) and diffusion kurtosis imaging (DKI) have shown promise in enhancing diagnostic capabilities by providing deeper insights into tissue microstructure. Conclusions: This review underscores the importance of ongoing standardization efforts and collaboration among institutions to improve the reliability and applicability of DWI in clinical settings. In conclusion, DWI represents a transformative approach to breast cancer management, allowing for enhanced differentiation of lesions while reducing the need for invasive procedures. Further research and standardization are essential to optimize its use in routine clinical practice.

SMDFnet: Saliency multiscale dense fusion network for MRI and CT image fusion.

MRI-CT image fusion technology combines magnetic resonance imaging (MRI) and computed tomography (CT) imaging to provide more comprehensive and accurate image information. This fusion technology can play an important role in medical diagnosis and surgical planning. However, there are several issues with current MRI-CT image fusion, such as the presence of artifacts in both MRI and CT images, which may affect the quality and accuracy of the images during the fusion process. The current fusion strategy of MRI and CT images is complex and prone to losing a large amount of information, which still needs further research and improvement. This article proposes a saliency multi-scale dense fusion network for MRI and CT image fusion to address existing issues. The proposed method first uses a pretrained network to extract depth information from MRI and CT images, which can effectively overcome the noise and artifacts caused by directly training and extracting features and improve the saliency information in the original images. Then, a multi-scale dense network is used to further enhance the extracted pre training features and achieve fusion, and multiple loss functions are used to optimize the network and improve the fusion quality. The experimental results show that the fusion results of the proposed method perform better than the objective indicators of the reference method, while retaining more significant information.

Diagnostic performance of depth of invasion, thickness, and styloglossus and hyoglossus muscle invasion on magnetic resonance imaging in predicting potential neck lymph node metastasis in clinical N0 tongue cancer.

To evaluate previously reported quantitative (tumor thickness 11mm and depth of invasion [DOI] 7.5mm) and qualitative (styloglossus/hyoglossus muscle invasion [SHMI]) magnetic resonance imaging (MRI) parameters for predicting occult neck node metastasis in clinical N0 oral tongue squamous cell carcinoma. This single-center retrospective study included 76 patients. MRI images were independently reviewed by two radiologists for tumor thickness, DOI, and SHMI. Statistical analysis assessed the predictive capability of these parameters for 2-year potential lymph node metastasis. Among the 76 cases, 30.2% developed 2-year potential lymph node metastasis. For tumor thickness ≥ 11mm, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy were 0.46, 0.68, 0.37, 0.75, and 0.61, respectively. DOI ≥ 7.5mm exhibited a sensitivity, specificity, PPV, NPV, and accuracy of 0.73, 0.59, 0.42, 0.84, and 0.63, respectively. SHMI demonstrated a sensitivity, specificity, PPV, NPV, and accuracy of 0.87, 0.51, 0.46, 0.89, and 0.63, respectively. DOI ≥ 7.5mm and SHMI demonstrated comparable diagnostic accuracy in predicting neck metastasis, surpassing tumor thickness of > 11mm. These findings underscore their potential utility in guiding decisions concerning elective neck dissection.

Exploring Generative Adversarial Network-Based Augmentation of Magnetic Resonance Brain Tumor Images

Background: A generative adversarial network (GAN) has gained popularity as a data augmentation technique in the medical field due to its efficiency in creating synthetic data for different machine learning models. In particular, the earlier literature suggests that the classification accuracy of a convolutional neural network (CNN) used for detecting brain tumors in magnetic resonance imaging (MRI) images increases when GAN-generated images are included in the training data together with the original images. However, there is little research about how the exact number of GAN-generated images and their ratio to the original images affects the results obtained. Materials and methods: Here, by using 1000 original images from a public repository with MRI images of patients with or without brain tumors, we built a GAN model to create synthetic brain MRI images. A modified U-Net CNN is trained multiple times with different training datasets and its classification accuracy is evaluated from a separate test set of another 1000 images. The Mann–Whitney U test is used to estimate whether the differences in the accuracy caused by different choices of training data are statistically significant. Results: According to our results, the use of GAN augmentation only sometimes produces a significant improvement. For instance, the classification accuracy significantly increases when 250–750 GAN-generated images are added to 1000 original images (p-values ≤ 0.0025) but decreases when 10 GAN-generated images are added to 500 original images (p-value: 0.03). Conclusions: Whenever GAN-based augmentation is used, the number of GAN-generated images should be carefully considered while accounting for the number of original images.