Abstract Objectives: Coupled with functional magnetic resonance imaging (fMRI), machine learning (ML), enables advanced pattern recognition and predictive modeling, driving progress in cognitive neuroscience and clinical diagnostics. Scientometrics offers a framework for examining the evolving intersection of fMRI and ML, uncovering knowledge gaps, trends, and collaborations. Materials and Methods: A scientometric analysis was conducted using data from the Scopus database, focusing on studies combining fMRI and ML. After filtering for peer-reviewed articles, reviews, and editorials, eligible studies were preprocessed using KNIME tools to standardize text. Data were analyzed with the Bibliometrix package in R-Studio and VOS-viewer to generate thematic maps, citation networks, and collaboration visualizations. These analyses identified key trends, influential contributors, and research gaps, offering a structured view of this interdisciplinary field. Results: The Scopus search yielded 4657 documents, with 2639 included after filtering. Scientific production in this field has grown steadily since 2004, with an upward trend resuming after 2021. Thematic mapping identified 4 types of research topics: motor (hot and well-developed), basic (foundational and cross-cutting), niche (specialized), and emerging or declining (less developed). Key active themes include “machine learning” and “functional connectivity,” while advanced techniques like generative adversarial networks are gaining attention. Collaboration among 139 researchers and 47 countries highlighted the United States and China as leaders in publications and citations. Among 582 journals, Neuroimage had the highest number of articles and citations. Conclusions: The evaluated studies depict steady progress, with notable contributions to clinical diagnostics, cognitive neuroscience, and therapy development. Thematic analysis highlights foundational methods, such as fMRI and support vector machines, alongside emerging innovations like generative adversarial networks and dynamic functional connectomics.
OBJECTIVES:Quantitative multinuclear MRI of bone is hindered by magnetic field (B0 and B1) inhomogeneities, calibration inconsistencies, and segmentation challenges. We developed a comprehensive and versatile MATLAB-based graphical interface that integrates voxel-wise B1 correction, auto and manual segmentation, coregistration, advanced visualization, and quantitative analysis with in-scan dual-density calibration, to generate reproducible bone matrix and mineral density maps from 1H and 31P ZTE MRI. MATERIALS AND METHODS:The postprocessing package was developed in MATLAB using a modular architecture comprising multiple m-files for B1 field correction, B0 bias correction, auto and manual image registration, and segmentation for quantitative data analysis. Otsu-based thresholding with min-max intensity normalization was employed for tissue segmentation and bias correction. Data from control, ovariectomized, and vitamin D-deficient rat femurs were analyzed using normality (Shapiro-Wilk) and variance (Levene) tests. Between-group comparisons used the Kruskal-Wallis test or analysis of variance with Bonferroni or Tukey post hoc tests, respectively. Cross-modality correlation analyses were conducted between MRI-derived measures and reference measures (µCT and gravimetry) using Pearson and Spearman coefficients. RESULTS:MRI-derived mineral density strongly correlated with µCT BMD (cortical [P = 0.22], trabecular [P = 0.31]), and the MRI matrix density correlated with the gravimetric data (cortical [P = 0.38], trabecular [P = 0.57]). No significant differences were observed between modalities for either cortical or trabecular bone. CONCLUSION:This standardized pipeline enables reproducible, calibrated bone density mapping for data sizes ranging from 64 × 64 × 64 to 512 × 512 × 512, with B1 and B0 corrections assessing matrix and mineral densities. Its implementation as a user-guided graphical user interface promotes adoption for preclinical and clinical quantitative bone imaging across experimental conditions.
Introduction:Breast cancer is a global issue impacting women's well-being, highlighting the importance of early detection to improve treatment outcomes and decrease mortality rates. This study aimed to assess various AI methodologies to classify breast images into normal, benign, and malignant. METHODS:A hybrid of 4 CNN-pertained networks-Res-Net18, Mobile-Net, Shuffle-Net, and Inception-V3-were applied on 269 mammograms and 267 dynamic contrast-enhanced MRI examinations. Transfer learning was used, adapting the last fully connected layers to classify 3 classes (normal, benign, and malignant) resulting in 12 features. Support vector machine was employed to categorize images. Classifier performances were evaluated using the confusion matrix, accuracy, precision, sensitivity, specificity, F1 score, and area under the receiver operating curve (AUC-ROC). RESULTS:Res-Net model achieved the highest accuracy, sensitivity, and specificity of 90.89%, 90.93%, and 95.39%, respectively. Whereas Shuffle-Net displayed the lowest accuracy of 84.76%. The AUC ranged between 0.95 and 0.97 among pretrained networks while it was higher (0.99) for the hybrid model. For MRI image classification, the Mobile-net network recorded the highest accuracy, sensitivity, and specificity of 88.55%, 88.49%, and 94.22%, respectively, while the Res-Net exhibited the lowest accuracy of 84.35%. The AUC ranged between 0.94 and 0.96 among pretrained networks while it was higher (0.98) for the Hybrid model. CONCLUSION:ResNet-18 showed the most optimal model for extracting features from mammograms compared with other CNN networks (Mobile-Net, Shuffle-Net, and Inception-V3) while Mobile-Net model was the most suitable in MRI. The effectiveness of deep learning in accurately classifying mammograms and MRI images can be improved by using a hybrid model.
BACKGROUND AND AIM:Brain metastases (BMs) are the most common intracranial malignancy, often arising from lung, breast, and melanoma cancers. Receptor tyrosine kinases, such as EGFR and HER2, drive tumor progression and resistance to therapy. Noninvasive detection of these biomarkers, especially in brain metastases, is crucial due to challenges with traditional biopsy methods. This systematic review and meta-analysis assess machine learning (ML)-based models for detecting EGFR mutations and HER2 overexpression in metastatic brain adenocarcinoma using MRI-derived radiomic features. METHODS:A systematic review and meta-analysis were conducted following PRISMA 2020 guidelines. Studies were identified via PubMed, Scopus, and Web of Science, focusing on ML applications to MRI radiomics for detecting EGFR and HER2 in brain metastases. Data on study design, imaging modality, model type, sample size, and performance metrics were extracted. Subgroup analyses were performed by model type (deep learning vs. classical ML) and sample size (<150 vs. ≥150 participants). A random-effects model was used to pool performance metrics, and risk of bias was assessed using the RoB 2 tool. STATA version 18 and Python 3.10 were used for analyses and visualizations. RESULTS:Of 383 identified studies, 31 (7925 participants) met the inclusion criteria. The pooled analysis showed strong diagnostic performance: AUC = 0.84, accuracy = 0.86, and sensitivity = 0.83. Subgroup analysis revealed higher AUC and accuracy in deep learning models compared with classical ML. Sensitivity analysis also indicated improved AUC in studies with larger sample sizes (≥150), though variability remained. No evidence of heterogeneity or publication bias was detected. CONCLUSION:ML models demonstrate strong diagnostic performance for detecting EGFR and HER2 in metastatic brain adenocarcinoma, supporting their potential as noninvasive diagnostic tools. However, these findings should be interpreted considering methodological heterogeneity and the limited use of external validation. Further prospective, multicenter studies are warranted to confirm their clinical applicability and generalizability.
Magnetic resonance imaging (MRI) is essential for diagnosis but often induces anxiety, especially in claustrophobic patients, potentially affecting image quality. This study compared oxygen saturation, heart rate, and anxiety levels between claustrophobic and non-claustrophobic patients undergoing closed and open MRI in Erbil, Iraq. The comparative study was conducted from October 2024 to April 2025 in the Radiology Departments of Consultant Medical City and Top Med Medical Complex Centers in Erbil using purposive sampling. The questionnaire contained 3 sections: sociodemographic variables, the Claustrophobia Questionnaire, and the State-Trait Anxiety Inventory-State Subscale. Physiological measures (oxygen saturation and heart rate) were recorded at 3 timepoints: pre-, mid-, and post-MRI. Statistical analyses included one-way ANOVA, repeated measures ANOVA, post hoc tests, and both univariate and multiple linear regression, using SPSS version 26. A total of 125 participants were involved in the study. The mean anxiety score was moderate, with higher levels in claustrophobic patients. Claustrophobia scores also fell within the moderate range, indicating psychological discomfort during the MRI procedure. Physiological measurements showed that claustrophobic patients, particularly those undergoing closed MRI, experienced elevated heart rates and reduced oxygen saturation compared to non-claustrophobic individuals. Statistical analysis indicated a strong positive association between anxiety and claustrophobia, with scan entry direction, age, and sex also being significant predictors of claustrophobic responses. Claustrophobic patients undergoing closed MRI experience increased anxiety and physiological distress. Open MRI systems and pre-scan anxiety screening are recommended to enhance patient comfort and diagnostic outcomes.
