Tomography

Background/Objectives: Mummy studies allow to reconstruct the characteristic of a population in a specific spatiotemporal context, in terms of living conditions, pathologies and death. Radiology represents an efficient diagnostic technique able to establish the preservation state of mummified organs and to estimate the patient's pathological conditions. However, the radiological approach shows some limitations. Although bone structures are easy to differentiate, soft tissue components are much more challenging, especially when they overlap. For this reason, computed tomography, a well-established approach that achieves optimal image contrast and three-dimensional reconstruction, has been introduced. This original article focuses attention on the role of virtual dissection as a promising technology for exploring human mummy anatomy and considers the potential of environmental scanning electron microscopy and X-ray spectroscopy as complementary approaches useful to understand the state of preservation of mummified remains. Methods: Ancient mummy corps have been analyzed through Anatomage Table 10 and environmental scanning electron microscope equipped with X-ray spectrometer; Results: Anatomage Table 10 through various volumetric renderings allows us to describe spine alteration due to osteoarthritis, dental state, and other clinical-pathological characteristics of different mummies. Environmental scanning electron microscope, with the advantage of observing mummified samples without prior specimen preparation, details on the state of tissue fragments. Skin, tendon and muscle show a preserved morphology and keratinocytes, collagen fibers and tendon structures are easily recognizable. Furthermore, X-ray spectrometer reveals in our tissue remains, the presence of compounds related to soil contamination. This investigation identifies a plethora of organic and inorganic substances where the mummies were found, providing crucial information about the mummification environment. Conclusions: These morphological and analytical techniques make it possible to study mummified bodies and describe their anatomical details in real size, in a non-invasive and innovative way, demonstrating that these interdisciplinary approaches could have great potential for improving knowledge in the study of ancient corpses. Full article

Background: Anesthesia can significantly impact positron emission tomography (PET) neuroimaging in preclinical studies. Therefore, understanding these effects is crucial for accurate interpretation of the results. In this experiment, we investigate the effect of [¹⁸F]-labeled glucose analog fluorodeoxyglucose ([¹⁸F]FDG) uptake in the brains of rats anesthetized with two commonly used anesthetics for rodents: isoflurane, an inhalation anesthetic, and Hypnorm–Dormicum, a combination injection anesthetic. Materials and Methods: Female adult Sprague Dawley rats were randomly assigned to one of two anesthesia groups: isoflurane or Hypnorm–Dormicum. The rats were submitted to dynamic [¹⁸F]FDG PET scan. The whole brain [¹⁸F]FDG standard uptake value (SUV) and the brain voxel-based analysis were performed. Results: The dynamic [¹⁸F]FDG data revealed that the brain SUV was 38% lower in the isoflurane group after 40 min of image (2.085 ± 0.3563 vs. 3.369 ± 0.5577, p = 0.0008). In voxel-based analysis between groups, the maps collaborate with SUV data, revealing a reduction in [¹⁸F]FDG uptake in the isoflurane group, primarily in the cortical regions, with additional small increases observed in the midbrain and cerebellum. Discussion and Conclusions: The observed differences in [¹⁸F]FDG uptake in the brain may be attributed to variations in metabolic activity. These results underscore the necessity for careful consideration of anesthetic choice and its impact on neuroimaging outcomes in future research. Full article

Objectives: Accurate kidney and tumor segmentation of computed tomography (CT) scans is vital for diagnosis and treatment, but manual methods are time-consuming and inconsistent, highlighting the value of AI automation. This study develops a fully automated AI model using vision transformers (ViTs) and convolutional neural networks (CNNs) to detect and segment kidneys and kidney tumors in Contrast-Enhanced (CECT) scans, with a focus on improving sensitivity for small, indistinct tumors. Methods: The segmentation framework employs a ViT-based model for the kidney organ, followed by a 3D UNet model with enhanced connections and attention mechanisms for tumor detection and segmentation. Two CECT datasets were used: a public dataset (KiTS23: 489 scans) and a private institutional dataset (Private: 592 scans). The AI model was trained on 389 public scans, with validation performed on the remaining 100 scans and external validation performed on all 592 private scans. Tumors were categorized by TNM staging as small (≤4 cm) (KiTS23: 54%, Private: 41%), medium (>4 cm to ≤7 cm) (KiTS23: 24%, Private: 35%), and large (>7 cm) (KiTS23: 22%, Private: 24%) for detailed evaluation. Results: Kidney and kidney tumor segmentations were evaluated against manual annotations as the reference standard. The model achieved a Dice score of 0.97 ± 0.02 for kidney organ segmentation. For tumor detection and segmentation on the KiTS23 dataset, the sensitivities and average false-positive rates per patient were as follows: 0.90 and 0.23 for small tumors, 1.0 and 0.08 for medium tumors, and 0.96 and 0.04 for large tumors. The corresponding Dice scores were 0.84 ± 0.11, 0.89 ± 0.07, and 0.91 ± 0.06, respectively. External validation on the private data confirmed the model’s effectiveness, achieving the following sensitivities and average false-positive rates per patient: 0.89 and 0.15 for small tumors, 0.99 and 0.03 for medium tumors, and 1.0 and 0.01 for large tumors. The corresponding Dice scores were 0.84 ± 0.08, 0.89 ± 0.08, and 0.92 ± 0.06. Conclusions: The proposed model demonstrates consistent and robust performance in segmenting kidneys and kidney tumors of various sizes, with effective generalization to unseen data. This underscores the model’s significant potential for clinical integration, offering enhanced diagnostic precision and reliability in radiological assessments. Full article

Background: Computed tomography scans are widely used in everyday medical practice due to speed, image reliability, and detectability of a wide range of pathologies. Each scan exposes the patient to a radiation dose, and performing a fast estimation of the effective dose (E) is an important step for radiological safety. The aim of this work is to estimate E from patient and CT acquisition parameters in the absence of a dose-tracking software exploiting machine learning. Methods: In total, 69,037 CT acquisitions were collected with the dose-tracking software (DTS) available at our institution. E calculated by DTS was chosen as the target value for prediction. Different machine learning algorithms were selected, optimizing parameters to achieve the best performance for each algorithm. Effective dose was also estimated using DLP and k-factors, and with multiple linear regression. Mean absolute error (MAE, mean absolute percentage error (MAPE), and R² were used to evaluate predictions in the test set and in an external dataset of 3800 acquisitions. Results: The random forest regressor (MAE: 0.416 mSv; MAPE: 7%; and R²: 0.98) showed best performances over the neural network and the support vector machine. However, all three machine learning algorithms outperformed effective dose estimation using k-factors (MAE: 2.06; MAPE: 26%) or multiple linear regression (MAE: 0.98; MAPE: 44.4%). The random forest regressor on the external dataset showed an MAE of 0.215 mSv and an MAPE of 7.1%. Conclusions: Our work demonstrated that machine learning models trained with data calculated by a dose-tracking software can provide good estimates of the effective dose just from patient and scanner parameters, without the need for a Monte Carlo approach. Full article

Due to the increasing number of people working at computers in professional settings, the incidence of lumbar disc herniation is increasing. Background/Objectives: The early diagnosis and treatment of lumbar disc herniation is much more likely to yield favorable results, allowing the hernia to be treated before it develops further. The aim of this study was to classify lumbar disc herniations in a computer-aided, fully automated manner using magnetic resonance images (MRIs). Methods: This study presents a hybrid method integrating residual network (ResNet50), grey wolf optimization (GWO), and machine learning classifiers such as multi-layer perceptron (MLP) and support vector machine (SVM) to improve classification performance. The proposed approach begins with feature extraction using ResNet50, a deep convolutional neural network known for its robust feature representation capabilities. ResNet50’s residual connections allow for effective training and high-quality feature extraction from input images. Following feature extraction, the GWO algorithm, inspired by the social hierarchy and hunting behavior of grey wolves, is employed to optimize the feature set by selecting the most relevant features. Finally, the optimized feature set is fed into machine learning classifiers (MLP and SVM) for classification. The use of various activation functions (e.g., ReLU, identity, logistic, and tanh) in MLP and various kernel functions (e.g., linear, rbf, sigmoid, and polynomial) in SVM allows for a thorough evaluation of the classifiers’ performance. Results: The proposed methodology demonstrates significant improvements in metrics such as accuracy, precision, recall, and F1 score, outperforming traditional approaches in several cases. These results highlight the effectiveness of combining deep learning-based feature extraction with optimization and machine learning classifiers. Conclusions: Compared to other methods, such as capsule networks (CapsNet), EfficientNetB6, and DenseNet169, the proposed ResNet50-GWO-SVM approach achieved superior performance across all metrics, including accuracy, precision, recall, and F1 score, demonstrating its robustness and effectiveness in classification tasks. Full article

Background/Objectives: Calculating the radiation dose from CT in ¹⁸F-PET/CT examinations poses a significant challenge. The objective of this study is to develop a deep learning-based automated program that standardizes the measurement of radiation doses. Methods: The torso CT was segmented into six distinct regions using TotalSegmentator. An automated program was employed to extract the necessary information and calculate the effective dose (ED) of PET/CT. The accuracy of our automated program was verified by comparing the EDs calculated by the program with those determined by a nuclear medicine physician (n = 30). Additionally, we compared the EDs obtained from an older PET/CT scanner with those from a newer PET/CT scanner (n = 42). Results: The CT ED calculated by the automated program was not significantly different from that calculated by the nuclear medicine physician (3.67 ± 0.61 mSv and 3.62 ± 0.60 mSv, respectively, p = 0.7623). Similarly, the total ED showed no significant difference between the two calculation methods (8.10 ± 1.40 mSv and 8.05 ± 1.39 mSv, respectively, p = 0.8957). A very strong correlation was observed in both the CT ED and total ED between the two measurements (r² = 0.9981 and 0.9996, respectively). The automated program showed excellent repeatability and reproducibility. When comparing the older and newer PET/CT scanners, the PET ED was significantly lower in the newer scanner than in the older scanner (4.39 ± 0.91 mSv and 6.00 ± 1.17 mSv, respectively, p < 0.0001). Consequently, the total ED was significantly lower in the newer scanner than in the older scanner (8.22 ± 1.53 mSv and 9.65 ± 1.34 mSv, respectively, p < 0.0001). Conclusions: We successfully developed an automated program for calculating the ED of torso ¹⁸F-PET/CT. By integrating a deep learning model, the program effectively eliminated inter-operator variability. Full article

Background: Medical imagesegmentation is an essential step in both clinical and research applications, and automated segmentation models—such as TotalSegmentator—have become ubiquitous. However, robust methods for validating the accuracy of these models remain limited, and manual inspection is often necessary before the segmentation masks produced by these models can be used. Methods: To address this gap, we have developed a novel validation framework for segmentation models, leveraging data augmentation to assess model consistency. We produced segmentation masks for both the original and augmented scans, and we calculated the alignment metrics between these segmentation masks. Results: Our results demonstrate strong correlation between the segmentation quality of the original scan and the average alignment between the masks of the original and augmented CT scans. These results were further validated by supporting metrics, including the coefficient of variance and the average symmetric surface distance, indicating that agreement with augmented-scan segmentation masks is a valid proxy for segmentation quality. Conclusions: Overall, our framework offers a pipeline for evaluating segmentation performance without relying on manually labeled ground truth data, establishing a foundation for future advancements in automated medical image analysis. Full article

Using a pediatric-focused lens, this review article briefly summarizes the presentation of several demyelinating and neuroinflammatory diseases using conventional magnetic resonance imaging (MRI) sequences, such as T1-weighted with and without an exogenous gadolinium-based contrast agent, T2-weighted, and fluid-attenuated inversion recovery (FLAIR). These conventional sequences exploit the intrinsic properties of tissue to provide a distinct signal contrast that is useful for evaluating disease features and monitoring treatment responses in patients by characterizing lesion involvement in the central nervous system and tracking temporal features with blood–brain barrier disruption. Illustrative examples are presented for pediatric-onset multiple sclerosis and neuroinflammatory diseases. This work also highlights findings from advanced MRI techniques, often infrequently employed due to the challenges involved in acquisition, post-processing, and interpretation, and identifies the need for future studies to extract the unique information, such as alterations in neurochemistry, disruptions of structural organization, or atypical functional connectivity, that may be relevant for the diagnosis and management of disease. Full article

Background: This study aimed to assess the interobserver variability of semi-automatic diameter and volumetric measurements versus manual diameter measurements for small lung nodules identified on computed tomography scans. Methods: The radiological patient database was searched for CT thorax examinations with at least one noncalcified solid nodule (∼3–10 mm). Three radiologists with four to six years of experience evaluated each nodule in accordance with the Fleischner Society guidelines using standard diameter measurements, semi-automatic lesion diameter measurements, and volumetric assessments. Spearman’s correlation coefficient measured intermeasurement agreement. We used descriptive Bland–Altman plots to visualize agreement in the measured data. Potential discrepancies were analyzed. Results: We studied a total of twenty-six nodules. Spearman’s test showed that there was a much stronger relationship (p < 0.05) between reviewers for the semi-automatic diameter and volume measurements (avg. r = 0.97 ± 0.017 and 0.99 ± 0.005, respectively) than for the manual method (avg. r = 0.91 ± 0.017). In the Bland–Altman test, the semi-automatic diameter measure outperformed the manual method for all comparisons, while the volumetric method had better results in two out of three comparisons. The incidence of reviewers modifying the software’s automatic outline varied between 62% and 92%. Conclusions: Semi-automatic techniques significantly reduced interobserver variability for small solid nodules, which has important implications for diagnostic assessments and screening. Both the semi-automatic diameter and semi-automatic volume measurements showed improvements over the manual measurement approach. Training could further diminish observer variability, given the considerable diversity in the number of adjustments among reviewers. Full article

Objectives: We evaluated the noise reduction effects of deep learning reconstruction (DLR) and hybrid iterative reconstruction (HIR) in brain computed tomography (CT). Methods: CT images of a 16 cm dosimetry phantom, a head phantom, and the brains of 11 patients were reconstructed using filtered backprojection (FBP) and various levels of DLR and HIR. The slice thickness was 5, 2.5, 1.25, and 0.625 mm. Phantom imaging was also conducted at various tube currents. The noise reduction ratio was calculated using FBP as the reference. For patient imaging, overall image quality was visually compared between DLR and HIR images that exhibited similar noise reduction ratios. Results: The noise reduction ratio increased with increasing levels of DLR and HIR in phantom and patient imaging. For DLR, noise reduction was more pronounced with decreasing slice thickness, while such thickness dependence was less evident for HIR. Although the noise reduction effects of DLR were similar between the head phantom and patients, they differed for the dosimetry phantom. Variations between imaging objects were small for HIR. The noise reduction ratio was low at low tube currents for the dosimetry phantom using DLR; otherwise, the influence of the tube current was small. In terms of visual image quality, DLR outperformed HIR in 1.25 mm thick images but not in thicker images. Conclusions: The degree of noise reduction using DLR depends on the slice thickness, tube current, and imaging object in addition to the level of DLR, which should be considered in the clinical use of DLR. DLR may be particularly beneficial for thin-slice imaging. Full article

This research introduces BAE-ViT, a specialized vision transformer model developed for bone age estimation (BAE). This model is designed to efficiently merge image and sex data, a capability not present in traditional convolutional neural networks (CNNs). BAE-ViT employs a novel data fusion method to facilitate detailed interactions between visual and non-visual data by tokenizing non-visual information and concatenating all tokens (visual or non-visual) as the input to the model. The model underwent training on a large-scale dataset from the 2017 RSNA Pediatric Bone Age Machine Learning Challenge, where it exhibited commendable performance, particularly excelling in handling image distortions compared to existing models. The effectiveness of BAE-ViT was further affirmed through statistical analysis, demonstrating a strong correlation with the actual ground-truth labels. This study contributes to the field by showcasing the potential of vision transformers as a viable option for integrating multimodal data in medical imaging applications, specifically emphasizing their capacity to incorporate non-visual elements like sex information into the framework. This tokenization method not only demonstrates superior performance in this specific task but also offers a versatile framework for integrating multimodal data in medical imaging applications. Full article

Background/Objectives: Breast cancer is a leading cause of mortality among women in Taiwan and globally. Non-invasive imaging methods, such as mammography and ultrasound, are critical for early detection, yet standalone modalities have limitations in regard to their diagnostic accuracy. This study aims to enhance breast cancer detection through a cross-modality fusion approach combining mammography and ultrasound imaging, using advanced convolutional neural network (CNN) architectures. Materials and Methods: Breast images were sourced from public datasets, including the RSNA, the PAS, and Kaggle, and categorized into malignant and benign groups. Data augmentation techniques were used to address imbalances in the ultrasound dataset. Three models were developed: (1) pre-trained CNNs integrated with machine learning classifiers, (2) transfer learning-based CNNs, and (3) a custom-designed 17-layer CNN for direct classification. The performance of the models was evaluated using metrics such as accuracy and the Kappa score. Results: The custom 17-layer CNN outperformed the other models, achieving an accuracy of 0.964 and a Kappa score of 0.927. The transfer learning model achieved moderate performance (accuracy 0.846, Kappa 0.694), while the pre-trained CNNs with machine learning classifiers yielded the lowest results (accuracy 0.780, Kappa 0.559). Cross-modality fusion proved effective in leveraging the complementary strengths of mammography and ultrasound imaging. Conclusions: This study demonstrates the potential of cross-modality imaging and tailored CNN architectures to significantly improve diagnostic accuracy and reliability in breast cancer detection. The custom-designed model offers a practical solution for early detection, potentially reducing false positives and false negatives, and improving patient outcomes through timely and accurate diagnosis. Full article

Background: Although it has been noticed that depressed patients show differences in processing emotions, the precise neural modulation mechanisms of positive and negative emotions remain elusive. FMRI is a cutting-edge medical imaging technology renowned for its high spatial resolution and dynamic temporal information, making it particularly suitable for the neural dynamics of depression research. Methods: To address this gap, our study firstly leveraged fMRI to delineate activated regions associated with positive and negative emotions in healthy individuals, resulting in the creation of the positive emotion atlas (PEA) and the negative emotion atlas (NEA). Subsequently, we examined neuroimaging changes in depression patients using these atlases and evaluated their diagnostic performance based on machine learning. Results: Our findings demonstrate that the classification accuracy of depressed patients based on PEA and NEA exceeded 0.70, a notable improvement compared to the whole-brain atlases. Furthermore, ALFF analysis unveiled significant differences between depressed patients and healthy controls in eight functional clusters during the NEA, focusing on the left cuneus, cingulate gyrus, and superior parietal lobule. In contrast, the PEA revealed more pronounced differences across fifteen clusters, involving the right fusiform gyrus, parahippocampal gyrus, and inferior parietal lobule. Conclusions: These findings emphasize the complex interplay between emotion modulation and depression, showcasing significant alterations in both PEA and NEA among depression patients. This research enhances our understanding of emotion modulation in depression, with implications for diagnosis and treatment evaluation. Full article

Evaluating altered mental status and suspected meningeal disorders in children often begins with imaging, typically before a lumbar puncture. The challenge is that meningeal enhancement is a common finding across a range of pathologies, making diagnosis complex. This review proposes a categorization of meningeal diseases based on their predominant imaging characteristics. It includes a detailed description of the clinical and imaging features of various conditions that lead to leptomeningeal or pachymeningeal enhancement in children and adolescents. These conditions encompass infectious meningitis (viral, bacterial, tuberculous, algal, and fungal), autoimmune diseases (such as anti-MOG demyelination, neurosarcoidosis, Guillain-Barré syndrome, idiopathic hypertrophic pachymeningitis, and NMDA-related encephalitis), primary and secondary tumors (including diffuse glioneuronal tumor of childhood, primary CNS rhabdomyosarcoma, primary CNS tumoral metastasis, extracranial tumor metastasis, and lymphoma), tumor-like diseases (Langerhans cell histiocytosis and ALK-positive histiocytosis), vascular causes (such as pial angiomatosis, ANCA-related vasculitis, and Moyamoya disease), and other disorders like spontaneous intracranial hypotension and posterior reversible encephalopathy syndrome. Despite the nonspecific nature of imaging findings associated with meningeal lesions, narrowing down the differential diagnoses is crucial, as each condition requires a tailored and specific treatment approach. Full article

Objective: Image-guided diagnosis and treatment of lung lesions is an active area of research. With the growing number of solutions proposed, there is also a growing need to establish a standard for the evaluation of these solutions. Thus, realistic phantom and preclinical environments must be established. Realistic study environments must include implanted lung nodules that are morphologically similar to real lung lesions under X-ray imaging. Methods: Various materials were injected into a phantom swine lung to evaluate the similarity to real lung lesions in size, location, density, and grayscale intensities in X-ray imaging. A combination of $n$-butyl cyanoacrylate (n-BCA) and ethiodized oil displayed radiopacity that was most similar to real lung lesions, and various injection techniques were evaluated to ensure easy implantation and to generate features mimicking malignant lesions. Results: The techniques used generated implanted nodules with properties mimicking solid nodules with features including pleural extensions and spiculations, which are typically present in malignant lesions. Using only n-BCA, implanted nodules mimicking ground glass opacity were also generated. These results are condensed into a set of recommendations that prescribe the materials and techniques that should be used to reproduce these nodules. Conclusions: Generated recommendations on the use of n-BCA and ethiodized oil can help establish a standard for the evaluation of new image-guided solutions and refinement of algorithms in phantom and animal studies with realistic nodules. Full article

Background: Femoroacetabular impingement (FAI) is a condition caused by abnormal contact between the femur head and the acetabulum, which damages the labrum and articular cartilage. While the prevalence and the type of impingement may vary across human groups, the variability among populations with short height or with a high prevalence of overweight has not yet been explored. Latin American studies have rarely been conducted in reference to this condition, including the Mayan and mestizo populations from the Yucatan Peninsula. Objective: We aimed to describe the prevalence of morphological changes in femoroacetabular impingement by measuring radiological angles in abdominopelvic tomography studies in a sample of patients from a population with short height. Methods: In this prospective study, patients with programmed abdominopelvic tomography unrelated to femoroacetabular impingement but with consistent symptoms were included. Among the 98 patients, the overall prevalence of unrelated femoroacetabular impingement was 47%, and the pincer-type was the most frequent. The cam-type occurred more frequently among individuals with taller stature compared to their peers. Alpha and Wiberg angles predicted cam- and pincer-type, respectively, with over 0.95 area under the curve values in ROC analyses. The inter-rater agreement in the study was >91%. Conclusions: In a patient population from Yucatan, Mexico, attending ambulatory consultations unrelated to femoroacetabular impingement, an overall morphological changes prevalence of 47% was observed. Angle measurements using tomographic techniques can be used to predict cam- and pincer-type femoroacetabular impingement. Average stature was observed to be shorter in patients with cam-type femoroacetabular impingement, but body mass index did not vary between groups. Full article

Background: Accurate reconstruction of internal temperature fields from surface temperature data is critical for applications such as non-invasive thermal imaging, particularly in scenarios involving small temperature gradients, like those in the human body. Methods: In this study, we employed 3D convolutional neural networks (CNNs) to predict internal temperature fields. The network’s performance was evaluated under both ideal and non-ideal conditions, incorporating noise and background temperature variations. A physics-informed loss function embedding the heat equation was used in conjunction with statistical uncertainty during training to simulate realistic scenarios. Results: The CNN achieved high accuracy for small phantoms (e.g., 10 cm in diameter). However, under non-ideal conditions, the network’s predictive capacity diminished in larger domains, particularly in regions distant from the surface. The introduction of physical constraints in the training processes improved the model’s robustness in noisy environments, enabling accurate reconstruction of hot-spots in deeper regions where traditional CNNs struggled. Conclusions: Combining deep learning with physical constraints provides a robust framework for non-invasive thermal imaging and other applications requiring high-precision temperature field reconstruction, particularly under non-ideal conditions. Full article

Background: Assessment of skeletal maturity is a common clinical practice to investigate adolescent growth and endocrine disorders. The distal radius and ulna (DRU) maturity classification is a practical and easy-to-use scheme that was designed for adolescent idiopathic scoliosis clinical management and presents high sensitivity in predicting the growth peak and cessation among adolescents. However, time-consuming and error-prone manual assessment limits DRU in clinical application. Methods: In this study, we propose a multi-task learning framework with an attention mechanism for the joint segmentation and classification of the distal radius and ulna in hand X-ray images. The proposed framework consists of two sub-networks: an encoder–decoder structure with attention gates for segmentation and a slight convolutional network for classification. Results: With a transfer learning strategy, the proposed framework improved DRU segmentation and classification over the single task learning counterparts and previously reported methods, achieving an accuracy of 94.3% and 90.8% for radius and ulna maturity grading. Findings: Our automatic DRU assessment platform covers the whole process of growth acceleration and cessation during puberty. Upon incorporation into advanced scoliosis progression prognostic tools, clinical decision making will be potentially improved in the conservative and operative management of scoliosis patients. Full article

Background: Early diagnosis of depression is crucial for effective treatment and suicide prevention. Traditional methods rely on self-report questionnaires and clinical assessments, lacking objective biomarkers. Combining functional magnetic resonance imaging (fMRI) with artificial intelligence can enhance depression diagnosis using neuroimaging indicators, but depression-specific fMRI datasets are often small and imbalanced, posing challenges for classification models. New Method: We propose the Spatio-Temporal Aggregation Network (STANet) for diagnosing depression by integrating convolutional neural networks (CNN) and recurrent neural networks (RNN) to capture both temporal and spatial features of brain activity. STANet comprises the following steps: (1) Aggregate spatio-temporal information via independent component analysis (ICA). (2) Utilize multi-scale deep convolution to capture detailed features. (3) Balance data using the synthetic minority over-sampling technique (SMOTE) to generate new samples for minority classes. (4) Employ the attention-Fourier gate recurrent unit (AFGRU) classifier to capture long-term dependencies, with an adaptive weight assignment mechanism to enhance model generalization. Results: STANet achieves superior depression diagnostic performance, with 82.38% accuracy and a 90.72% AUC. The Spatio-Temporal Feature Aggregation module enhances classification by capturing deeper features at multiple scales. The AFGRU classifier, with adaptive weights and a stacked Gated Recurrent Unit (GRU), attains higher accuracy and AUC. SMOTE outperforms other oversampling methods. Additionally, spatio-temporal aggregated features achieve better performance compared to using only temporal or spatial features. Comparison with existing methods: STANet significantly outperforms traditional classifiers, deep learning classifiers, and functional connectivity-based classifiers. Conclusions: The successful performance of STANet contributes to enhancing the diagnosis and treatment assessment of depression in clinical settings on imbalanced and small fMRI. Full article