Data Brief. 2024 Feb 6;53:110165. doi: 10.1016/j.dib.2024.110165. eCollection 2024 Apr.

ABSTRACT

Like other crops, different types of diseases affect apple trees. These diseases cause ugly cosmetic changes on the fruit and hence reduce its shelf life and value. To eliminate their impact, they need to be detected well in advance before any control measures are applied. The manual method of disease detection and subsequent classification has flaws as it involves manual scouting and analysis of the affected leaves through the naked eye. Besides, the manual method may result in wrong judgment as the knowledge of an expert limits the accuracy. Deep Learning Models have been successfully implemented for automated disease detection and classification. However, these models need massive datasets for training, testing and validation. This study proposes one such dataset that has been built indigenously by collecting images from the apple cultivation fields of Kashmir valley and subjecting it to cleaning and subsequent annotation by experts. Augmentation techniques have been used to enhance the size and quality of the dataset to prevent over-fitting of deep learning models.

PMID:38379888 | PMC:PMC10877676 | DOI:10.1016/j.dib.2024.110165

J Med Imaging (Bellingham). 2024 Jan;11(1):014008. doi: 10.1117/1.JMI.11.1.014008. Epub 2024 Feb 19.

ABSTRACT

PURPOSE: In recent years, the continuous advancement of convolutional neural networks (CNNs) has led to the widespread integration of deep neural networks as a mainstream approach in clinical diagnostic support. Particularly, the utilization of CNN-based medical image segmentation has delivered favorable outcomes for aiding clinical diagnosis. Within this realm, network architectures based on the U-shaped structure and incorporating skip connections, along with their diverse derivatives, have gained extensive utilization across various medical image segmentation tasks. Nonetheless, two primary challenges persist. First, certain organs or tissues present considerable complexity, substantial morphological variations, and size discrepancies, posing significant challenges for achieving highly accurate segmentation. Second, the predominant focus of current deep neural networks on single-resolution feature extraction limits the effective extraction of feature information from complex medical images, thereby contributing to information loss via continuous pooling operations and contextual information interaction constraints within the U-shaped structure.

APPROACH: We proposed a five-layer pyramid segmentation network (PS5-Net), a multiscale segmentation network with diverse resolutions that is founded on the U-Net architecture. Initially, this network effectively leverages the distinct features of images at varying resolutions across different dimensions, departing from prior single-resolution feature extraction methods to adapt to intricate and variable segmentation scenarios. Subsequently, to comprehensively integrate feature information from diverse resolutions, a kernel selection module is proposed to assign weights to features across different dimensions, enhancing the fusion of feature information from various resolutions. Within the feature extraction network denoted as PS-UNet, we preserve the classical structure of the traditional U-Net while enhancing it through the incorporation of dilated convolutions.

RESULTS: PS5-Net attains a Dice score of 0.9613 for liver segmentation on the CHLISC dataset and 0.8587 on the ISIC2018 dataset for skin lesion segmentation. Comparative analysis with diverse medical image segmentation methodologies in recent years reveals that PS5-Net has achieved the highest scores and substantial advancements.

CONCLUSIONS: PS5-Net effectively harnesses the rich semantic information available at different resolutions, facilitating a comprehensive and nuanced understanding of the input medical images. By capitalizing on global contextual connections, the network adeptly captures the intricate interplay of features and dependencies across the entire image, resulting in more accurate and robust segmentation outcomes. The experimental validation of PS5-Net underscores its superior performance in medical image segmentation tasks, offering promising prospects for enhancing diagnostic and analytical processes within clinical settings. These results highlight the potential of PS5-Net to significantly contribute to the advancement of medical imaging technologies and ultimately improve patient care through more precise and reliable image analysis.

PMID:38379775 | PMC:PMC10876014 | DOI:10.1117/1.JMI.11.1.014008

Clin Nutr. 2024 Feb 6;43(3):881-891. doi: 10.1016/j.clnu.2024.02.005. Online ahead of print.

ABSTRACT

OBJECTIVE: The aim of this study is using clinical factors and non-enhanced computed tomography (CT) deep features of the psoas muscles at third lumbar vertebral (L3) level to construct a model to predict malnutrition in gastric cancer before surgery, and to provide a new nutritional status assessment and survival assessment tool for gastric cancer patients.

METHODS: A retrospective analysis of 312 patients of gastric cancer were divided into malnutrition group and normal group based on Nutrition Risk Screening 2002(NRS-2002). 312 regions of interest (ROI) of the psoas muscles at L3 level of non-enhanced CT were delineated. Deep learning (DL) features were extracted from the ROI using a deep migration model and were screened by principal component analysis (PCA) and least-squares operator (LASSO). The clinical predictors included Body Mass Index (BMI), lymphocyte and albumin. Both deep learning model (including deep learning features) and mixed model (including selected deep learning features and selected clinical predictors) were constructed by 11 classifiers. The model was evaluated and selected by calculating receiver operating characteristic (ROC), area under curve (AUC), accuracy, sensitivity and specificity, calibration curve and decision curve analysis (DCA). The Cohen's Kappa coefficient (κ) was using to compare the diagnostic agreement for malnutrition between the mixed model and the GLIM in gastric cancer patients.

RESULT: The results of logistics multivariate analysis showed that BMI [OR = 0.569 (95% CI 0.491-0.660)], lymphocyte [OR = 0.638 (95% CI 0.408-0.998)], and albumin [OR = 0.924 (95% CI 0.859-0.994)] were clinically independent malnutrition of gastric cancer predictor(P < 0.05). Among the 11 classifiers, the Multilayer Perceptron (MLP)were selected as the best classifier. The AUC of the training and test sets for deep learning model were 0.806 (95% CI 0.7485-0.8635) and 0.769 (95% CI 0.673-0.863) and with accuracies were 0.734 and 0.766, respectively. The AUC of the training and test sets for the mixed model were 0.909 (95% CI 0.869-0.948) and 0.857 (95% CI 0.782-0.931) and with accuracies of 0.845 and 0.861, respectively. The DCA confirmed the clinical benefit of the both models. The Cohen's Kappa coefficient (κ) was 0.647 (P < 0.001). Diagnostic agreement for malnutrition between the mixed model and GLIM criteria was good. The mixed model was used to calculate the predicted probability of malnutrition in gastric cancer patients, which was divided into high-risk and low-risk groups by median, and the survival analysis showed that the overall survival time of the high-risk group was significantly lower than that of the low-risk group (P = 0.005).

CONCLUSION: Deep learning based on mixed model may be a potential tool for predicting malnutrition in gastric cancer patients.

PMID:38377634 | DOI:10.1016/j.clnu.2024.02.005

Comput Med Imaging Graph. 2024 Feb 16;113:102355. doi: 10.1016/j.compmedimag.2024.102355. Online ahead of print.

ABSTRACT

Automatic retinal arteriovenous classification can assist ophthalmologists in disease early diagnosis. Deep learning-based methods and topological graph-based methods have become the main solutions for retinal arteriovenous classification in recent years. This paper reviews the automatic retinal arteriovenous classification methods from 2003 to 2022. Firstly, we compare different methods and provide comparison tables of the summary results. Secondly, we complete the classification of the public arteriovenous classification datasets and provide the annotation development tables of different datasets. Finally, we sort out the challenges of evaluation methods and provide a comprehensive evaluation system. Quantitative and qualitative analysis shows the changes in research hotspots over time, Quantitative and qualitative analyses reveal the evolution of research hotspots over time, highlighting the significance of exploring the integration of deep learning with topological information in future research.

PMID:38377630 | DOI:10.1016/j.compmedimag.2024.102355

Anal Chem. 2024 Feb 20. doi: 10.1021/acs.analchem.3c05002. Online ahead of print.

ABSTRACT

Mass spectrometry imaging (MSI) is a high-throughput imaging technique capable of the qualitative and quantitative in situ detection of thousands of ions in biological samples. Ion image representation is a technique that produces a low-dimensional vector embedded with significant spectral and spatial information on an ion image, which further facilitates the distance-based similarity measurement for the identification of colocalized ions. However, given the low signal-to-noise ratios inherent in MSI data coupled with the scarcity of annotated data sets, achieving an effective ion image representation for each ion image remains a challenge. In this study, we propose DeepION, a novel deep learning-based method designed specifically for ion image representation, which is applied to the identification of colocalized ions and isotope ions. In DeepION, contrastive learning is introduced to ensure that the model can generate the ion image representation in a self-supervised manner without manual annotation. Since data augmentation is a crucial step in contrastive learning, a unique data augmentation strategy is designed by considering the characteristics of MSI data, such as the Poisson distribution of ion abundance and a random pattern of missing values, to generate plentiful ion image pairs for DeepION model training. Experimental results of rat brain tissue MSI show that DeepION outperforms other methods for both colocalized ion and isotope ion identification, demonstrating the effectiveness of ion image representation. The proposed model could serve as a crucial tool in the biomarker discovery and drug development of the MSI technique.

PMID:38377545 | DOI:10.1021/acs.analchem.3c05002

Proc Natl Acad Sci U S A. 2024 Feb 27;121(9):e2310012121. doi: 10.1073/pnas.2310012121. Epub 2024 Feb 20.

ABSTRACT

Sex plays a crucial role in human brain development, aging, and the manifestation of psychiatric and neurological disorders. However, our understanding of sex differences in human functional brain organization and their behavioral consequences has been hindered by inconsistent findings and a lack of replication. Here, we address these challenges using a spatiotemporal deep neural network (stDNN) model to uncover latent functional brain dynamics that distinguish male and female brains. Our stDNN model accurately differentiated male and female brains, demonstrating consistently high cross-validation accuracy (>90%), replicability, and generalizability across multisession data from the same individuals and three independent cohorts (N ~ 1,500 young adults aged 20 to 35). Explainable AI (XAI) analysis revealed that brain features associated with the default mode network, striatum, and limbic network consistently exhibited significant sex differences (effect sizes > 1.5) across sessions and independent cohorts. Furthermore, XAI-derived brain features accurately predicted sex-specific cognitive profiles, a finding that was also independently replicated. Our results demonstrate that sex differences in functional brain dynamics are not only highly replicable and generalizable but also behaviorally relevant, challenging the notion of a continuum in male-female brain organization. Our findings underscore the crucial role of sex as a biological determinant in human brain organization, have significant implications for developing personalized sex-specific biomarkers in psychiatric and neurological disorders, and provide innovative AI-based computational tools for future research.

PMID:38377194 | DOI:10.1073/pnas.2310012121

PLoS Comput Biol. 2024 Feb 20;20(2):e1011890. doi: 10.1371/journal.pcbi.1011890. Online ahead of print.

ABSTRACT

Recent advances in computer vision have led to significant progress in the generation of realistic image data, with denoising diffusion probabilistic models proving to be a particularly effective method. In this study, we demonstrate that diffusion models can effectively generate fully-annotated microscopy image data sets through an unsupervised and intuitive approach, using rough sketches of desired structures as the starting point. The proposed pipeline helps to reduce the reliance on manual annotations when training deep learning-based segmentation approaches and enables the segmentation of diverse datasets without the need for human annotations. We demonstrate that segmentation models trained with a small set of synthetic image data reach accuracy levels comparable to those of generalist models trained with a large and diverse collection of manually annotated image data, thereby offering a streamlined and specialized application of segmentation models.

PMID:38377165 | DOI:10.1371/journal.pcbi.1011890

PLoS Genet. 2024 Feb 20;20(2):e1010657. doi: 10.1371/journal.pgen.1010657. Online ahead of print.

ABSTRACT

A growing body of evidence suggests that gene flow between closely related species is a widespread phenomenon. Alleles that introgress from one species into a close relative are typically neutral or deleterious, but sometimes confer a significant fitness advantage. Given the potential relevance to speciation and adaptation, numerous methods have therefore been devised to identify regions of the genome that have experienced introgression. Recently, supervised machine learning approaches have been shown to be highly effective for detecting introgression. One especially promising approach is to treat population genetic inference as an image classification problem, and feed an image representation of a population genetic alignment as input to a deep neural network that distinguishes among evolutionary models (i.e. introgression or no introgression). However, if we wish to investigate the full extent and fitness effects of introgression, merely identifying genomic regions in a population genetic alignment that harbor introgressed loci is insufficient-ideally we would be able to infer precisely which individuals have introgressed material and at which positions in the genome. Here we adapt a deep learning algorithm for semantic segmentation, the task of correctly identifying the type of object to which each individual pixel in an image belongs, to the task of identifying introgressed alleles. Our trained neural network is thus able to infer, for each individual in a two-population alignment, which of those individual's alleles were introgressed from the other population. We use simulated data to show that this approach is highly accurate, and that it can be readily extended to identify alleles that are introgressed from an unsampled "ghost" population, performing comparably to a supervised learning method tailored specifically to that task. Finally, we apply this method to data from Drosophila, showing that it is able to accurately recover introgressed haplotypes from real data. This analysis reveals that introgressed alleles are typically confined to lower frequencies within genic regions, suggestive of purifying selection, but are found at much higher frequencies in a region previously shown to be affected by adaptive introgression. Our method's success in recovering introgressed haplotypes in challenging real-world scenarios underscores the utility of deep learning approaches for making richer evolutionary inferences from genomic data.

PMID:38377104 | DOI:10.1371/journal.pgen.1010657

IEEE Trans Biomed Eng. 2024 Feb 20;PP. doi: 10.1109/TBME.2024.3367530. Online ahead of print.

ABSTRACT

Cochlear implants (CIs) provide a solution for individuals with severe sensorineural hearing loss to regain their hearing abilities. When someone experiences this form of hearing impairment in both ears, they may be equipped with two separate CI devices, which will typically further improve the CI benefits. This spatial hearing is particularly crucial when tackling the challenge of understanding speech in noisy environments, a common issue CI users face. Currently, extensive research is dedicated to developing algorithms that can autonomously filter out undesired background noises from desired speech signals. At present, some research focuses on achieving end-to-end denoising, either as an integral component of the initial CI signal processing or by fully integrating the denoising process into the CI sound coding strategy. This work is presented in the context of bilateral CI (BiCI) systems, where we propose a deep-learning-based bilateral speech enhancement model that shares information between both hearing sides. Specifically, we connect two monaural end-to-end deep denoising sound coding techniques through intermediary latent fusion layers. These layers amalgamate the latent representations generated by these techniques by multiplying them together, resulting in an enhanced ability to reduce noise and improve learning generalization. The objective instrumental results demonstrate that the proposed fused BiCI sound coding strategy achieves higher interaural coherence, superior noise reduction, and enhanced predicted speech intelligibility scores compared to the baseline methods. Furthermore, our speech-in-noise intelligibility results in BiCI users reveal that the deep denoising sound coding strategy can attain scores similar to those achieved in quiet conditions.

PMID:38376983 | DOI:10.1109/TBME.2024.3367530

IEEE J Biomed Health Inform. 2024 Feb 20;PP. doi: 10.1109/JBHI.2024.3367756. Online ahead of print.

ABSTRACT

Deep learning-based methods have been widely used in medical image segmentation recently. However, existing works are usually difficult to simultaneously capture global long-range information from images and topological correlations among feature maps. Further, medical images often suffer from blurred target edges. Accordingly, this paper proposes a novel medical image segmentation framework named a label-decoupled network with spatial-channel graph convolution and dual attention enhancement mechanism (LADENet for short). It constructs learnable adjacency matrices and utilizes graph convolutions to effectively capture global long-range information on spatial locations and topological dependencies between different channels in an image. Then a label-decoupled strategy based on distance transformation is introduced to decouple an original segmentation label into a body label and an edge label for supervising the body branch and edge branch. Again, a dual attention enhancement mechanism, designing a body attention block in the body branch and an edge attention block in the edge branch, is built to promote the learning ability of spatial region and boundary features. Besides, a feature interactor is devised to fully consider the information interaction between the body and edge branches to improve segmentation performance. Experiments on benchmark datasets reveal the superiority of LADENet compared to state-of-the-art approaches.

PMID:38376972 | DOI:10.1109/JBHI.2024.3367756

IEEE Trans Pattern Anal Mach Intell. 2024 Feb 20;PP. doi: 10.1109/TPAMI.2024.3367773. Online ahead of print.

ABSTRACT

In recent years, the security of deep learning models achieves more and more attentions with the rapid development of neural networks, which are vulnerable to adversarial examples. Almost all existing gradient-based attack methods use the sign function in the generation to meet the requirement of perturbation budget on L∞ norm. However, we find that the sign function may be improper for generating adversarial examples since it modifies the exact gradient direction. Instead of using the sign function, we propose to directly utilize the exact gradient direction with a scaling factor for generating adversarial perturbations, which improves the attack success rates of adversarial examples even with fewer perturbations. At the same time, we also theoretically prove that this method can achieve better black-box transferability. Moreover, considering that the best scaling factor varies across different images, we propose an adaptive scaling factor generator to seek an appropriate scaling factor for each image, which avoids the computational cost for manually searching the scaling factor. Our method can be integrated with almost all existing gradient-based attack methods to further improve their attack success rates. Extensive experiments on the CIFAR10 and ImageNet datasets show that our method exhibits higher transferability and outperforms the state-of-the-art methods.

PMID:38376968 | DOI:10.1109/TPAMI.2024.3367773

Med Biol Eng Comput. 2024 Feb 20. doi: 10.1007/s11517-024-03047-6. Online ahead of print.

ABSTRACT

Challenges arise in accessing archived signal outputs due to proprietary software limitations. There is a notable lack of exploration in open-source mandibular EMG signal conversion for continuous access and analysis, hindering tasks such as pattern recognition and predictive modelling for temporomandibular joint complex function. To Develop a workflow to extract normalised signal parameters from images of mandibular muscle EMG and identify optimal clustering methods for quantifying signal intensity and activity durations. A workflow utilising OpenCV, variational encoders and Neurokit2 generated and augmented 866 unique EMG signals from jaw movement exercises. k-means, GMM and DBSCAN were employed for normalisation and cluster-centric signal processing. The workflow was validated with data collected from 66 participants, measuring temporalis, masseter and digastric muscles. DBSCAN (0.35 to 0.54) and GMM (0.09 to 0.24) exhibited lower silhouette scores for mouth opening, anterior protrusion and lateral excursions, while K-means performed best (0.10 to 0.11) for temporalis and masseter muscles during chewing activities. The current study successfully developed a deep learning workflow capable of extracting normalised signal data from EMG images and generating quantifiable parameters for muscle activity duration and general functional intensity.

PMID:38376739 | DOI:10.1007/s11517-024-03047-6

Hepatol Int. 2024 Feb 20. doi: 10.1007/s12072-023-10630-w. Online ahead of print.

ABSTRACT

Liver disease is regarded as one of the major health threats to humans. Radiographic assessments hold promise in terms of addressing the current demands for precisely diagnosing and treating liver diseases, and artificial intelligence (AI), which excels at automatically making quantitative assessments of complex medical image characteristics, has made great strides regarding the qualitative interpretation of medical imaging by clinicians. Here, we review the current state of medical-imaging-based AI methodologies and their applications concerning the management of liver diseases. We summarize the representative AI methodologies in liver imaging with focusing on deep learning, and illustrate their promising clinical applications across the spectrum of precise liver disease detection, diagnosis and treatment. We also address the current challenges and future perspectives of AI in liver imaging, with an emphasis on feature interpretability, multimodal data integration and multicenter study. Taken together, it is revealed that AI methodologies, together with the large volume of available medical image data, might impact the future of liver disease care.

PMID:38376649 | DOI:10.1007/s12072-023-10630-w

Ann Nucl Med. 2024 Feb 20. doi: 10.1007/s12149-024-01907-7. Online ahead of print.

ABSTRACT

OBJECTIVE: Accurate delineation of renal regions of interest (ROIs) is critical for the assessment of renal function in pediatric dynamic renal scintigraphy (DRS). The purpose of this study was to develop and evaluate a deep learning (DL) model that can fully automatically delineate renal ROIs and calculate renal function in pediatric 99mTechnetium-ethylenedicysteine (99mTc-EC) DRS.

METHODS: This study retrospectively analyzed 1,283 pediatric DRS data at a single center from January to December 2018. These patients were divided into training set (n = 1027), validation set (n = 128), and testing set (n = 128). A fully automatic segmentation of ROIs (FASR) model was developed and evaluated. The pixel values of the automatically segmented ROIs were calculated to predict renal blood perfusion rate (BPR) and differential renal function (DRF). Precision, recall rate, intersection over union (IOU), and Dice similarity coefficient (DSC) were used to evaluate the performance of FASR model. Intraclass correlation (ICC) and Pearson correlation analysis were used to compare the consistency of automatic and manual method in assessing the renal function parameters in the testing set.

RESULTS: The FASR model achieved a precision of 0.88, recall rate of 0.94, IOU of 0.83, and DSC of 0.91. In the testing set, the r values of BPR and DRF calculated by the two methods were 0.94 (P < 0.01) and 0.97 (P < 0.01), and the ICCs (95% confidence interval CI) were 0.94 (0.90-0.96) and 0.94 (0.91-0.96).

CONCLUSION: We propose a reliable and stable DL model that can fully automatically segment ROIs and accurately predict renal function in pediatric 99mTc-EC DRS.

PMID:38376629 | DOI:10.1007/s12149-024-01907-7

Acta Neurochir (Wien). 2024 Feb 20;166(1):92. doi: 10.1007/s00701-024-05973-8.

ABSTRACT

PURPOSE: This study evaluates the nnU-Net for segmenting brain, skin, tumors, and ventricles in contrast-enhanced T1 (T1CE) images, benchmarking it against an established mesh growing algorithm (MGA).

METHODS: We used 67 retrospectively collected annotated single-center T1CE brain scans for training models for brain, skin, tumor, and ventricle segmentation. An additional 32 scans from two centers were used test performance compared to that of the MGA. The performance was measured using the Dice-Sørensen coefficient (DSC), intersection over union (IoU), 95th percentile Hausdorff distance (HD95), and average symmetric surface distance (ASSD) metrics, with time to segment also compared.

RESULTS: The nnU-Net models significantly outperformed the MGA (p < 0.0125) with a median brain segmentation DSC of 0.971 [95CI: 0.945-0.979], skin: 0.997 [95CI: 0.984-0.999], tumor: 0.926 [95CI: 0.508-0.968], and ventricles: 0.910 [95CI: 0.812-0.968]. Compared to the MGA's median DSC for brain: 0.936 [95CI: 0.890, 0.958], skin: 0.991 [95CI: 0.964, 0.996], tumor: 0.723 [95CI: 0.000-0.926], and ventricles: 0.856 [95CI: 0.216-0.916]. NnU-Net performance between centers did not significantly differ except for the skin segmentations Additionally, the nnU-Net models were faster (mean: 1139 s [95CI: 685.0-1616]) than the MGA (mean: 2851 s [95CI: 1482-6246]).

CONCLUSIONS: The nnU-Net is a fast, reliable tool for creating automatic deep learning-based segmentation pipelines, reducing the need for extensive manual tuning and iteration. The models are able to achieve this performance despite a modestly sized training set. The ability to create high-quality segmentations in a short timespan can prove invaluable in neurosurgical settings.

PMID:38376564 | DOI:10.1007/s00701-024-05973-8

Eur Heart J Cardiovasc Imaging. 2024 Feb 20:jeae045. doi: 10.1093/ehjci/jeae045. Online ahead of print.

ABSTRACT

AIMS: Vessel specific coronary artery calcification (CAC) is additive to global CAC for prognostic assessment. We assessed accuracy and prognostic implications of vessel-specific automated deep learning (DL) CAC analysis on electrocardiogram gated and attenuation correction computed tomography (CT) in a large multicenter registry.

METHODS AND RESULTS: Vessel-specific CAC was assessed in the left main/left anterior descending (LM/LAD), left circumflex (LCX) and right coronary artery (RCA) using a DL model trained on 3000 gated CT and tested on 2094 gated CT and 5969 non-gated attenuation correction CT. Vessel-specific agreement was assessed with linear weighted Cohen's Kappa for CAC zero, 1-100, 101-400 and >400 Agatston units (AU). Risk of major adverse cardiovascular events (MACE) was assessed during 2.4±1.4 years follow-up, with hazard ratios (HR) and 95% confidence intervals (CI). There was strong to excellent agreement between DL and expert ground truth for CAC in LM/LAD, LCX and RCA on gated CT [0.90 (95% CI 0.89 to 0.92); 0.70 (0.68 to 0.73); 0.79 (0.77 to 0.81)] and attenuation correction CT [(0.78 (0.77 to 0.80); 0.60 (0.58 to 0.62); 0.70 (0.68 to 0.71)]. MACE occurred in 242 (12%) undergoing gated CT and 841(14%) of undergoing attenuation correction CT. LM/LAD CAC >400 AU was associated with the highest risk of MACE on gated (HR 12.0, 95% CI 7.96, 18.0, p<0.001) and attenuation correction CT (HR 4.21, 95% CI 3.48, 5.08, p<0.001).

CONCLUSION: Vessel-specific CAC assessment with DL can be performed accurately and rapidly on gated CT and attenuation correction CT and provides important prognostic information.

PMID:38376471 | DOI:10.1093/ehjci/jeae045

Magn Reson Med. 2024 Feb 20. doi: 10.1002/mrm.30046. Online ahead of print.

ABSTRACT

PURPOSE: To develop a new electromagnetic interference (EMI) elimination strategy for RF shielding-free MRI via active EMI sensing and deep learning direct MR signal prediction (Deep-DSP).

METHODS: Deep-DSP is proposed to directly predict EMI-free MR signals. During scanning, MRI receive coil and EMI sensing coils simultaneously sample data within two windows (i.e., for MR data and EMI characterization data acquisition, respectively). Afterward, a residual U-Net model is trained using synthetic MRI receive coil data and EMI sensing coil data acquired during EMI signal characterization window, to predict EMI-free MR signals from signals acquired by MRI receive and EMI sensing coils. The trained model is then used to directly predict EMI-free MR signals from data acquired by MRI receive and sensing coils during the MR signal-acquisition window. This strategy was evaluated on an ultralow-field 0.055T brain MRI scanner without any RF shielding and a 1.5T whole-body scanner with incomplete RF shielding.

RESULTS: Deep-DSP accurately predicted EMI-free MR signals in presence of strong EMI. It outperformed recently developed EDITER and convolutional neural network methods, yielding better EMI elimination and enabling use of few EMI sensing coils. Furthermore, it could work well without dedicated EMI characterization data.

CONCLUSION: Deep-DSP presents an effective EMI elimination strategy that outperforms existing methods, advancing toward truly portable and patient-friendly MRI. It exploits electromagnetic coupling between MRI receive and EMI sensing coils as well as typical MR signal characteristics. Despite its deep learning nature, Deep-DSP framework is computationally simple and efficient.

PMID:38376455 | DOI:10.1002/mrm.30046

Med Phys. 2024 Feb 20. doi: 10.1002/mp.16997. Online ahead of print.

ABSTRACT

BACKGROUND: Auto-segmentation of organs-at-risk (OARs) in the head and neck (HN) on computed tomography (CT) images is a time-consuming component of the radiation therapy pipeline that suffers from inter-observer variability. Deep learning (DL) has shown state-of-the-art results in CT auto-segmentation, with larger and more diverse datasets showing better segmentation performance. Institutional CT auto-segmentation datasets have been small historically (n < 50) due to the time required for manual curation of images and anatomical labels. Recently, large public CT auto-segmentation datasets (n > 1000 aggregated) have become available through online repositories such as The Cancer Imaging Archive. Transfer learning is a technique applied when training samples are scarce, but a large dataset from a closely related domain is available.

PURPOSE: The purpose of this study was to investigate whether a large public dataset could be used in place of an institutional dataset (n > 500), or to augment performance via transfer learning, when building HN OAR auto-segmentation models for institutional use.

METHODS: Auto-segmentation models were trained on a large public dataset (public models) and a smaller institutional dataset (institutional models). The public models were fine-tuned on the institutional dataset using transfer learning (transfer models). We assessed both public model generalizability and transfer model performance by comparison with institutional models. Additionally, the effect of institutional dataset size on both transfer and institutional models was investigated. All DL models used a high-resolution, two-stage architecture based on the popular 3D U-Net. Model performance was evaluated using five geometric measures: the dice similarity coefficient (DSC), surface DSC, 95th percentile Hausdorff distance, mean surface distance (MSD), and added path length.

RESULTS: For a small subset of OARs (left/right optic nerve, spinal cord, left submandibular), the public models performed significantly better (p < 0.05) than, or showed no significant difference to, the institutional models under most of the metrics examined. For the remaining OARs, the public models were inferior to the institutional models, although performance differences were small (DSC ≤ 0.03, MSD < 0.5 mm) for seven OARs (brainstem, left/right lens, left/right parotid, mandible, right submandibular). The transfer models performed significantly better than the institutional models for seven OARs (brainstem, right lens, left/right optic nerve, left/right parotid, spinal cord) with a small margin of improvement (DSC ≤ 0.02, MSD < 0.4 mm). When numbers of institutional training samples were limited, public and transfer models outperformed the institutional models for most OARs (brainstem, left/right lens, left/right optic nerve, left/right parotid, spinal cord, and left/right submandibular).

CONCLUSION: Training auto-segmentation models with public data alone was suitable for a small number of OARs. Using only public data incurred a small performance deficit for most other OARs, when compared with institutional data alone, but may be preferable over time-consuming curation of a large institutional dataset. When a large institutional dataset was available, transfer learning with models pretrained on a large public dataset provided a modest performance improvement for several OARs. When numbers of institutional samples were limited, using the public dataset alone, or as a pretrained model, was beneficial for most OARs.

PMID:38376454 | DOI:10.1002/mp.16997

Radiology. 2024 Feb;310(2):e231938. doi: 10.1148/radiol.231938.

ABSTRACT

Background Deep learning (DL)-accelerated MRI can substantially reduce examination times. However, studies prospectively evaluating the diagnostic performance of DL-accelerated MRI reconstructions in acute suspected stroke are lacking. Purpose To investigate the interchangeability of DL-accelerated MRI with conventional MRI in patients with suspected acute ischemic stroke at 1.5 T. Materials and Methods In this prospective study, 211 participants with suspected acute stroke underwent clinically indicated MRI at 1.5 T between June 2022 and March 2023. For each participant, conventional MRI (including T1-weighted, T2-weighted, T2*-weighted, T2 fluid-attenuated inversion-recovery, and diffusion-weighted imaging; 14 minutes 18 seconds) and DL-accelerated MRI (same sequences; 3 minutes 4 seconds) were performed. The primary end point was the interchangeability between conventional and DL-accelerated MRI for acute ischemic infarction detection. Secondary end points were interchangeability regarding the affected vascular territory and clinically relevant secondary findings (eg, microbleeds, neoplasm). Three readers evaluated the overall occurrence of acute ischemic stroke, affected vascular territory, clinically relevant secondary findings, overall image quality, and diagnostic confidence. For acute ischemic lesions, size and signal intensities were assessed. The margin for interchangeability was chosen as 5%. For interrater agreement analysis and interrater reliability analysis, multirater Fleiss κ and the intraclass correlation coefficient, respectively, was determined. Results The study sample consisted of 211 participants (mean age, 65 years ± 16 [SD]); 123 male and 88 female). Acute ischemic stroke was confirmed in 79 participants. Interchangeability was demonstrated for all primary and secondary end points. No individual equivalence indexes (IEIs) exceeded the interchangeability margin of 5% (IEI, -0.002 [90% CI: -0.007, 0.004]). Almost perfect interrater agreement was observed (P > .91). DL-accelerated MRI provided higher overall image quality (P < .001) and diagnostic confidence (P < .001). The signal properties of acute ischemic infarctions were similar in both techniques and demonstrated good to excellent interrater reliability (intraclass correlation coefficient, ≥0.8). Conclusion Despite being four times faster, DL-accelerated brain MRI was interchangeable with conventional MRI for acute ischemic lesion detection. © RSNA, 2024 Supplemental material is available for this article. See also the editorial by Haller in this issue.

PMID:38376403 | DOI:10.1148/radiol.231938

Brain Behav. 2024 Jan;14(1):e3348. doi: 10.1002/brb3.3348.

ABSTRACT

BACKGROUND: Predicting suicide is a pressing issue among older adults; however, predicting its risk is difficult. Capitalizing on the recent development of machine learning, considerable progress has been made in predicting complex behavior such as suicide. As depression remained the strongest risk for suicide, we aimed to apply deep learning algorithms to identify suicidality in a group with late-life depression (LLD).

METHODS: We enrolled 83 patients with LLD, 35 of which were non-suicidal and 48 were suicidal, including 26 with only suicidal ideation and 22 with past suicide attempts, for resting-state functional magnetic resonance imaging (MRI). Cross-sample entropy (CSE) analysis was conducted to examine the complexity of MRI signals among brain regions. Three-dimensional (3D) convolutional neural networks (CNNs) were used, and the classification accuracy in each brain region was averaged to predict suicidality after sixfold cross-validation.

RESULTS: We found brain regions with a mean accuracy above 75% to predict suicidality located mostly in default mode, fronto-parietal, and cingulo-opercular resting-state networks. The models with right amygdala and left caudate provided the most reliable accuracy in all cross-validation folds, indicating their neurobiological importance in late-life suicide.

CONCLUSION: Combining CSE analysis and the 3D CNN, several brain regions were found to be associated with suicidality.

PMID:38376042 | DOI:10.1002/brb3.3348