Comput Biol Med. 2024 Jan 24;170:108018. doi: 10.1016/j.compbiomed.2024.108018. Online ahead of print.

ABSTRACT

In histopathology practice, scanners, tissue processing, staining, and image acquisition protocols vary from center to center, resulting in subtle variations in images. Vanilla convolutional neural networks are sensitive to such domain shifts. Data augmentation is a popular way to improve domain generalization. Currently, state-of-the-art domain generalization in computational pathology is achieved using a manually curated set of augmentation transforms. However, manual tuning of augmentation parameters is time-consuming and can lead to sub-optimal generalization performance. Meta-learning frameworks can provide efficient ways to find optimal training hyper-parameters, including data augmentation. In this study, we hypothesize that an automated search of augmentation hyper-parameters can provide superior generalization performance and reduce experimental optimization time. We select four state-of-the-art automatic augmentation methods from general computer vision and investigate their capacity to improve domain generalization in histopathology. We analyze their performance on data from 25 centers across two different tasks: tumor metastasis detection in lymph nodes and breast cancer tissue type classification. On tumor metastasis detection, most automatic augmentation methods achieve comparable performance to state-of-the-art manual augmentation. On breast cancer tissue type classification, the leading automatic augmentation method significantly outperforms state-of-the-art manual data augmentation.

PMID:38281317 | DOI:10.1016/j.compbiomed.2024.108018

Comput Biol Med. 2024 Jan 21;170:108001. doi: 10.1016/j.compbiomed.2024.108001. Online ahead of print.

ABSTRACT

Intracranial arterial stenosis (ICAS) is characterized by the pathological narrowing or occlusion of the inner lumen of intracranial blood vessels. However, the retina can indirectly react to cerebrovascular disease. Therefore, retinal fundus images (RFI) serve as valuable noninvasive and easily accessible screening tools for early detection and diagnosis of ICAS. This paper introduces an adversarial learning-based domain adaptation algorithm (ALDA) specifically designed for ICAS detection in multi-source datasets. The primary objective is to achieve accurate detection and enhanced generalization of ICAS based on RFI. Given the limitations of traditional algorithms in meeting the accuracy and generalization requirements, ALDA overcomes these challenges by leveraging RFI datasets from multiple sources and employing the concept of adversarial learning to facilitate feature representation sharing and distinguishability learning. In order to evaluate the performance of the ALDA algorithm, we conducted experimental validation on multi-source datasets. We compared its results with those obtained from other deep learning algorithms in the ICAS detection task. Furthermore, we validated the potential of ALDA for detecting diabetic retinopathy. The experimental results clearly demonstrate the significant improvements achieved by the ALDA algorithm. By leveraging information from diverse datasets, ALDA learns feature representations that exhibit enhanced generalizability and robustness. This makes it a reliable auxiliary diagnostic tool for clinicians, thereby facilitating the prevention and treatment of cerebrovascular diseases.

PMID:38280254 | DOI:10.1016/j.compbiomed.2024.108001

Jpn J Radiol. 2024 Jan 27. doi: 10.1007/s11604-023-01527-7. Online ahead of print.

ABSTRACT

PURPOSE: To develop a convolutional neural network (CNN) model to diagnose skull-base invasion by nasopharyngeal malignancies in CT images and evaluate the model's diagnostic performance.

MATERIALS AND METHODS: We divided 100 malignant nasopharyngeal tumor lesions into a training (n = 70) and a test (n = 30) dataset. Two head/neck radiologists reviewed CT and MRI images and determined the positive/negative skull-base invasion status of each case (training dataset: 29 invasion-positive and 41 invasion-negative; test dataset: 13 invasion-positive and 17 invasion-negative). Preprocessing involved extracting continuous slices of the nasopharynx and clivus. The preprocessed training dataset was used for transfer learning with Residual Neural Networks 50 to create a diagnostic CNN model, which was then tested on the preprocessed test dataset to determine the invasion status and model performance. Original CT images from the test dataset were reviewed by a radiologist with extensive head/neck imaging experience (senior reader: SR) and another less-experienced radiologist (junior reader: JR). Gradient-weighted class activation maps (Grad-CAMs) were created to visualize the explainability of the invasion status classification.

RESULTS: The CNN model's diagnostic accuracy was 0.973, significantly higher than those of the two radiologists (SR: 0.838; JR: 0.595). Receiver operating characteristic curve analysis gave an area under the curve of 0.953 for the CNN model (versus 0.832 and 0.617 for SR and JR; both p < 0.05). The Grad-CAMs suggested that the invasion-negative cases were present predominantly in bone marrow, while the invasion-positive cases exhibited osteosclerosis and nasopharyngeal masses.

CONCLUSIONS: This CNN technique would be useful for CT-based diagnosis of skull-base invasion by nasopharyngeal malignancies.

PMID:38280100 | DOI:10.1007/s11604-023-01527-7

Acta Otolaryngol. 2024 Jan 27:1-7. doi: 10.1080/00016489.2023.2301648. Online ahead of print.

ABSTRACT

Background: The mortality rates of laryngeal squamous cell carcinoma cancer (LSCC) have not significantly decreased in the last decades.Objectives: We primarily aimed to compare the predictive performance of DeepTables with the state-of-the-art machine learning (ML) algorithms (Voting ensemble, Stack ensemble, and XGBoost) to stratify patients with LSCC into chance of overall survival (OS). In addition, we complemented the developed model by providing interpretability using both global and local model-agnostic techniques.Methods: A total of 2792 patients in the Surveillance, Epidemiology, and End Results (SEER) database diagnosed with LSCC were reviewed. The global model-agnostic interpretability was examined using SHapley Additive exPlanations (SHAP) technique. Likewise, individual interpretation of the prediction was made using Local Interpretable Model Agnostic Explanations (LIME).Results: The state-of-the-art ML ensemble algorithms outperformed DeepTables. Specifically, the examined ensemble algorithms showed comparable weighted area under receiving curve of 76.9, 76.8, and 76.1 with an accuracy of 71.2%, 70.2%, and 71.8%, respectively. The global methods of interpretability (SHAP) demonstrated that the age of the patient at diagnosis, N-stage, T-stage, tumor grade, and marital status are among the prominent parameters.Conclusions: A ML model for OS prediction may serve as an ancillary tool for treatment planning of LSCC patients.

PMID:38279817 | DOI:10.1080/00016489.2023.2301648

Network. 2024 Jan 27:1-28. doi: 10.1080/0954898X.2024.2306988. Online ahead of print.

ABSTRACT

Cardiovascular diseases (CVD) represent a significant global health challenge, often remaining undetected until severe cardiac events, such as heart attacks or strokes, occur. In regions like Qatar, research focused on non-invasive CVD identification methods, such as retinal imaging and dual-energy X-ray absorptiometry (DXA), is limited. This study presents a groundbreaking system known as Multi-Modal Artificial Intelligence for Cardiovascular Disease (M2AI-CVD), designed to provide highly accurate predictions of CVD. The M2AI-CVD framework employs a four-fold methodology: First, it rigorously evaluates image quality and processes lower-quality images for further analysis. Subsequently, it uses the Entropy-based Fuzzy C Means (EnFCM) algorithm for precise image segmentation. The Multi-Modal Boltzmann Machine (MMBM) is then employed to extract relevant features from various data modalities, while the Genetic Algorithm (GA) selects the most informative features. Finally, a ZFNet Convolutional Neural Network (ZFNetCNN) classifies images, effectively distinguishing between CVD and Non-CVD cases. The research's culmination, tested across five distinct datasets, yields outstanding results, with an accuracy of 95.89%, sensitivity of 96.89%, and specificity of 98.7%. This multi-modal AI approach offers a promising solution for the accurate and early detection of cardiovascular diseases, significantly improving the prospects of timely intervention and improved patient outcomes in the realm of cardiovascular health.

PMID:38279811 | DOI:10.1080/0954898X.2024.2306988

NPJ Precis Oncol. 2024 Jan 27;8(1):21. doi: 10.1038/s41698-024-00514-z.

ABSTRACT

Deep learning (DL) has been widely investigated in breast ultrasound (US) for distinguishing between benign and malignant breast masses. This systematic review of test diagnosis aims to examine the accuracy of DL, compared to human readers, for the diagnosis of breast cancer in the US under clinical settings. Our literature search included records from databases including PubMed, Embase, Scopus, and Cochrane Library. Test accuracy outcomes were synthesized to compare the diagnostic performance of DL and human readers as well as to evaluate the assistive role of DL to human readers. A total of 16 studies involving 9238 female participants were included. There were no prospective studies comparing the test accuracy of DL versus human readers in clinical workflows. Diagnostic test results varied across the included studies. In 14 studies employing standalone DL systems, DL showed significantly lower sensitivities in 5 studies with comparable specificities and outperformed human readers at higher specificities in another 4 studies; in the remaining studies, DL models and human readers showed equivalent test outcomes. In 12 studies that assessed assistive DL systems, no studies proved the assistive role of DL in the overall diagnostic performance of human readers. Current evidence is insufficient to conclude that DL outperforms human readers or enhances the accuracy of diagnostic breast US in a clinical setting. Standardization of study methodologies is required to improve the reproducibility and generalizability of DL research, which will aid in clinical translation and application.

PMID:38280946 | DOI:10.1038/s41698-024-00514-z

Neuroimage. 2024 Jan 25:120519. doi: 10.1016/j.neuroimage.2024.120519. Online ahead of print.

ABSTRACT

Functional brain networks (FBNs) are spatial patterns of brain function that play a critical role in understanding human brain function. There are many proposed methods for mapping the spatial patterns of brain function, however they oversimplify the underlying assumptions of brain function and have various limitations such as linearity and independence. Additionally, current methods fail to account for the dynamic nature of FBNs, which limits their effectiveness in accurately characterizing these networks. To address these limitations, we present a novel deep learning and spatial-wise attention based model called Spatial-Temporal Convolutional Attention (STCA) to accurately model dynamic FBNs. Specifically, we train STCA in a self-supervised manner by utilizing a Convolutional Autoencoder to guide the STCA module in assigning higher attention weights to regions of functional activity. To validate the reliability of the results, we evaluate our approach on the HCP- task motor behavior dataset, the experimental results demonstrate that the STCA derived FBNs have higher spatial similarity with the templates and that the spatial similarity between the templates and the FBNs derived by STCA fluctuates with the task design over time, suggesting that STCA can reflect the dynamic changes of brain function, providing a powerful tool to better understand human brain function. Code is available at https://github.com/SNNUBIAI/STCAE.

PMID:38280690 | DOI:10.1016/j.neuroimage.2024.120519

Lab Invest. 2024 Jan 25:100341. doi: 10.1016/j.labinv.2024.100341. Online ahead of print.

ABSTRACT

Ki-67 is a nuclear protein associated with proliferation, and a strong potential biomarker in breast cancer, but is not routinely measured in current clinical management due to a lack of standardization. Digital image analysis (DIA) is a promising technology that could allow high throughput analysis and standardization. There is a dearth of data on the clinical reliability as well as intra- and inter-algorithmic variability of different DIA methods. In this study, we scored and compared a set of breast cancer cases in which manually counted Ki-67 has already been demonstrated to have prognostic value (n=278) to five DIA methods; namely, Aperio ePathology, Definiens Tissue Studio, Qupath, an unsupervised IHC color histogram (IHCCH) algorithm and a deep learning pipeline piNET. The piNET system achieved high agreement (ICC: 0.850) and correlation (R= 0.85) with the reference score. The Qupath algorithm exhibited a high degree of reproducibility between all rater instances (ICC: 0.889). Although piNET performed well against absolute manual counts, none of the tested DIA methods classified common Ki-67 cutoffs with high agreement or reached the clinically relevant Cohen's kappa of at least 0.8. The highest agreement achieved was Cohen's kappa statistic of 0.73 for cutoffs 20% and 25% by the piNET system. The main contributors to inter-algorithmic variation and poor cutoff characterization included heterogeneous tumor biology, varying algorithm implementation, and setting assignments. It appears that image segmentation is the primary explanation for semi-automated intra-algorithmic variation, which involves significant manual intervention to correct. Automated pipelines such as piNET may be crucial in developing robust and reproducible unbiased DIA approaches to accurately quantify Ki-67 for clinical diagnosis in the future.

PMID:38280634 | DOI:10.1016/j.labinv.2024.100341

Heart Rhythm. 2024 Jan 25:S1547-5271(24)00085-7. doi: 10.1016/j.hrthm.2024.01.031. Online ahead of print.

ABSTRACT

BACKGROUND: Patients with hypertrophic cardiomyopathy (HCM) are at risk for sudden death and individuals with ≥1 major risk markers are considered for primary prevention implantable cardioverter defibrillators. Guidelines recommend cardiac magnetic resonance imaging (CMR) to identify high-risk imaging features. However, CMR is resource intensive and is not widely accessible worldwide.

OBJECTIVE: To develop electrocardiogram (ECG) deep-learning (DL) models for identification of HCM patients with high-risk features.

METHODS: HCM patients evaluated at Tufts Medical Center (N=1,930; Boston, United States) were used to develop ECG-DL models for prediction of high-risk imaging features: systolic dysfunction, massive hypertrophy (≥30mm), apical aneurysm, and extensive late-gadolinium enhancement (LGE). ECG-DL models were externally validated in an HCM cohort from Amrita Hospital (N=233; Kochi, India).

RESULTS: ECG-DL models reliably identified high-risk features (systolic dysfunction, massive hypertrophy, apical aneurysm, and extensive LGE) during hold-out model testing (c-statistics 0.72, 0.83, 0.93, and 0.76) and external validation (c-statistics 0.71, 0.76, 0.91, and 0.68). A hypothetical screening strategy employing echocardiography combined with ECG-DL guided selective CMR use demonstrated sensitivity of 97% for identifying patients with high-risk features, while reducing the number of recommended CMRs by 61%. Negative predictive value with this screening strategy for absence of high-risk features in patients without ECG-DL recommendation for CMR was 99.5%.

CONCLUSIONS: In HCM, novel ECG-DL models reliably identified patients with high-risk imaging features while offering the potential to reduce CMR testing requirements in under-resourced areas.

PMID:38280624 | DOI:10.1016/j.hrthm.2024.01.031

Sci Total Environ. 2024 Jan 25:170383. doi: 10.1016/j.scitotenv.2024.170383. Online ahead of print.

ABSTRACT

Dissolved oxygen (DO) depletion is a severe threat to aquatic ecosystems. Hence, using easily available routine water environmental variables without DO as inputs to predict the daily minimum DO concentration in rivers has huge practical significance in the watershed management. The daily minimum DO concentrations at the outlet of the Oyster River watershed in New Hampshire, USA, were predicted by a set of deep learning neural networks using meteorological data and high-frequency water level, water temperature, and specific conductance (CTD) data measured within the watershed. The dependent variable, DO concentration, was measured at the outlet. From April 2013 to March 2018, the dataset was separated into training, validation, and test portions with a ratio of 5:3:3. A Long Short-Term Memory (LSTM) model and a hybrid Convolutional Neural Networks (CNN-LSTM) model were trained and evaluated for predicting the daily minimum DO concentration. The hybrid CNN-LSTM model exhibited the better predictive stability but the comparable accuracy (the mean R2 value = 0.865) compared with the pure LSTM model (the mean R2 value = 0.839). The model performance (both the stability and accuracy) was improved by aggregating the input data frequency from 15-min of raw data to 24-h. Likewise, the modeling performance didn't benefit from including 'forecasted' meteorological data at the predicted time step in the input dataset. This study provided an efficient and low-cost approach to predict the water quality in rivers and streams to realize the scientific watershed management.

PMID:38280612 | DOI:10.1016/j.scitotenv.2024.170383

Sci Total Environ. 2024 Jan 26:170375. doi: 10.1016/j.scitotenv.2024.170375. Online ahead of print.

ABSTRACT

Dryland biodiversity is decreasing at an alarming rate. Advanced intelligent tools are urgently needed to rapidly, automatedly, and precisely detect dryland threatened species on a large scale for biological conservation. Here, we explored the performance of three deep convolutional neural networks (Deeplabv3+, Unet, and Pspnet models) on the intelligent recognition of rare species based on high-resolution (0.3 m) satellite images taken by an unmanned aerial vehicle (UAV). We focused on a threatened species, Populus euphratica, in the Tarim River Basin (China), where there has been a severe population decline in the 1970s and restoration has been carried out since 2000. The testing results showed that Unet outperforms Deeplabv3+ and Pspnet when the training samples are lower, while Deeplabv3+ performs best as the dataset increases. Overall, when training samples are 80, Deeplabv3+ had the best overall performance for Populus euphratica identification, with mean pixel accuracy (MPA) between 87.31 % and 90.2 %, which, on average is 3.74 % and 11.29 % higher than Unet and Pspnet, respectively. Deeplabv3+ can accurately detect the boundaries of Populus euphratica even in areas of dense vegetation, with lower identification uncertainty for each pixel than other models. This study developed a UAV imagery-based identification framework using deep learning with high resolution in large-scale regions. This approach can accurately capture the variation in dryland threatened species, especially those in inaccessible areas, thereby fostering rapid and efficient conservation actions.

PMID:38280598 | DOI:10.1016/j.scitotenv.2024.170375

J Affect Disord. 2024 Jan 25:S0165-0327(24)00162-9. doi: 10.1016/j.jad.2024.01.146. Online ahead of print.

ABSTRACT

INTRODUCTION: Mnemonic discrimination (MD), the ability to discriminate new stimuli from similar memories, putatively involves dentate gyrus pattern separation. Since lithium may normalize dentate gyrus functioning in lithium-responsive bipolar disorder (BD), we hypothesized that lithium treatment would be associated with better MD in lithium-responsive BD patients.

METHODS: BD patients (N = 69; NResponders = 16 [23 %]) performed the Continuous Visual Memory Test (CVMT), which requires discriminating between novel and previously seen images. Before testing, all patients had prophylactic lithium responsiveness assessed over ≥1 year of therapy (with the Alda Score), although only thirty-eight patients were actively prescribed lithium at time of testing (55 %; 12/16 responders, 26/53 nonresponders). We then used computational modelling to extract patient-specific MD indices. Linear models were used to test how (A) lithium treatment, (B) lithium responsiveness via the continuous Alda score, and (C) their interaction, affected MD.

RESULTS: Superior MD performance was associated with lithium treatment exclusively in lithium-responsive patients (Lithium x AldaScore β = 0.257 [SE 0.078], p = 0.002). Consistent with prior literature, increased age was associated with worse MD (β = -0.03 [SE 0.01], p = 0.005).

LIMITATIONS: Secondary pilot analysis of retrospectively collected data in a cross-sectional design limits generalizability.

CONCLUSION: Our study is the first to examine MD performance in BD. Lithium is associated with better MD performance only in lithium responders, potentially due to lithium's effects on dentate gyrus granule cell excitability. Our results may influence the development of behavioural probes for dentate gyrus neuronal hyperexcitability in BD.

PMID:38280568 | DOI:10.1016/j.jad.2024.01.146

Methods. 2024 Jan 25:S1046-2023(24)00029-X. doi: 10.1016/j.ymeth.2024.01.011. Online ahead of print.

ABSTRACT

MicroRNAs (miRNAs) are vital in regulating gene expression through binding to specific target sites on messenger RNAs (mRNAs), a process closely tied to cancer pathogenesis. Identifying miRNA functional targets is essential but challenging, due to incomplete genome annotation and an emphasis on known miRNA-mRNA interactions, restricting predictions of unknown ones. To address those challenges, we have developed a deep learning model based on miRNA functional target identification, named miTDS, to investigate miRNA-mRNA interactions. miTDS first employs a scoring mechanism to eliminate unstable sequence pairs and then utilizes a dynamic word embedding model based on the transformer architecture, enabling a comprehensive analysis of miRNA-mRNA interaction sites by harnessing the global contextual associations of each nucleotide. On this basis, miTDS fuses extended seed alignment representations learned in the multi-scale attention mechanism module with dynamic semantic representations extracted in the RNA-based dual-path module, which can further elucidate and predict miRNA and mRNA functions and interactions. To validate the effectiveness of miTDS, we conducted a thorough comparison with state-of-the-art miRNA-mRNA functional target prediction methods. The evaluation, performed on a dataset cross-referenced with entries from MirTarbase and Diana-TarBase, revealed that miTDS surpasses current methods in accurately predicting functional targets. In addition, our model exhibited proficiency in identifying A-to-I RNA editing sites, which represents an aberrant interaction that yields valuable insights into the suppression of cancerous processes. Moreover, all supporting source code and data can be downloaded from https://github.com/mingziiz/miTDS.

PMID:38280472 | DOI:10.1016/j.ymeth.2024.01.011

Ophthalmol Retina. 2024 Jan 25:S2468-6530(24)00045-9. doi: 10.1016/j.oret.2024.01.019. Online ahead of print.

ABSTRACT

OBJECTIVE: We aimed to develop a deep learning system capable of identifying subjects with cognitive impairment quickly and easily based on multimodal ocular images.

DESIGN: Cross-sectional study SUBJECTS: Participants of Beijing Eye Study 2011 and patients attending Beijing Tongren Eye Center and Beijing Tongren Hospital Physical Examination Center.

METHODS: We trained and validated a deep learning algorithm to assess cognitive impairment using retrospectively collected data from the Beijing Eye Study 2011. Cognitive impairment was defined as a Mini-Mental State Examination (MMSE) score <24. Based on fundus photographs and optical coherence tomography (OCT) images, we developed five models based on the following sets of images: macula-centered fundus photographs, optic disc-centered fundus photographs, fundus photographs of both fields, optical coherence tomography (OCT) images, and fundus photographs of both fields with OCT (multi-modal). The performance of the models was evaluated and compared in an external validation dataset, which was collected from patients attending Beijing Tongren Eye Center and Beijing Tongren Hospital Physical Examination Center.

MAIN OUTCOME MEASURES: Area under the curve (AUC).

RESULTS: A total of 9,424 retinal photographs and 4,712 OCT images were used to develop the model. The external validation sets from each center included 1,180 fundus photographs and 590 OCT images. Model comparison revealed that the multi-modal performed best, achieving an AUC of 0.820 in the internal validation set, 0.786 in external validation set 1 and 0.784 in external validation set 2. We evaluated the performance of the multi-model in different sexes and different age groups; there were no significant differences. The heatmap analysis showed that signals around the optic disc in fundus photographs and the retina and choroid around the macular and optic disc regions in OCT images were used by the multi-modal to identify participants with cognitive impairment.

CONCLUSIONS: Fundus photographs and OCT can provide valuable information on cognitive function. Multi-modal models provide richer information compared to single-mode models. Deep learning algorithms based on multimodal retinal images may be capable of screening cognitive impairment. This technique has potential value for broader implementation in community-based screening or clinic settings.

PMID:38280426 | DOI:10.1016/j.oret.2024.01.019

Ophthalmol Retina. 2024 Jan 25:S2468-6530(24)00044-7. doi: 10.1016/j.oret.2024.01.018. Online ahead of print.

ABSTRACT

OBJECTIVE: To review recent technological advancement in imaging, surgical visualization, robotics technology and the use of artificial intelligence in surgical vitreoretinal (VR) diseases.

BACKGROUND: Technological advancements in imaging enhance both pre-operative and intra-operative management of surgical VR diseases. Widefield imaging in fundal photography and optical coherence tomography (OCT) can improve assessment of peripheral retinal disorders like retinal detachments, degeneration, and tumors. OCT angiography provides a rapid and non-invasive imaging of the retinal and choroidal vasculature. Surgical visualization has also improved with intraoperative OCT providing a detailed real-time assessment of retinal layers to guide surgical decisions. Heads-up display and head-mounted display utilize three-dimensional technology to provide surgeons with enhanced visual guidance and improved ergonomics during surgery. Intraocular robotics technology allows for greater surgical precision and is shown to be useful in retinal vein cannulation and subretinal drug delivery. In addition, deep learning techniques leverages on diverse data including widefield retinal photography and OCT for better predictive accuracy in classification, segmentation, and prognostication of many surgical VR diseases.

CONCLUSION: This review article summarized the latest updates in these areas and highlights the importance of continuous innovation and improvement in technology within the field. These advancements have the potential to reshape management of surgical VR diseases in the very near future and to ultimately improve patient care.

PMID:38280425 | DOI:10.1016/j.oret.2024.01.018

Brief Bioinform. 2024 Jan 22;25(2):bbad531. doi: 10.1093/bib/bbad531.

ABSTRACT

Rare antinuclear antibody (ANA) pattern recognition has been a widely applied technology for routine ANA screening in clinical laboratories. In recent years, the application of deep learning methods in recognizing ANA patterns has witnessed remarkable advancements. However, the majority of studies in this field have primarily focused on the classification of the most common ANA patterns, while another subset has concentrated on the detection of mitotic metaphase cells. To date, no prior research has been specifically dedicated to the identification of rare ANA patterns. In the present paper, we introduce a novel attention-based enhancement framework, which was designed for the recognition of rare ANA patterns in ANA-indirect immunofluorescence images. More specifically, we selected the algorithm with the best performance as our target detection network by conducting comparative experiments. We then further developed and enhanced the chosen algorithm through a series of optimizations. Then, attention mechanism was introduced to facilitate neural networks in expediting the learning process, extracting more essential and distinctive features for the target features that belong to the specific patterns. The proposed approach has helped to obtained high precision rate of 86.40%, 82.75% recall, 84.24% F1 score and 84.64% mean average precision for a 9-category rare ANA pattern detection task on our dataset. Finally, we evaluated the potential of the model as medical technologist assistant and observed that the technologist's performance improved after referring to the results of the model prediction. These promising results highlighted its potential as an efficient and reliable tool to assist medical technologists in their clinical practice.

PMID:38279651 | DOI:10.1093/bib/bbad531

Brief Bioinform. 2024 Jan 22;25(2):bbae005. doi: 10.1093/bib/bbae005.

ABSTRACT

The identification of human-herpesvirus protein-protein interactions (PPIs) is an essential and important entry point to understand the mechanisms of viral infection, especially in malignant tumor patients with common herpesvirus infection. While natural language processing (NLP)-based embedding techniques have emerged as powerful approaches, the application of multi-modal embedding feature fusion to predict human-herpesvirus PPIs is still limited. Here, we established a multi-modal embedding feature fusion-based LightGBM method to predict human-herpesvirus PPIs. In particular, we applied document and graph embedding approaches to represent sequence, network and function modal features of human and herpesviral proteins. Training our LightGBM models through our compiled non-rigorous and rigorous benchmarking datasets, we obtained significantly better performance compared to individual-modal features. Furthermore, our model outperformed traditional feature encodings-based machine learning methods and state-of-the-art deep learning-based methods using various benchmarking datasets. In a transfer learning step, we show that our model that was trained on human-herpesvirus PPI dataset without cytomegalovirus data can reliably predict human-cytomegalovirus PPIs, indicating that our method can comprehensively capture multi-modal fusion features of protein interactions across various herpesvirus subtypes. The implementation of our method is available at https://github.com/XiaodiYangpku/MultimodalPPI/.

PMID:38279649 | DOI:10.1093/bib/bbae005

Adv Sci (Weinh). 2024 Jan 26:e2307609. doi: 10.1002/advs.202307609. Online ahead of print.

ABSTRACT

Noncommunicable diseases (NCD), such as obesity, diabetes, and cardiovascular disease, are defining healthcare challenges of the 21st century. Medical infrastructure, which for decades sought to reduce the incidence and severity of communicable diseases, has proven insufficient in meeting the intensive, long-term monitoring needs of many NCD disease patient groups. In addition, existing portable devices with rigid electronics are still limited in clinical use due to unreliable data, limited functionality, and lack of continuous measurement ability. Here, a wearable system for at-home cardiovascular monitoring of postpartum women-a group with urgently unmet NCD needs in the United States-using a cloud-integrated soft sternal device with conformal nanomembrane sensors is introduced. A supporting mobile application provides device data to a custom cloud architecture for real-time waveform analytics, including medical device-grade blood pressure prediction via deep learning, and shares the results with both patient and clinician to complete a robust and highly scalable remote monitoring ecosystem. Validated in a month-long clinical study with 20 postpartum Black women, the system demonstrates its ability to remotely monitor existing disease progression, stratify patient risk, and augment clinical decision-making by informing interventions for groups whose healthcare needs otherwise remain unmet in standard clinical practice.

PMID:38279514 | DOI:10.1002/advs.202307609

Exp Biol Med (Maywood). 2024 Jan 26:15353702231220664. doi: 10.1177/15353702231220664. Online ahead of print.

ABSTRACT

This study aims to construct and evaluate a deep learning model, utilizing ultrasound images, to accurately differentiate benign and malignant thyroid nodules. The objective includes visualizing the model's process for interpretability and comparing its diagnostic precision with a cohort of 80 radiologists. We employed ResNet as the classification backbone for thyroid nodule prediction. The model was trained using 2096 ultrasound images of 655 distinct thyroid nodules. For performance evaluation, an independent test set comprising 100 cases of thyroid nodules was curated. In addition, to demonstrate the superiority of the artificial intelligence (AI) model over radiologists, a Turing test was conducted with 80 radiologists of varying clinical experience. This was meant to assess which group of radiologists' conclusions were in closer alignment with AI predictions. Furthermore, to highlight the interpretability of the AI model, gradient-weighted class activation mapping (Grad-CAM) was employed to visualize the model's areas of focus during its prediction process. In this cohort, AI diagnostics demonstrated a sensitivity of 81.67%, a specificity of 60%, and an overall diagnostic accuracy of 73%. In comparison, the panel of radiologists on average exhibited a diagnostic accuracy of 62.9%. The AI's diagnostic process was significantly faster than that of the radiologists. The generated heat-maps highlighted the model's focus on areas characterized by calcification, solid echo and higher echo intensity, suggesting these areas might be indicative of malignant thyroid nodules. Our study supports the notion that deep learning can be a valuable diagnostic tool with comparable accuracy to experienced senior radiologists in the diagnosis of malignant thyroid nodules. The interpretability of the AI model's process suggests that it could be clinically meaningful. Further studies are necessary to improve diagnostic accuracy and support auxiliary diagnoses in primary care settings.

PMID:38279511 | DOI:10.1177/15353702231220664

BMC Genomics. 2024 Jan 26;25(1):117. doi: 10.1186/s12864-024-10019-5.

ABSTRACT

BACKGROUND: In cellular activities, essential proteins play a vital role and are instrumental in comprehending fundamental biological necessities and identifying pathogenic genes. Current deep learning approaches for predicting essential proteins underutilize the potential of gene expression data and are inadequate for the exploration of dynamic networks with limited evaluation across diverse species.

RESULTS: We introduce ECDEP, an essential protein identification model based on evolutionary community discovery. ECDEP integrates temporal gene expression data with a protein-protein interaction (PPI) network and employs the 3-Sigma rule to eliminate outliers at each time point, constructing a dynamic network. Next, we utilize edge birth and death information to establish an interaction streaming source to feed into the evolutionary community discovery algorithm and then identify overlapping communities during the evolution of the dynamic network. SVM recursive feature elimination (RFE) is applied to extract the most informative communities, which are combined with subcellular localization data for classification predictions. We assess the performance of ECDEP by comparing it against ten centrality methods, four shallow machine learning methods with RFE, and two deep learning methods that incorporate multiple biological data sources on Saccharomyces. Cerevisiae (S. cerevisiae), Homo sapiens (H. sapiens), Mus musculus, and Caenorhabditis elegans. ECDEP achieves an AP value of 0.86 on the H. sapiens dataset and the contribution ratio of community features in classification reaches 0.54 on the S. cerevisiae (Krogan) dataset.

CONCLUSIONS: Our proposed method adeptly integrates network dynamics and yields outstanding results across various datasets. Furthermore, the incorporation of evolutionary community discovery algorithms amplifies the capacity of gene expression data in classification.

PMID:38279081 | DOI:10.1186/s12864-024-10019-5