Am J Clin Pathol. 2024 Aug 13:aqae107. doi: 10.1093/ajcp/aqae107. Online ahead of print.

ABSTRACT

OBJECTIVES: This review summarizes the current and potential uses of artificial intelligence (AI) in the current state of clinical microbiology with a focus on replacement of labor-intensive tasks.

METHODS: A search was conducted on PubMed using the key terms clinical microbiology and artificial intelligence. Studies were reviewed for relevance to clinical microbiology, current diagnostic techniques, and potential advantages of AI in routine microbiology workflows.

RESULTS: Numerous studies highlight potential labor, as well as diagnostic accuracy, benefits to the implementation of AI for slide-based and macroscopic digital image analyses. These range from Gram stain interpretation to categorization and quantitation of culture growth.

CONCLUSIONS: Artificial intelligence applications in clinical microbiology significantly enhance diagnostic accuracy and efficiency, offering promising solutions to labor-intensive tasks and staffing shortages. More research efforts and US Food and Drug Administration clearance are still required to fully incorporate these AI applications into routine clinical laboratory practices.

PMID:39136261 | DOI:10.1093/ajcp/aqae107

Int J Neural Syst. 2024 Oct;34(10):2450055. doi: 10.1142/S0129065724500552.

ABSTRACT

Automatic seizure detection from Electroencephalography (EEG) is of great importance in aiding the diagnosis and treatment of epilepsy due to the advantages of convenience and economy. Existing seizure detection methods are usually patient-specific, the training and testing are carried out on the same patient, limiting their scalability to other patients. To address this issue, we propose a cross-subject seizure detection method via unsupervised domain adaptation. The proposed method aims to obtain seizure specific information through shallow and deep feature alignments. For shallow feature alignment, we use convolutional neural network (CNN) to extract seizure-related features. The distribution gap of the shallow features between different patients is minimized by multi-kernel maximum mean discrepancies (MK-MMD). For deep feature alignment, adversarial learning is utilized. The feature extractor tries to learn feature representations that try to confuse the domain classifier, making the extracted deep features more generalizable to new patients. The performance of our method is evaluated on the CHB-MIT and Siena databases in epoch-based experiments. Additionally, event-based experiments are also conducted on the CHB-MIT dataset. The results validate the feasibility of our method in diminishing the domain disparities among different patients.

PMID:39136190 | DOI:10.1142/S0129065724500552

Ecol Lett. 2024 Aug;27(8):e14495. doi: 10.1111/ele.14495.

ABSTRACT

In the realm of biological image analysis, deep learning (DL) has become a core toolkit, for example for segmentation and classification. However, conventional DL methods are challenged by large biodiversity datasets characterized by unbalanced classes and hard-to-distinguish phenotypic differences between them. Here we present BioEncoder, a user-friendly toolkit for metric learning, which overcomes these challenges by focussing on learning relationships between individual data points rather than on the separability of classes. BioEncoder is released as a Python package, created for ease of use and flexibility across diverse datasets. It features taxon-agnostic data loaders, custom augmentation options, and simple hyperparameter adjustments through text-based configuration files. The toolkit's significance lies in its potential to unlock new research avenues in biological image analysis while democratizing access to advanced deep metric learning techniques. BioEncoder focuses on the urgent need for toolkits bridging the gap between complex DL pipelines and practical applications in biological research.

PMID:39136114 | DOI:10.1111/ele.14495

Front Oral Health. 2024 Jul 29;5:1442100. doi: 10.3389/froh.2024.1442100. eCollection 2024.

ABSTRACT

The current review focuses on how artificial intelligence (AI) and robotics can be applied to the field of Prosthodontics and oral implantology. The classification and methodologies of AI and application of AI and robotics in various aspects of Prosthodontics is summarized. The role of AI has potentially expanded in dentistry. It plays a vital role in data management, diagnosis, and treatment planning and administrative tasks. It has widespread applications in Prosthodontics owing to its immense diagnostic capability and possible therapeutic application. AI and robotics are next-generation technologies that are opening new avenues of growth and exploration for Prosthodontics. The current surge in digital human-centered automation has greatly benefited the dental field, as it transforms towards a new robotic, machine learning, and artificial intelligence era. The application of robotics and AI in the dental field aims to improve dependability, accuracy, precision, and efficiency by enabling the widespread adoption of cutting-edge dental technologies in future. Hence, the objective of the current review was to represent literature relevant to the applications of robotics and AI and in the context of diagnosis and clinical decision-making and predict successful treatment in Prosthodontics and oral implantology.

PMID:39135907 | PMC:PMC11317471 | DOI:10.3389/froh.2024.1442100

Comput Struct Biotechnol J. 2024 Jul 6;23:2995-3018. doi: 10.1016/j.csbj.2024.06.037. eCollection 2024 Dec.

ABSTRACT

The 4,6-substituted-1,3,5-triazin-2(1H)-ones are promising inhibitors of human DNA topoisomerase IIα. To further develop this chemical class targeting the enzyme´s ATP binding site, the triazin-2(1H)-one substitution position 6 was optimized. Inspired by binding of preclinical substituted 9H-purine derivative, bicyclic substituents were incorporated at position 6 and the utility of this modification was validated by a combination of molecular simulations, dynamic pharmacophores, and free energy calculations. Considering also predictions of Deepfrag, a software developed for structure-based lead optimization based on deep learning, compounds with both bicyclic and monocyclic substitutions were synthesized and investigated for their inhibitory activity. The SAR data showed that the bicyclic substituted compounds exhibited good inhibition of topo IIα, comparable to their mono-substituted counterparts. Further evaluation on a panel of human protein kinases showed selectivity for the inhibition of topo IIα. Mechanistic studies indicated that the compounds acted predominantly as catalytic inhibitors, with some exhibiting topo IIα poison effects at higher concentrations. Integration of STD NMR experiments and molecular simulations, provided insights into the binding model and highlighted the importance of the Asn120 interaction and hydrophobic interactions with substituents at positions 4 and 6. In addition, NCI-60 screening demonstrated cytotoxicity of the compounds with bicyclic substituents and identified sensitive human cancer cell lines, underlining the translational relevance of our findings for further preclinical development of this class of compounds. The study highlights the synergy between simulation and AI-based approaches in efficiently guiding molecular design for drug optimization, which has implications for further preclinical development of this class of compounds.

PMID:39135887 | PMC:PMC11318567 | DOI:10.1016/j.csbj.2024.06.037

Imeta. 2024 Jun 15;3(4):e214. doi: 10.1002/imt2.214. eCollection 2024 Aug.

ABSTRACT

Rapid and accurate diagnostic tests are fundamental for improving patient outcomes and combating infectious diseases. The Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) Cas12a-based detection system has emerged as a promising solution for on-site nucleic acid testing. Nonetheless, the effective design of CRISPR RNA (crRNA) for Cas12a-based detection remains challenging and time-consuming. In this study, we propose an enhanced crRNA design system with deep learning for Cas12a-mediated diagnostics, referred to as EasyDesign. This system employs an optimized convolutional neural network (CNN) prediction model, trained on a comprehensive data set comprising 11,496 experimentally validated Cas12a-based detection cases, encompassing a wide spectrum of prevalent pathogens, achieving Spearman's ρ = 0.812. We further assessed the model performance in crRNA design for four pathogens not included in the training data: Monkeypox Virus, Enterovirus 71, Coxsackievirus A16, and Listeria monocytogenes. The results demonstrated superior prediction performance compared to the traditional experiment screening. Furthermore, we have developed an interactive web server (https://crispr.zhejianglab.com/) that integrates EasyDesign with recombinase polymerase amplification (RPA) primer design, enhancing user accessibility. Through this web-based platform, we successfully designed optimal Cas12a crRNAs for six human papillomavirus (HPV) subtypes. Remarkably, all the top five predicted crRNAs for each HPV subtype exhibited robust fluorescent signals in CRISPR assays, thereby suggesting that the platform could effectively facilitate clinical sample testing. In conclusion, EasyDesign offers a rapid and reliable solution for crRNA design in Cas12a-based detection, which could serve as a valuable tool for clinical diagnostics and research applications.

PMID:39135699 | PMC:PMC11316927 | DOI:10.1002/imt2.214

Psychophysiology. 2024 Aug 12:e14667. doi: 10.1111/psyp.14667. Online ahead of print.

ABSTRACT

Pupillometry has been used in the studies of postural control to assess cognitive load during dual tasks, but its response to increased balance task intensity has not been investigated. Furthermore, it is unknown whether side-specific changes in pupil diameter occur with more demanding balance tasks providing additional insights into postural control. The two aims of this study were to analyze differences in steady-state pupil diameter between balance tasks with increased intensity and to determine whether there are side-specific changes. Forty-eight healthy subjects performed parallel and left and right one-legged stances on a force plate with and without foam with right and left pupil diameters measured with a mobile infrared eye-tracker. Differences between balance tasks in parameters (average pupil diameter of each eye, average of both pupil diameters and the difference between the left and right pupil diameter) were analyzed using a two-way repeated measures analysis of variance, and deep learning neural network models were used to investigate how pupillometry predicted each balance task. The pupil diameter of the left eye, the average pupil diameter of both eyes and the difference in pupil diameters increased statistically significantly from simpler to more demanding balance tasks, with this being more pronounced for the left eye. The deep learning neural network models revealed side-specific changes in pupil diameter with more demanding balance tasks. This study confirms pupillary responses to increased intensity of balance task and indicates side-specific pupil responses that could be related to task-specific involvement of higher levels of postural control.

PMID:39135357 | DOI:10.1111/psyp.14667

Comput Biol Med. 2024 Aug 11;180:109000. doi: 10.1016/j.compbiomed.2024.109000. Online ahead of print.

ABSTRACT

The fetus's health is evaluated with the biometric parameters obtained from the low-resolution ultrasound images. The accuracy of biometric parameters in existing protocols typically depends on conventional image processing approaches and hence, is prone to error. This study introduces the Attention Gate Double U-Net with Guided Decoder (ADU-GD) model specifically crafted for fetal biometric parameter prediction. The attention network and guided decoder are specifically designed to dynamically merge local features with their global dependencies, enhancing the precision of parameter estimation. The ADU-GD displays superior performance with Mean Absolute Error of 0.99 mm and segmentation accuracy of 99.1 % when benchmarked against the well-established models. The proposed model consistently achieved a high Dice index score of about 99.1 ± 0.8, with a minimal Hausdorff distance of about 1.01 ± 1.07 and a low Average Symmetric Surface Distance of about 0.25 ± 0.21, demonstrating the model's excellence. In a comprehensive evaluation, ADU-GD emerged as a frontrunner, outperforming existing deep-learning models such as Double U-Net, DeepLabv3, FCN-32s, PSPNet, SegNet, Trans U-Net, Swin U-Net, Mask-R2CNN, and RDHCformer models in terms of Mean Absolute Error for crucial fetal dimensions, including Head Circumference, Abdomen Circumference, Femur Length, and BiParietal Diameter. It achieved superior accuracy with MAE values of 2.2 mm, 2.6 mm, 0.6 mm, and 1.2 mm, respectively.

PMID:39133952 | DOI:10.1016/j.compbiomed.2024.109000

BMC Public Health. 2024 Aug 12;24(1):2171. doi: 10.1186/s12889-024-19590-8.

ABSTRACT

BACKGROUND: Influenza, an acute infectious respiratory disease, presents a significant global health challenge. Accurate prediction of influenza activity is crucial for reducing its impact. Therefore, this study seeks to develop a hybrid Convolution Neural Network-Long Short Term Memory neural network (CNN-LSTM) model to forecast the percentage of influenza-like-illness (ILI) rate in Hebei Province, China. The aim is to provide more precise guidance for influenza prevention and control measures.

METHODS: Using ILI% data from 28 national sentinel hospitals in the Hebei Province, spanning from 2010 to 2022, we employed the Python deep learning framework PyTorch to develop the CNN-LSTM model. Additionally, we utilized R and Python to develop four other models commonly used for predicting infectious diseases. After constructing the models, we employed these models to make retrospective predictions, and compared each model's prediction performance using mean absolute error (MAE), root mean square error (RMSE), mean absolute percentage error (MAPE), and other evaluation metrics.

RESULTS: Based on historical ILI% data from 28 national sentinel hospitals in Hebei Province, the Seasonal Auto-Regressive Indagate Moving Average (SARIMA), Extreme Gradient Boosting (XGBoost), Convolution Neural Network (CNN), Long Short Term Memory neural network (LSTM) models were constructed. On the testing set, all models effectively predicted the ILI% trends. Subsequently, these models were used to forecast over different time spans. Across various forecasting periods, the CNN-LSTM model demonstrated the best predictive performance, followed by the XGBoost model, LSTM model, CNN model, and SARIMA model, which exhibited the least favorable performance.

CONCLUSION: The hybrid CNN-LSTM model had better prediction performances than the SARIMA model, CNN model, LSTM model, and XGBoost model. This hybrid model could provide more accurate influenza activity projections in the Hebei Province.

PMID:39135162 | DOI:10.1186/s12889-024-19590-8

J Transl Med. 2024 Aug 12;22(1):756. doi: 10.1186/s12967-024-05567-z.

ABSTRACT

BACKGROUND: Decoding human genomic sequences requires comprehensive analysis of DNA sequence functionality. Through computational and experimental approaches, researchers have studied the genotype-phenotype relationship and generate important datasets that help unravel complicated genetic blueprints. Thus, the recently developed artificial intelligence methods can be used to interpret the functions of those DNA sequences.

METHODS: This study explores the use of deep learning, particularly pre-trained genomic models like DNA_bert_6 and human_gpt2-v1, in interpreting and representing human genome sequences. Initially, we meticulously constructed multiple datasets linking genotypes and phenotypes to fine-tune those models for precise DNA sequence classification. Additionally, we evaluate the influence of sequence length on classification results and analyze the impact of feature extraction in the hidden layers of our model using the HERV dataset. To enhance our understanding of phenotype-specific patterns recognized by the model, we perform enrichment, pathogenicity and conservation analyzes of specific motifs in the human endogenous retrovirus (HERV) sequence with high average local representation weight (ALRW) scores.

RESULTS: We have constructed multiple genotype-phenotype datasets displaying commendable classification performance in comparison with random genomic sequences, particularly in the HERV dataset, which achieved binary and multi-classification accuracies and F1 values exceeding 0.935 and 0.888, respectively. Notably, the fine-tuning of the HERV dataset not only improved our ability to identify and distinguish diverse information types within DNA sequences but also successfully identified specific motifs associated with neurological disorders and cancers in regions with high ALRW scores. Subsequent analysis of these motifs shed light on the adaptive responses of species to environmental pressures and their co-evolution with pathogens.

CONCLUSIONS: These findings highlight the potential of pre-trained genomic models in learning DNA sequence representations, particularly when utilizing the HERV dataset, and provide valuable insights for future research endeavors. This study represents an innovative strategy that combines pre-trained genomic model representations with classical methods for analyzing the functionality of genome sequences, thereby promoting cross-fertilization between genomics and artificial intelligence.

PMID:39135093 | DOI:10.1186/s12967-024-05567-z

J Nanobiotechnology. 2024 Aug 13;22(1):482. doi: 10.1186/s12951-024-02686-z.

ABSTRACT

Treatment-induced ototoxicity and accompanying hearing loss are a great concern associated with chemotherapeutic or antibiotic drug regimens. Thus, prophylactic cure or early treatment is desirable by local delivery to the inner ear. In this study, we examined a novel way of intratympanically delivered sustained nanoformulation by using crosslinked hybrid nanoparticle (cHy-NPs) in a thermoresponsive hydrogel i.e. thermogel that can potentially provide a safe and effective treatment towards the treatment-induced or drug-induced ototoxicity. The prophylactic treatment of the ototoxicity can be achieved by using two therapeutic molecules, Flunarizine (FL: T-type calcium channel blocker) and Honokiol (HK: antioxidant) co-encapsulated in the same delivery system. Here we investigated, FL and HK as cytoprotective molecules against cisplatin-induced toxic effects in the House Ear Institute - Organ of Corti 1 (HEI-OC1) cells and in vivo assessments on the neuromast hair cell protection in the zebrafish lateral line. We observed that cytotoxic protective effect can be enhanced by using FL and HK in combination and developing a robust drug delivery formulation. Therefore, FL-and HK-loaded crosslinked hybrid nanoparticles (FL-cHy-NPs and HK-cHy-NPs) were synthesized using a quality-by-design approach (QbD) in which design of experiment-central composite design (DoE-CCD) following the standard least-square model was used for nanoformulation optimization. The physicochemical characterization of FL and HK loaded-NPs suggested the successful synthesis of spherical NPs with polydispersity index < 0.3, drugs encapsulation (> 75%), drugs loading (~ 10%), stability (> 2 months) in the neutral solution, and appropriate cryoprotectant selection. We assessed caspase 3/7 apopototic pathway in vitro that showed significantly reduced signals of caspase 3/7 activation after the FL-cHy-NPs and HK-cHy-NPs (alone or in combination) compared to the CisPt. The final formulation i.e. crosslinked-hybrid-nanoparticle-embedded-in-thermogel was developed by incorporating drug-loaded cHy-NPs in poloxamer-407, poloxamer-188, and carbomer-940-based hydrogel. A combination of artificial intelligence (AI)-based qualitative and quantitative image analysis determined the particle size and distribution throughout the visible segment. The developed formulation was able to release the FL and HK for at least a month. Overall, a highly stable nanoformulation was successfully developed for combating treatment-induced or drug-induced ototoxicity via local administration to the inner ear.

PMID:39135039 | DOI:10.1186/s12951-024-02686-z

Curr Gastroenterol Rep. 2024 Aug 12. doi: 10.1007/s11894-024-00942-8. Online ahead of print.

ABSTRACT

PURPOSE OF REVIEW: In this review, we aim to summarize the existing literature and future directions on the use of artificial intelligence (AI) for the diagnosis and treatment of PB (pancreaticobiliary) disorders. RECENT FINDINGS: AI models have been developed to aid in the diagnosis and management of PB disorders such as pancreatic adenocarcinoma (PDAC), pancreatic neuroendocrine tumors (pNETs), acute pancreatitis, chronic pancreatitis, autoimmune pancreatitis, choledocholithiasis, indeterminate biliary strictures, cholangiocarcinoma and endoscopic procedures such as ERCP, EUS, and cholangioscopy. Recent studies have integrated radiological, endoscopic and pathological data to develop models to aid in better detection and prognostication of these disorders. AI is an indispensable proponent in the future practice of medicine. It has been extensively studied and approved for use in the detection of colonic polyps. AI models based on clinical, laboratory, and radiomics have been developed to aid in the diagnosis and management of various PB disorders and its application is ever expanding. Despite promising results, these AI-based models need further external validation to be clinically applicable.

PMID:39134866 | DOI:10.1007/s11894-024-00942-8

Methods Mol Biol. 2024;2831:179-197. doi: 10.1007/978-1-0716-3969-6_12.

ABSTRACT

Digital reconstruction of neuronal structures from 3D neuron microscopy images is critical for the quantitative investigation of brain circuits and functions. Currently, neuron reconstructions are mainly obtained by manual or semiautomatic methods. However, these ways are labor-intensive, especially when handling the huge volume of whole brain microscopy imaging data. Here, we present a deep-learning-based neuron morphology analysis toolbox (DNeuroMAT) for automated analysis of neuron microscopy images, which consists of three modules: neuron segmentation, neuron reconstruction, and neuron critical points detection.

PMID:39134850 | DOI:10.1007/978-1-0716-3969-6_12

Sci Rep. 2024 Aug 12;14(1):18736. doi: 10.1038/s41598-024-69617-8.

ABSTRACT

Monkeypox (Mpox), a zoonotic illness triggered by the monkeypox virus (MPXV), poses a significant threat since it may be transmitted and has no cure. This work introduces a computational method to predict Protein-Protein Interactions (PPIs) during MPXV infection. The objective is to discover prospective drug targets and repurpose current potential Food and Drug Administration (FDA) drugs for therapeutic purposes. In this work, ensemble features, comprising 2-5 node graphlet attributes and protein composition-based features are utilized for Deep Learning (DL) models to predict PPIs. The technique that is used here demonstrated an excellent prediction performance for PPI on both the Human Integrated Protein-Protein Interaction Reference (HIPPIE) and MPXV-Human PPI datasets. In addition, the human protein targets for MPXV have been identified accurately along with the detection of possible therapeutic targets. Furthermore, the validation process included conducting docking research studies on potential FDA drugs like Nicotinamide Adenine Dinucleotide and Hydrogen (NADH), Fostamatinib, Glutamic acid, Cannabidiol, Copper, and Zinc in DrugBank identified via research on drug repurposing and the Drug Consensus Score (DCS) for MPXV. This has been achieved by employing the primary crystal structures of MPXV, which are now accessible. The docking study is also supported by Molecular Dynamics (MD) simulation. The results of our study emphasize the effectiveness of using ensemble feature-based PPI prediction to understand the molecular processes involved in viral infection and to aid in the development of repurposed drugs for emerging infectious diseases such as, but not limited to, Mpox. The source code and link to data used in this work is available at: https://github.com/CMATERJU-BIOINFO/In-Silico-Drug-Repurposing-Methodology-To-Suggest-Therapies-For-Emerging-Threats-like-Mpox .

PMID:39134619 | DOI:10.1038/s41598-024-69617-8

Light Sci Appl. 2024 Aug 13;13(1):188. doi: 10.1038/s41377-024-01556-5.

ABSTRACT

The surge in interest regarding the next generation of optical fiber transmission has stimulated the development of digital signal processing (DSP) schemes that are highly cost-effective with both high performance and low complexity. As benchmarks for nonlinear compensation methods, however, traditional DSP designed with block-by-block modules for linear compensations, could exhibit residual linear effects after compensation, limiting the nonlinear compensation performance. Here we propose a high-efficient design thought for DSP based on the learnable perspectivity, called learnable DSP (LDSP). LDSP reuses the traditional DSP modules, regarding the whole DSP as a deep learning framework and optimizing the DSP parameters adaptively based on backpropagation algorithm from a global scale. This method not only establishes new standards in linear DSP performance but also serves as a critical benchmark for nonlinear DSP designs. In comparison to traditional DSP with hyperparameter optimization, a notable enhancement of approximately 1.21 dB in the Q factor for 400 Gb/s signal after 1600 km fiber transmission is experimentally demonstrated by combining LDSP and perturbation-based nonlinear compensation algorithm. Benefiting from the learnable model, LDSP can learn the best configuration adaptively with low complexity, reducing dependence on initial parameters. The proposed approach implements a symbol-rate DSP with a small bit error rate (BER) cost in exchange for a 48% complexity reduction compared to the conventional 2 samples/symbol processing. We believe that LDSP represents a new and highly efficient paradigm for DSP design, which is poised to attract considerable attention across various domains of optical communications.

PMID:39134543 | DOI:10.1038/s41377-024-01556-5

Nephrology (Carlton). 2024 Aug 12. doi: 10.1111/nep.14376. Online ahead of print.

ABSTRACT

AIM: This study aimed to explore the value of ultrasound (US) images in chronic kidney disease (CKD) screening by constructing a CKD screening model based on grey-scale US images.

METHODS: According to the CKD diagnostic criteria, 1049 patients from Tongde Hospital of Zhejiang Province were retrospectively enrolled in the study. A total of 4365 renal US images were collected from these patients. Convolutional neural networks were used for feature extractions and a screening model was constructed by fusing ResNet34 and texture features to identify CKD and its stage. A comparative analysis was performed to compare the diagnosis results of the model with physicians.

RESULTS: When diagnosing CKD or non-CKD, the receiver operating characteristic curve (AUC) of our model was 0.918 and that of the senior physician group was 0.869 (p < .05). For the diagnosis of CKD stage, the AUC of our model for CKD G1-G3 was 0.781, 0.880, and 0.905, respectively, while the AUC of the senior physician group for CKD G1-G3 was 0.506, 0.586, and 0.796, respectively; all differences were statistically significant (p < .05). The diagnostic efficiency of our model for CKD G4 and G5 reached the level of the senior physicians group. Specifically, the AUC of our model for CKD G4-G5 was 0.867 and 0.931, respectively, while the AUC of the senior physician group for CKD G4-G5 was 0.838 and 0.963, respectively (all p > .05).

CONCLUSIONS: Our deep learning radiomics model is more effective than senior physicians in the diagnosis of early CKD.

PMID:39134509 | DOI:10.1111/nep.14376

Clin Exp Optom. 2024 Aug 12:1-6. doi: 10.1080/08164622.2024.2385565. Online ahead of print.

ABSTRACT

CLINICAL RELEVANCE: The challenges of establishing retinal screening programs in rural settings may be mitigated by the emergence of deep-learning systems for early disease detection.

BACKGROUND: Deep-learning systems have demonstrated promising results in retinal disease detection and may be particularly useful in rural settings where accessibility remains a barrier to equitable service provision. This study aims to evaluate the real-world performance of Thirona RetCAD for the detection of referable diabetic retinopathy and age-related macular degeneration in a rural Australian population.

METHODS: Colour fundus images from participants with known diabetic retinopathy or age-related macular degeneration were randomly selected from ophthalmology clinics in four rural Australian centres. Grading was confirmed retrospectively by two retinal specialists. RetCAD produced a quantitative measure (0-100) for DR and AMD severity. The area under the ROC curve (AUC) was calculated. Sensitivity, specificity, and positive and negative predictive values were calculated at a pre-defined cut-point of ≥50.

RESULTS: A total of 150 images from 82 participants were included. The mean age (SD) was 64.0 (12.8) years. Seventy-nine (52.7%) eyes had evidence of referable DR, while 54 (36.0%) had evidence of referable AMD. The AUC for referable DR detection was 0.971 (95% CI 0.950-0.936) with a sensitivity of 86.1% (76.8%-92.0%) and a specificity of 91.6% (82.8%-96.1%) at the pre-defined cut-point. Using the Youden Index method, the optimal cut-point was 41.2 (sensitivity 93.7%, specificity 90.1%). The AUC for the detection of referable AMD was 0.880 (0.824-0.936). At the pre-defined cut-point sensitivity was 88.9% (77.8%-94.8%) and specificity was 66.7% (56.8%-75.3%). The optimal cut-point was 52.6 (sensitivity 87.0%, specificity 75.0%).

CONCLUSION: RetCAD is comparable with but does not outperform equivalent deep-learning systems for retinal disease detection. RetCAD may be suitable as an automated screening tool in a rural Australian setting.

PMID:39134384 | DOI:10.1080/08164622.2024.2385565

Sci Total Environ. 2024 Aug 10:175451. doi: 10.1016/j.scitotenv.2024.175451. Online ahead of print.

ABSTRACT

Long-term trend forecast of chlorophyll-a concentration (Chla) holds significant implications for eutrophication management and pollution control planning on lakes, especially under the background of climate change. However, it is a challenging task due to the mixture of trend, seasonal and residual components in time series and the nonlinear relationships between Chla and the hydro-environmental factors. Here we developed a hybrid approach for long-term trend forecast of Chla in lakes, taking the Lake Taihu as an instantiation case, by the integration of Seasonal and Trend decomposition using Loess (STL), wavelet coherence, and Convolutional Neural Network with Bidirectional Long Short-Term Memory (CNN-BiLSTM). The results showed that long-term trends of Chla and the hydro-environmental factors could be effectively separated from the seasonal and residual terms by STL method, thereby enhancing the characterization of long-term variation. The resonance pattern and time lag between Chla and the hydro-environmental factors in the time-frequency domain were accurately identified by wavelet coherence. Chla responded quickly to variations in TP, but showed a time lag response to variation in WT in Lake Taihu. The forecasting method using multivariate and CNN-BiLSTM largely outperformed the other methods for Lake Taihu with regards to R2, RMSE, IOA and peak capture capability, owning to the combination of CNN for extracting local features and the integration of bidirectional propagation mechanism for the acquisition of higher-level features. The proposed hybrid deep learning approach offers an effective solution for the long-term trend forecast of algal blooms in eutrophic lakes and is capable of addressing the complex attributes of hydro-environmental data.

PMID:39134277 | DOI:10.1016/j.scitotenv.2024.175451

Med Phys. 2024 Aug 12. doi: 10.1002/mp.17343. Online ahead of print.

ABSTRACT

BACKGROUND: The volume measurement of intracerebral hemorrhage (ICH) and intraventricular hemorrhage (IVH) provides critical information for precise treatment of patients with spontaneous ICH but remains a big challenge, especially for IVH segmentation. However, the previously proposed ICH and IVH segmentation tools lack external validation and segmentation quality assessment.

PURPOSE: This study aimed to develop a robust deep learning model for the segmentation of ICH and IVH with external validation, and to provide quality assessment for IVH segmentation.

METHODS: In this study, a Residual Encoding Unet (REUnet) for the segmentation of ICH and IVH was developed using a dataset composed of 977 CT images (all contained ICH, and 338 contained IVH; a five-fold cross-validation procedure was adopted for training and internal validation), and externally tested using an independent dataset consisting of 375 CT images (all contained ICH, and 105 contained IVH). The performance of REUnet was compared with six other advanced deep learning models. Subsequently, three approaches, including Prototype Segmentation (ProtoSeg), Test Time Dropout (TTD), and Test Time Augmentation (TTA), were employed to derive segmentation quality scores in the absence of ground truth to provide a way to assess the segmentation quality in real practice.

RESULTS: For ICH segmentation, the median (lower-quantile-upper quantile) of Dice scores obtained from REUnet were 0.932 (0.898-0.953) for internal validation and 0.888 (0.859-0.916) for external test, both of which were better than those of other models while comparable to that of nnUnet3D in external test. For IVH segmentation, the Dice scores obtained from REUnet were 0.826 (0.757-0.868) for internal validation and 0.777 (0.693-0.827) for external tests, which were better than those of all other models. The concordance correlation coefficients between the volumes estimated from the REUnet-generated segmentations and those from the manual segmentations for both ICH and IVH ranged from 0.944 to 0.987. For IVH segmentation quality assessment, the segmentation quality score derived from ProtoSeg was correlated with the Dice Score (Spearman r = 0.752 for the external test) and performed better than those from TTD (Spearman r = 0.718) and TTA (Spearman r = 0.260) in the external test. By setting a threshold to the segmentation quality score, we were able to identify low-quality IVH segmentation results by ProtoSeg.

CONCLUSIONS: The proposed REUnet offers a promising tool for accurate and automated segmentation of ICH and IVH, and for effective IVH segmentation quality assessment, and thus exhibits the potential to facilitate therapeutic decision-making for patients with spontaneous ICH in clinical practice.

PMID:39133935 | DOI:10.1002/mp.17343

J Appl Clin Med Phys. 2024 Aug 12:e14483. doi: 10.1002/acm2.14483. Online ahead of print.

ABSTRACT

PURPOSE: In recent years, the use of deep learning for medical image segmentation has become a popular trend, but its development also faces some challenges. Firstly, due to the specialized nature of medical data, precise annotation is time-consuming and labor-intensive. Training neural networks effectively with limited labeled data is a significant challenge in medical image analysis. Secondly, convolutional neural networks commonly used for medical image segmentation research often focus on local features in images. However, the recognition of complex anatomical structures or irregular lesions often requires the assistance of both local and global information, which has led to a bottleneck in its development. Addressing these two issues, in this paper, we propose a novel network architecture.

METHODS: We integrate a shift window mechanism to learn more comprehensive semantic information and employ a semi-supervised learning strategy by incorporating a flexible amount of unlabeled data. Specifically, a typical U-shaped encoder-decoder structure is applied to obtain rich feature maps. Each encoder is designed as a dual-branch structure, containing Swin modules equipped with windows of different size to capture features of multiple scales. To effectively utilize unlabeled data, a level set function is introduced to establish consistency between the function regression and pixel classification.

RESULTS: We conducted experiments on the COVID-19 CT dataset and DRIVE dataset and compared our approach with various semi-supervised and fully supervised learning models. On the COVID-19 CT dataset, we achieved a segmentation accuracy of up to 74.56%. Our segmentation accuracy on the DRIVE dataset was 79.79%.

CONCLUSIONS: The results demonstrate the outstanding performance of our method on several commonly used evaluation metrics. The high segmentation accuracy of our model demonstrates that utilizing Swin modules with different window sizes can enhance the feature extraction capability of the model, and the level set function can enable semi-supervised models to more effectively utilize unlabeled data. This provides meaningful insights for the application of deep learning in medical image segmentation. Our code will be released once the manuscript is accepted for publication.

PMID:39133901 | DOI:10.1002/acm2.14483