Deep generative spatiotemporal learning for integrating fracture mechanics in composite materials: inverse design, discovery, and optimization.

The trade-off between strength and toughness presents a fundamental challenge in engineering material design. Composite materials (CMs) can strategically arrange different materials to enhance both strength and toughness by optimizing the distribution of loads and increasing resistance to crack propagation. However, current data-driven computational modeling approaches for CM configuration optimization suffer from limitations of "substantial computational cost" and "poor predictive power over extrapolation spaces", making it difficult to integrate with global optimization algorithms, and ultimately limiting the discovery of materials with optimal tradeoffs. As a breakthrough, we propose a data-driven design framework with a multi-task DL architecture capable of accurately predicting local fields' spatiotemporal behavior, including stress evolution and crack propagation, alongside homogenized mechanical properties. Our model, trained on datasets generated from crack phase fields simulations of random configurations, demonstrated exceptional predictive performance even for unseen configurations with well organized patterns exploiting nature-inspired morphological features. Importantly, solely from composite material (CM) configurations, our model effectively predicts long-term spatiotemporal fields with an accuracy comparable to FEM but with a substantial reduction in computational time. By coupling the model's predictive power with genetic optimization algorithms, we demonstrated the framework's applicability in two representative inverse design tasks: devising CM configurations with mechanical properties beyond the training set and guiding desired crack pattern formation. Our research highlights the potential of artificial intelligence as a feasible alternative to conventional computational approaches for straightforward configurational and structural optimization.

Machine learning-based inverse design methods considering data characteristics and design space size in materials design and manufacturing: a review.

In the last few decades, the influence of machine learning has permeated many areas of science and technology, including the field of materials science. This toolkit of data driven methods accelerated the discovery and production of new materials by accurately predicting the complicated physical processes and mechanisms that are not fully described by existing materials theories. However, the availability of a growing number of increasingly complex machine learning models confronts us with the question of "which machine learning algorithm to employ". In this review, we provide a comprehensive review of common machine learning algorithms used for materials design, as well as a guideline for selecting the most appropriate model considering the nature of the design problem. To this end, we classify the material design problems into four categories of: (i) the training data set being sufficiently large to capture the trend of design space (interpolation problem), (ii) a vast design space that cannot be explored thoroughly with the initial training data set alone (extrapolation problem), (iii) multi-fidelity datasets (small accurate dataset and large approximate dataset), and (iv) only a small dataset available. The most successful machine learning-based surrogate models and design approaches will be discussed for each case along with pertinent literature. This review focuses mostly on the use of ML algorithms for the inverse design of complicated composite structures, a topic that has received a lot of attention recently with the rise of additive manufacturing.

Novel next generation sequencing panel method for the multiple detection and identification of foodborne pathogens in agricultural wastewater.

Detecting and identifying the origins of foodborne pathogen outbreaks is a challenging. The Next-Generation Sequencing (NGS) panel method offers a potential solution by enabling efficient screening and identification of various bacteria in one reaction. In this study, new NGS panel primer sets that target 18 specific virulence factor genes from six target pathogens (Bacillus cereus, Yersinia enterocolitica, Staphylococcus aureus, Vibrio cholerae, Vibrio parahaemolyticus, and Vibrio vulnificus) were developed and optimized. The primer sets were validated for specificity and selectivity through singleplex PCR, confirming the expected amplicon size. Crosscheck and multiplex PCR showed no interference in the primer set or pathogenic DNA mixture. The NGS panel analysis of spiked water samples detected all 18 target genes in a single reaction, with pathogen concentrations ranging from 108 to 105 colony-forming units (CFUs) per target pathogen. Notably, the total sequence read counts from the virulence factor genes showed a positive association with the CFUs per target pathogen. However, the method exhibited relatively low sensitivity and occasional false positive results at low pathogen concentrations of 105 CFUs. To validate the detection and identification results, two sets of quantitative real-time PCR (qPCR) analyses were independently performed on the same spiked water samples, yielding almost the same efficiency and specificity compared to the NGS panel analysis. Comparative statistical analysis and Spearman correlation analysis further supported the similarity of the results by showing a negative association between the NGS panel sequence read counts and qPCR cycle threshold (Ct) values. To enhance NGS panel analysis for better detection, optimization of primer sets and real-time NGS sequencing technology are essential. Nonetheless, this study provides valuable insights into applying NGS panel analysis for multiple foodborne pathogen detection, emphasizing its potential in ensuring food safety.

Development and Evaluation of a Next-Generation Sequencing Panel for the Multiple Detection and Identification of Pathogens in Fermented Foods.

These days, bacterial detection methods have some limitations in sensitivity, specificity, and multiple detection. To overcome these, novel detection and identification method is necessary to be developed. Recently, NGS panel method has been suggested to screen, detect, and even identify specific foodborne pathogens in one reaction. In this study, new NGS panel primer sets were developed to target 13 specific virulence factor genes from five types of pathogenic Escherichia coli, Listeria monocytogenes, and Salmonella enterica serovar Typhimurium, respectively. Evaluation of the primer sets using singleplex PCR, crosscheck PCR and multiplex PCR revealed high specificity and selectivity without interference of primers or genomic DNAs. Subsequent NGS panel analysis with six artificially contaminated food samples using those primer sets showed that all target genes were multi-detected in one reaction at 108-105 CFU of target strains. However, a few false-positive results were shown at 106-105 CFU. To validate this NGS panel analysis, three sets of qPCR analyses were independently performed with the same contaminated food samples, showing the similar specificity and selectivity for detection and identification. While this NGS panel still has some issues for detection and identification of specific foodborne pathogens, it has much more advantages, especially multiple detection and identification in one reaction, and it could be improved by further optimized NGS panel primer sets and even by application of a new real-time NGS sequencing technology. Therefore, this study suggests the efficiency and usability of NGS panel for rapid determination of origin strain in various foodborne outbreaks in one reaction.

A Deep Learning Algorithm for Classifying Diabetic Retinopathy Using Optical Coherence Tomography Angiography.

Purpose: To develop an automated diabetic retinopathy (DR) staging system using optical coherence tomography angiography (OCTA) images with a convolutional neural network (CNN) and to verify the feasibility of the system. Methods: In this retrospective cross-sectional study, a total of 918 data sets of 3 × 3 mm2 OCTA images and 917 data sets of 6 × 6 mm2 OCTA images were obtained from 1118 eyes. A deep CNN and four traditional machine learning models were trained with annotations made by a retinal specialist based on ultra-widefield fluorescein angiography. Separately, the same images of the test data sets were independently graded by two human experts. The results of the CNN algorithm were compared with those of traditional machine learning-based classifiers and human experts. Results: The proposed CNN achieved an accuracy of 0.728, a sensitivity of 0.675, a specificity of 0.944, an F1 score of 0.683, and a quadratic weighted κ of 0.908 for a six-level staging task, which were far superior to the results of traditional machine learning methods or human experts. The CNN algorithm showed a better performance using 6 × 6 mm2 rather than 3 × 3 mm2 sized OCTA images and using combined data rather than a separate OCTA layer alone. Conclusions: CNN-based classification using OCTA images can provide reliable assistance to clinicians for DR classification. Translational Relevance: This CNN algorithm can guide the clinical decision for invasive angiography or referrals to ophthalmology specialists, helping to create more efficient diagnostic workflow in primary care settings.

A deep learning model for identifying diabetic retinopathy using optical coherence tomography angiography.

As the prevalence of diabetes increases, millions of people need to be screened for diabetic retinopathy (DR). Remarkable advances in technology have made it possible to use artificial intelligence to screen DR from retinal images with high accuracy and reliability, resulting in reducing human labor by processing large amounts of data in a shorter time. We developed a fully automated classification algorithm to diagnose DR and identify referable status using optical coherence tomography angiography (OCTA) images with convolutional neural network (CNN) model and verified its feasibility by comparing its performance with that of conventional machine learning model. Ground truths for classifications were made based on ultra-widefield fluorescein angiography to increase the accuracy of data annotation. The proposed CNN classifier achieved an accuracy of 91-98%, a sensitivity of 86-97%, a specificity of 94-99%, and an area under the curve of 0.919-0.976. In the external validation, overall similar performances were also achieved. The results were similar regardless of the size and depth of the OCTA images, indicating that DR could be satisfactorily classified even with images comprising narrow area of the macular region and a single image slab of retina. The CNN-based classification using OCTA is expected to create a novel diagnostic workflow for DR detection and referral.

Prospective trial of treat-and-extend regimen with aflibercept for branch retinal vein occlusion: 1-year results of the PLATON trial.

PURPOSE: To evaluate the functional and anatomical outcomes of a treat-and-extend (TAE) regimen with aflibercept for treatment-naive macular edema (ME) secondary to branch retinal vein occlusion (BRVO). METHODS: This was a prospective, multicenter, noncomparative, open-label clinical trial. Forty-eight eyes of 48 patients received three monthly intravitreal aflibercept injections prior to the TAE regimen. However, if the best-corrected visual acuity (BCVA) was ≥ 20/20 and the central macular thickness (CMT) was < 250 µm during the loading phase, the patient immediately proceeded to the TAE regimen. The treatment interval was adjusted by 4 weeks based on changes in CMT. The primary outcome was the mean change in BCVA from baseline to 52 weeks. RESULTS: The mean change in BCVA was 23.6 ± 14.2 letters. The proportion of patients with BCVA gain ≥ 15 letters was 77.1% at 24 weeks and 72.9% at 52 weeks. The mean reduction in CMT was 326.2 ± 235.6 µm at 24 weeks and 324.2 ± 238.0 µm at 52 weeks. The mean number of injections was 6.7 ± 1.2 (range: 6-11, all patients received three monthly intravitreal aflibercept injections) over 52 weeks, and 34 patients (70.8%) reached the maximal extension interval of 16 weeks at 52 weeks. CONCLUSIONS: The TAE regimen using aflibercept for ME secondary to BRVO, which has a treatment interval of up to 16 weeks, showed comparable efficacy to the fixed-dosing regimen along with reduced treatment burden.

Macular Microvascular Changes and Their Correlation With Peripheral Nonperfusion in Branch Retinal Vein Occlusion.

PURPOSE: To investigate the correlation between macular microvascular alterations on optical coherence tomography angiography (OCTA) and retinal ischemia on ultra-widefield fluorescein angiography (UWF FA) in eyes with branch retinal vein occlusion (BRVO). DESIGN: Cross-sectional study. METHODS: This prospective study was performed from September 2019 to June 2020 at Yeungnam University Medical Center. We included 60 patients with treatment-naïve BRVO. Two independent, masked graders analyzed OCTA parameters, including vessel density, skeletal density, and fractal dimension (FD), and UWF FA parameters, including retinal nonperfusion area (NPA) and ischemic index (ISI), from various concentric regions (perimacular region, 0.5-3 mm radius; near-peripheral region, 3-10 mm; midperipheral region, 10-15 mm; far-peripheral region, >15 mm). A repeated-measures analysis of variance test and a paired t test were performed for inter-visit and inter-regional comparisons, and Pearson correlation coefficient and multivariate regression analyses were performed to examine the correlation between UWF FA and OCTA parameters. RESULTS: The OCTA parameters from both the superficial and deep capillary plexuses (DCP) were significantly correlated with NPA and ISI in all concentric regions. Even after adjusting for several covariates, all OCTA parameters revealed a significant association with ISI on UWF FA. Moreover, OCTA parameters from DCP were significantly correlated with concentrations of placental growth factor and vascular endothelial growth factor. Although all OCTA parameters achieved excellent results of area under the curve (AUC) > 0.9 for detecting severe retinal ischemia, defined as ISI >10%, FD reduction in DCP was the most reliable parameter (AUC = 0.948, P < .001), and 5.39% was the best cut-off point for predicting ISI > 10%. CONCLUSIONS: OCTA is a useful noninvasive tool not only for evaluation of macular microvasculature but for supposition of peripheral nonperfusion in eyes with BRVO.

Optimal fluence rate of photodynamic therapy for chronic central serous chorioretinopathy.

AIMS: To investigate the lowest effective fluence rate of photodynamic therapy (PDT) for treating chronic central serous chorioretinopathy (CSC). METHODS: Fifty-one eyes of 51 patients with chronic CSC were randomly treated with 30% (n=15), 40% (n=16) or 50% (n=17) of the standard-fluence rate of PDT and followed up for 12 months. The success rate, recurrence rate, mean best-corrected visual acuity (BCVA), central foveal thickness (CFT), subfoveal choroidal thickness (SFCT), integrity of the outer retinal layer and complications were evaluated at baseline and at the follow-up periods after PDT. RESULTS: The rate of complete subretinal fluid (SRF) resolution in the 30%-fluence, 40%-fluence and 50%-fluence groups was 60.0%, 81.2% and 100.0%, respectively, at 3 months (p=0.009), and 80.0%, 94.0% and 100.0%, respectively, at 12 months (p=0.06). The recurrence rate in the 50%-fluence group was lower than that in the 30%- and 40%-fluence groups at 12 months (30% vs 50%, 40% vs 50%; p=0.002, p=0.030, respectively (log-rank test)). The mean BCVA improved significantly 12 months after PDT only in the 40%- and 50%-fluence groups (p=0.005, p=0.003, respectively). Mean CFT and SFCT decreased significantly at 12 months in the three groups. The rate of complications did not differ significantly among the three groups. CONCLUSIONS: A 50%-fluence rate of PDT seems to be the most effective for treating chronic CSC, considering the low recurrence rate and high rate of complete SRF resolution, compared with other low-fluence PDT. TRIAL REGISTRATION NUMBER: NCT01630863.

Management of a Submacular Hemorrhage Secondary to Age-Related Macular Degeneration: A Comparison of Three Treatment Modalities.

This paper aims to compare the effects of three treatment modalities for a submacular hemorrhage (SMH) secondary to exudative age-related macular degeneration (AMD). Seventy-seven patients with an SMH were divided into three groups: small-sized (optic disc diameter (ODD) ≥ 1 to < 4), medium-sized (ODD ≥ 4 within the temporal arcade) and large-sized (ODD ≥ 4, exceeding the temporal arcade). Patients received anti-vascular endothelial growth factor (anti-VEGF) monotherapy, pneumatic displacement (PD) with anti-VEGF or a vitrectomy with a subretinal tissue plasminogen activator (tPA) and gas tamponade based on the surgeon's discretion. The functional and anatomical outcomes were evaluated. Among the 77 eyes, 45 eyes had a small-sized, 21 eyes had a medium-sized and 11 eyes had a large-sized SMH. In the small-sized group, all treatment modalities showed a gradual best-corrected visual acuity (BCVA) improvement with high hemorrhagic regression or displacement rates (over 75%). In the medium-sized group, PD and surgery were associated with better BCVA with more displacement than anti-VEGF monotherapy (67% and 83%, respectively, vs. 33%). In the large-sized group, surgery showed a better visual improvement with a higher displacement rate than PD (86% vs. 25%). Our findings demonstrated that visual improvement can be expected through appropriate treatment strategy regardless of the SMH size. In cases with a larger SMH, invasive techniques including PD or surgery were more advantageous than anti-VEGF monotherapy.

Major constituents and antimicrobial activity of Korean herb Acorus calamus.

The constituents and antimicrobial activity of the essential oil from Acorus calamus were analysed. Methyl isoeugenol and cyclohexanone were identified as the major constituents of the essential oil. The essential oil was tested for antimicrobial activity against bacteria and yeast, and has shown strong antibiotic activities against most of the tested microbes, except Escherichia coli. The hexane extract has shown a similar pattern of antimicrobial activity as the essential oil. Methyl isoeugenol, the most abundant constituent in the essential oil, has also shown similar antimicrobial activity, except against Bacillus subtilis. The essential oil as well as the hexane extract and methyl isoeugenol have shown antimicrobial activity against Propionibacterium acne, which is known to be involved in acne vulgaris.