Dual Channel H2O2 Photosynthesis in Pure Water over S-Scheme Heterojunction Cs3PMo12/CC Boosted by Proton and Electron Reservoirs.

Dual channel photo-driven H2O2 production in pure water on small-scale on-site setups is a promising strategy to provide low-concentrated H2O2 whenever needed. This process suffers, however, strongly from the fast recombination of photo-generated charge carriers and the sluggish oxidation process. Here, insoluble Keggin-type cesium phosphomolybdate Cs3PMo12O40 (abbreviated to Cs3PMo12) is introduced to carbonized cellulose (CC) to construct S-scheme heterojunction Cs3PMo12/CC. Dual channel H2O2 photosynthesis from both H2O oxidation and O2 reduction in pure water has been thus achieved with the production rate of 20.1 mmol L-1 gcat. -1 h-1, apparent quantum yield (AQY) of 2.1% and solar-to-chemical conversion (SCC) efficiency of 0.050%. H2O2 accumulative concentration reaches 4.9 mmol L-1. This high photocatalytic activity is guaranteed by unique features of Cs3PMo12/CC, namely, S-scheme heterojunction, electron reservoir, and proton reservoir. The former two enhance the separation of photo-generated charge carriers, while the latter speeds up the torpid oxidation process. In situ experiments reveal that H2O2 is formed via successive single-electron transfer in both channels. In real practice, exposing the reaction system under natural sunlight outdoors successfully results in 0.24 mmol L-1 H2O2. This work provides a key practical strategy for designing photocatalysts in modulating redox half-reactions in photosynthesis.

Cascade Electrocatalytic Reduction of Nitrate to Ammonia Using a Heterobimetallic Covalent Organic Framework Composed of Cu-Porphyrin and Co-Bipyridine.

The electrocatalytic reduction of nitrate (NO3-) to ammonia (NH3) not only offers an effective solution to environmental problems caused by the accumulation of NO3- but also provides a sustainable alternative to the Haber-Bosch process. However, the conversion of NO3- to NH3 is a complicated process involving multiple steps, leading to a low Faradaic efficiency (FE) for NH3 production. The structural designability of covalent organic frameworks (COFs) renders feasible and precise modulation at the molecular level, facilitating the incorporation of multiple well-defined catalytic sites with different reactivities into a cohesive entity. This promotes the efficiency of the overall reaction through the coupling of multistep reactions. Herein, heterobimetallic CuP-CoBpy was prepared by postmodification, involving the anchoring of cobalt ions to the CuP-Bpy structure. As a result of the cascade effect of the bimetallic sites, CuP-CoBpy achieved an outstanding NH3 yield of 13.9 mg h-1 mgcat.-1 with a high FE of 96.7% at -0.70 V versus the reversible hydrogen electrode and exhibited excellent stability during catalysis. A series of experimental and theoretical studies revealed that the CuP unit facilitates the conversion of NO3- to NO2-, while the CoBpy moiety significantly prompts the reduction of NO2- to NH3. This study demonstrates that tailoring the structural units for the construction of COFs based on each step in the multistep reaction can enhance both the catalytic activity and product selectivity of the overall process.

A Novel Central Camera Calibration Method Recording Point-to-Point Distortion for Vision-Based Human Activity Recognition.

The camera is the main sensor of vison-based human activity recognition, and its high-precision calibration of distortion is an important prerequisite of the task. Current studies have shown that multi-parameter model methods achieve higher accuracy than traditional methods in the process of camera calibration. However, these methods need hundreds or even thousands of images to optimize the camera model, which limits their practical use. Here, we propose a novel point-to-point camera distortion calibration method that requires only dozens of images to get a dense distortion rectification map. We have designed an objective function based on deformation between the original images and the projection of reference images, which can eliminate the effect of distortion when optimizing camera parameters. Dense features between the original images and the projection of the reference images are calculated by digital image correlation (DIC). Experiments indicate that our method obtains a comparable result with the multi-parameter model method using a large number of pictures, and contributes a 28.5% improvement to the reprojection error over the polynomial distortion model.

Research on Fault Diagnosis of Rolling Bearings Based on Variational Mode Decomposition Improved by the Niche Genetic Algorithm.

Due to the influence of signal-to-noise ratio in the early failure stage of rolling bearings in rotating machinery, it is difficult to effectively extract feature information. Variational Mode Decomposition (VMD) has been widely used to decompose vibration signals which can reflect more fault omens. In order to improve the efficiency and accuracy, a method to optimize VMD by using the Niche Genetic Algorithm (NGA) is proposed in this paper. In this method, the optimal Shannon entropy of modal components in a VMD algorithm is taken as the optimization objective, by using the NGA to constantly update and optimize the combination of influencing parameters composed of α and K so as to minimize the local minimum entropy. According to the obtained optimization results, the optimal input parameters of the VMD algorithm were set. The method mentioned is applied to the fault extraction of a simulated signal and a measured signal of a rolling bearing. The decomposition process of the rolling-bearing fault signal was transferred to the variational frame by the NGA-VMD algorithm, and several eigenmode function components were obtained. The energy feature extracted from the modal component containing the main fault information was used as the input vector of a particle swarm optimized support vector machine (PSO-SVM) and used to identify the fault type of the rolling bearing. The analysis results of the simulation signal and measured signal show that: the NGA-VMD algorithm can decompose the vibration signal of a rolling bearing accurately and has a better robust performance and correct recognition rate than the VMD algorithm. It can highlight the local characteristics of the original sample data and reduce the interference of the parameters selected artificially in the VMD algorithm on the processing results, improving the fault-diagnosis efficiency of rolling bearings.

A randomized, controlled comparison of a stannous-containing dentifrice for reducing gingival bleeding and balancing the oral microbiome relative to a positive control.

PURPOSE: To evaluate the effect of a stannous-containing fluoride dentifrice on gingival health and on the composition of the oral microbiome versus a positive control dentifrice over 2 weeks, in a population of healthy adults with self-reported sub-optimal oral health at baseline. METHODS: This was a randomized, controlled, double-blind clinical study. 87 subjects with self-reported sub-optimal oral health at enrollment were randomized to brush twice daily with either an experimental dentifrice (n= 43) or a marketed positive control dentifrice (n= 43), both containing stannous chloride and 0.321% sodium fluoride. All subjects used a soft, manual toothbrush that was provided. The Mazza modification of gingival papillary bleeding Index (Mazza GI) was used to assess gingivitis at baseline and at Week 2. Supragingival plaque was collected for microbiome composition analyses at baseline, Week 1, and Week 2. RESULTS: 83 subjects completed the study. Baseline means were balanced between the treatment groups (P> 0.34). At Week 2, the positive control dentifrice demonstrated a 63.8% statistically significant (P< 0.0001) reduction relative to baseline for Mazza number of gingival bleeding sites. The experimental stannous containing dentifrice provided a comparable 63.5% gingival bleeding reduction versus baseline. There was no significant (P= 0.96) difference between the two dentifrices for either Mazza GI score or number of bleeding sites measured. The microbiome composition analysis at Week 1 found that 28 gingivitis-associated bacterial genera, including Porphyromonas, Tannerella, and Fusobacterium, were significantly inhibited in both dentifrice groups when compared to baseline, while the relative abundance of genera associated with oral health, such as Rothia, Streptococcus, Haemophilus, and Lautropia, was significantly elevated after treatment. These improvements in the oral ecosystem were sustained at Week 2. CLINICAL SIGNIFICANCE: An experimental stannous-containing sodium fluoride dentifrice significantly reduced gingival bleeding comparable to a positive control, and both dentifrices promoted a shift in the oral microbiome towards those genera associated with oral health in a subject population with self-reported sub-optimal oral health at baseline.

Preparation and Post-Assembly Modification of Metallosupramolecular Assemblies from Poly( N-Heterocyclic Carbene) Ligands.

Although the coordination chemistry of N-heterocyclic carbenes (NHCs) with transition metals has been explored for half a century, only in the past ten years has the chemistry of metallosupramolecular assemblies based on poly-NHC ligands been studied more extensively. Remarkable discrete assemblies featuring poly-NHC ligands including two-dimensional metallacycles and three-dimensional metallaprisms/cages have since emerged. These assemblies are mostly obtained starting from various imidazolium or benzimidazolium salts. Driven by the increasing interest in new supramolecular architectures from carbon donor ligands, design, and construction of poly-NHC metal assemblies has become a rapidly growing area of research. The metal-carbene bond length is fixed to approximately 2.0 Å in linear NHC-M-NHC complexes. This allows the use of such complexes bearing olefin-substituted NHC ligands as templates for subsequent photochemical [2 + 2] cycloaddition reactions. The postassembly modification of such assemblies has been actively explored in recent years. In this review, we focus on the synthetic methods, characterization, structural features, and postassembly modifications of metallosupramolecular assemblies obtained from poly-NHC ligands.

Skin benefits of moisturising body wash formulas for children with atopic dermatitis: A randomised controlled clinical study in China.

BACKGROUND: It acknowledged that skin care is an important part of atopic dermatitis therapy. However, clinical evidences are limited for the best bathing practices, especially the skin health performance of cleansing products on children's atopic dermatitis skin. METHODS: A randomised controlled clinical study was conducted in China among 4- to 18-year-old children with mild-to-moderate atopic dermatitis to evaluate the skin health effect of three cleansing systems (a mild synthetic bar, an ultra-mild body wash with lipids, and an ultra-mild body wash with lipids and zinc pyrithione) by measuring SCORing of Atopic Dermatitis (SCORAD), consumption of topical corticosteroid and the characteristics of microbiome. RESULTS: Increased Staphylococcus aureus abundance and decreased microbial diversity were observed in atopic dermatitis lesion sites compared with healthy control sites. After 4 weeks of treatment, all three treatments showed clinically important improvement from baseline in SCORAD. Four-week corticosteroid consumption was significantly lower for the two body wash groups than the bar group. A significant decrease in S. aureus abundance and increase in microbial diversity were observed in the lesion sites for the two body wash formulas, while the microbial diversity was statistically insignificant for the mild cleansing bar group. However, there were no incremental benefits provided by the body wash formulas based on the assessment of SCORAD. CONCLUSIONS: These results demonstrated the safety and efficacy of using the investigational body wash formulas with lipids in reducing the needs for corticosteroid and improving the healthy composition of skin microbiome vs. the mild synthetic bar soap.

Genome sequencing and heterologous expression of antiporters reveal alkaline response mechanisms of Halomonas alkalicola.

Halomonas alkalicola CICC 11012s is an alkaliphilic and halotolerant bacterium isolated from a soap-making tank (pH > 10) from a household-product plant. This strain can propagate at pH 12.5, which is fatal to most bacteria. Genomic analysis revealed that the genome size was 3,511,738 bp and contained 3295 protein-coding genes, including a complete cell wall and plasma membrane lipid biosynthesis pathway. Furthermore, four putative Na+/H+ and K+/H+ antiporter genes, or gene clusters, designated as HaNhaD, HaNhaP, HaMrp and HaPha, were identified within the genome. Heterologous expression of these genes in antiporter-deficient Escherichia coli indicated that HaNhaD, an Na+/H+ antiporter, played a dominant role in Na+ tolerance and pH homeostasis in acidic, neutral and alkaline environments. In addition, HaMrp exhibited Na+ tolerance; however, it functioned mainly in alkaline conditions. Both HaNhaP and HaPha were identified as K+/H+ antiporters that played an important role in high alkalinity and salinity. In summary, genome analysis and heterologous expression experiments demonstrated that a complete set of adaptive strategies have been developed by the double extremophilic strain CICC 11012s in response to alkalinity and salinity. Specifically, four antiporters exhibiting different physiological roles for different situations worked together to support the strain in harsh surroundings.

[Analysis on SSR information in transcriptome of Astragalus membranaceus var. mongholicus and its polymorphism].

In order to enrich the library of SSR and provide more powerful tools for molecular marker-assisted breeding in Astragalus membranaceus var. mongholicus, simple sequence repeats (SSR) loci in its transcriptome were searched in 18 040 unigenes (>=1 kb) by using MISA. SSR loci information was analyzed and SSR primers were designed by Primer 3. Furthermore, 110 pairs of primers were randomly selected for the polymorphic analysis on 20 plants collected from different habitats. A total of 5 640 SSRs were found in the transcriptome of A. membranaceus var. mongholicus, distributed in 4 462 unigenes with the distribution frequency of 31.26%. SSR loci occurred every 6 514 bp. Mono-nucleotide repeat was the main type, accounted for as much as 36.72% of all SSRs, followed by tri-nucleotide(32.57%) and di-nucleotide(27.73%) repeat motif. Among all 75 repeat types, A/T(2 026) was the predominant one followed by AG/CT(1 179), AAG/CTT(477). For validating the availability of the SSR primers designed using Primer 3, 110 pairs of primers were randomly selected for PCR amplification. Among them, 97 pairs of primers (88.18%) produced clear and reproductive bands. Using 19 pairs of primers showed polymorphism, 20 plants were divded into two groups by UPGMA. There are numerous SSRs in A. membranaceus var. mongholicus transcriptome with high frequency and various types, this will provide the abundant candidate molecular markers for genetic diversity, molecular identification, and marker-assisted breeding study for this plant.

Halomonas alkalicola sp. nov., isolated from a household product plant.

A Gram-stain-negative, alkaliphilic and moderately halophilic bacterium, designated 56-L4-10aEnT, was isolated from a household product plant in China. Cells of the novel isolate were rod-shaped, non-spore-forming and non-motile. Phylogenetic analysis based on 16S rRNA gene sequences showed that strain 56-L4-10aEnT belongs to the genus Halomonas, with the six closest neighbours being Halomonas mongoliensis Z-7009T (97.59 % 16S rRNA gene sequence similarity), Halomonas ventosae Al12T (97.35 %), Halomonas campaniensis 5AGT (97.22 %), Halomonas alimentaria YKJ-16T (97.22 %), Halomonas shengliensis SL014B-85T (97.12 %) and Halomonas fontilapidosi 5CRT (97.09 %). The main polar lipids of strain 56-L4-10aEnT contained diphosphatidylglycerol, phosphatidylglycerol and phosphatidylethanolamine. The predominant respiratory quinone was Q-9, with Q-8 as a minor component. The major fatty acids were C18 : 1ω7c/C18 : 1ω6c and C16 : 0. Strain 56-L4-10aEnT was clearly distinguished from the type strains mentioned above through phylogenetic analysis, DNA-DNA hybridization, fatty acid composition data and a range of physiological and biochemical characteristics comparisons. It is evident from the genotypic and phenotypic data that strain 56-L4-10aEnTcould be classified as a representative of a novel species of the genus Halomonas, for which the name Halomonas alkalicola sp. nov. is proposed. The type strain is 56-L4-10aEnT (=CICC 11012sT=DSM 103354T).

Novosphingobium clariflavum sp. nov., isolated from a household product plant.

A Gram-stain-negative, rod-shaped, bright-yellow-pigmented bacterium, designated 164T, was isolated from a used sponge for equipment cleaning at a household product plant in China. The 16S rRNA gene sequence comparisons indicated that strain 164T was most closely related to Novosphingobium panipatense DSM 22890T (98.28 % similarity) and shared sequence similarities of 97.73-98.27 % with other members of the genus Novosphingobium. In DNA-DNA hybridization studies the relatedness between strain 164T and its closest phylogenetic neighbours was <70 %, which indicated that strain 164T represented a novel species of the genus Novosphingobium. The DNA G+C content of strain 164T was 65.9 mol%. The major respiratory quinone was ubiquinone Q-10 (83.5 %) with minor amounts of Q-9 (16.5 %). The polar lipid profile included diphosphatidylglycerol, phosphatidylglycerol, phosphatidylethanolamine, phosphatidyldimethylethanolamine, sphingoglycolipid, phosphatidylcholine, unidentified aminolipids and unidentified aminophospholipids. Spermidine was the major polyamine. The major fatty acids were summed feature 8 (consisting of C18 : 1ω7c and/or C18 : 1ω6c) and C14 : 0 2-OH. The results obtained from phylogenetic analysis, DNA-DNA hybridization, and chemotaxonomic and phenotypic analysis support the conclusion that strain 164T represents a novel species of the genus Novosphingobium, for which the name Novosphingobium clariflavum sp. nov. is proposed. The type strain is 164T (=CICC 11035sT=DSM 103351T).

Natural abundance 15N†NMR by dynamic nuclear polarization: fast analysis of binding sites of a novel amine-carboxyl-linked immobilized dirhodium catalyst.

A novel heterogeneous dirhodium catalyst has been synthesized. This stable catalyst is constructed from dirhodium acetate dimer (Rh2(OAc)4) units, which are covalently linked to amine- and carboxyl-bifunctionalized mesoporous silica (SBA-15-NH2-COOH). It shows good efficiency in catalyzing the cyclopropanation reaction of styrene and ethyl diazoacetate (EDA) forming cis- and trans-1-ethoxycarbonyl-2-phenylcyclopropane. To characterize the structure of this catalyst and to confirm the successful immobilization, heteronuclear solid-state NMR experiments have been performed. The high application potential of dynamic nuclear polarization (DNP) NMR for the analysis of binding sites in this novel catalyst is demonstrated. Signal-enhanced (13)C CP MAS and (15)N CP MAS techniques have been employed to detect different carboxyl and amine binding sites in natural abundance on a fast time scale. The interpretation of the experimental chemical shift values for different binding sites has been corroborated by quantum chemical calculations on dirhodium model complexes.

Design of a Heterogeneous Catalyst Based on Cellulose Nanocrystals for Cyclopropanation: Synthesis and Solid-State NMR Characterization.

Heterogeneous dirhodium(II) catalysts based on environmentally benign and biocompatible cellulose nanocrystals (CNC-Rh2) as support material were obtained by ligand exchange between carboxyl groups on the CNC surface and Rh2(OOCCF3)4, as was confirmed by solid-state (19)F and (13)C NMR spectroscopy. On average, two CF3COO(-) groups are replaced during ligand exchange, which is consistent with quantitative analysis by a combination of (19)F NMR spectroscopy and thermogravimetry. CNC-Rh2 catalysts performed well in a model cyclopropanation reaction, in spite of the low dirhodium(II) content on the CNC surface (0.23 mmol g(-1)). The immobilization through covalent bonding combined with the separate locations of binding positions and active sites of CNC-Rh2 guarantees a high stability against leaching and allows the recovery and reuse of the catalyst during the cyclopropanation reaction.

Fine tuning dual active sites in modulating cascade electrocatalytic nitrate reduction over covalent organic framework.

Conversion of NO3- to NH3 proceeds stepwise in natural system under two different enzymes involving intermediate NO2-. Artificial electro-driven NO3- reduction also faces the obstacle of low faradaic efficiency due to insufficient utilization of this intermediate. Herein, we demonstrate a bimetallic COF-based electrocatalyst for the cascade catalysis of NO3--to-NO2--to-NH3 for the first time. TpBpy-Cu2Co4 exhibits a significantly improved performance, with an enhancement factor of 1.4-2 compared to monometallic TpBpy-M. The NH3 yield rate achieves 25.6 mg h-1 mgcat.-1 at -0.55 V vs RHE over TpBpy-Cu2Co4, together with excellent faradaic efficiency (93.4 %). This achievement demonstrates cascade catalysis between Co and Cu units, and their distinct roles are investigated through electrochemical experiments and theory calculations. In electrocatalytic process, Cu site facilities *NO3-to-*NO3H step, while the Co site significantly decreases the energy barrier of *NHOH-to-*NH. The present work provides a valuable inspiration in designing efficient catalysts for cascade reaction.

Interactions and effects of a stannous-containing sodium fluoride dentifrice on oral pathogens and the oral microbiome.

Numerous studies have investigated the effects of stannous ions on specific microbes and their efficacy in reducing dental plaque. Nonetheless, our understanding of their impact on the oral microbiome is still a subject of ongoing exploration. Therefore, this study sought to evaluate the effects of a stannous-containing sodium fluoride dentifrice in comparison to a zinc-containing sodium fluoride dentifrice and a control group on intact, healthy oral biofilms. Utilizing the novel 2bRAD-M approach for species-resolved metagenomics, and FISH/CLSM with probes targeting periodontal and caries associated species alongside Sn2+ and Zn2+ ions, we collected and analyzed in situ biofilms from 15 generally healthy individuals with measurable dental plaque and treated the biofilms with dentifrices to elucidate variations in microbial distribution. Although significant shifts in the microbiome upon treatment were not observed, the use of a stannous-containing sodium fluoride dentifrice primarily led to an increase in health-associated commensal species and decrease in pathogenic species. Notably, FISH/CLSM analysis highlighted a marked reduction in representative species associated with periodontitis and caries following treatment with the use of a stannous-containing sodium fluoride dentifrice, as opposed to a zinc-containing sodium fluoride dentifrice and the control group. Additionally, Sn2+ specific intracellular imaging reflected the colocalization of Sn2+ ions with P. gingivalis but not with other species. In contrast, Zn2+ ions exhibited non-specific binding, thus suggesting that Sn2+ could exhibit selective binding toward pathogenic species. Altogether, our results demonstrate that stannous ions could help to maintain a healthy oral microbiome by preferentially targeting certain pathogenic bacteria to reverse dysbiosis and underscores the importance of the continual usage of such products as a preventive measure for oral diseases and the maintenance of health.

SimCol3D - 3D reconstruction during colonoscopy challenge.

Colorectal cancer is one of the most common cancers in the world. While colonoscopy is an effective screening technique, navigating an endoscope through the colon to detect polyps is challenging. A 3D map of the observed surfaces could enhance the identification of unscreened colon tissue and serve as a training platform. However, reconstructing the colon from video footage remains difficult. Learning-based approaches hold promise as robust alternatives, but necessitate extensive datasets. Establishing a benchmark dataset, the 2022 EndoVis sub-challenge SimCol3D aimed to facilitate data-driven depth and pose prediction during colonoscopy. The challenge was hosted as part of MICCAI 2022 in Singapore. Six teams from around the world and representatives from academia and industry participated in the three sub-challenges: synthetic depth prediction, synthetic pose prediction, and real pose prediction. This paper describes the challenge, the submitted methods, and their results. We show that depth prediction from synthetic colonoscopy images is robustly solvable, while pose estimation remains an open research question.

Effects of chloroform extract of Dryopteris crassirhizoma on the ultramicroscopic structures of Meloidogyne incognita.

In our early experiments, the chloroform extract of D. crassirhizoma was demonstrated to contain the highest concentrations of total phloroglucinols among several extract fractions and possessed the most effective nematicidal activity. This study aimed to ascertain the ultrastructural changes in M. incognita after treatment with a D. crassirhizoma chloroform extract at 1 mg·mL⁻¹ for 24 h. It was found that the extract exhibited significant destructive effects on the worm's ultrastructure and caused distinctive damage to body surfaces and internal structures. These results will contribute to a deeper understanding of the nematicidal mechanism of D. crassirhizoma, as well as in the design of efficient bionematicides.

Facile and controllable preparation of carbon microsphere for electro-driven nitrogen reduction: Accommodating nitrogen doping with hierarchical porous structure.

Driven by sustainable electricity, electrochemical nitrogen fixation under ambient conditions is considered as a promising strategy to generate low-concentrated NH3/NH4+. Under the principle of doping and porous engineering, nitrogen-doped carbon microsphere with hierarchical pores (NC-HP) is fabricated via pyrolyzing polymer microsphere. Hierarchical structure with macro-, meso- and micropores is obtained by assembling melamine/phenol-formaldehyde oligomers in Pickering droplets, with the assistance of triblock copolymer Pluronic F127. The regularity of mesopores is strongly affected by melamine to phenol mass ratio. For NC-HP, nitrogen content (N-content) in the carbon matrix can reach as high as 19.1 wt%, yet trade-off effect is observed between N-content and regularity of mesopores. As consequence, NC-HP-3 with N-content of 15.6 wt% and distinct mesopores exhibits the highest catalytic performance. At -0.5 V vs. RHE, NH3/NH4+ production rate and Faradaic efficiency (FE) value reach 15.6 µg∙mgcat.-1∙h-1 and 15.5%, respectively. It shows excellent recyclability, and no degradations are observed with respect to morphology and porous structure. In this hierarchical porous structure, mesopores are expected to facilitate mass transfer for both electrolyte ions and nitrogen, and hence catalytic active sites (e.g. pyrrolic- and pyridinic-N species) in hierarchically mutually connected pores can be well utilized.

Impact of antibacterial detergent on used-towel microbiomes at species-level and its effect on malodor control.

The impact of antibacterial detergent on microbial exchanges and its subsequent effect on malodor in used towels were examined. Homogenization of microbiome among postwashed and indoor dried towels that was dominated by known malodor-producing bacteria. The microbial exchange was attenuated, and the abundance of malodor-producing bacteria was reduced in towels laundered with antibacterial detergent. Reduction of malodorous volatile organic compounds produced from towels laundered with antibacterial detergent.

Real-Time Evaluation of Helicobacter pylori Infection by Convolution Neural Network During White-Light Endoscopy: A Prospective, Multicenter Study (With Video).

INTRODUCTION: Convolutional neural network during endoscopy may facilitate evaluation of Helicobacter pylori infection without obtaining gastric biopsies. The aim of the study was to evaluate the diagnosis accuracy of a computer-aided decision support system for H. pylori infection (CADSS-HP) based on convolutional neural network under white-light endoscopy. METHODS: Archived video recordings of upper endoscopy with white-light examinations performed at Sir Run Run Shaw Hospital (January 2019-September 2020) were used to develop CADSS-HP. Patients receiving endoscopy were prospectively enrolled (August 2021-August 2022) from 3 centers to calculate the diagnostic property. Accuracy of CADSS-HP for H. pylori infection was also compared with endoscopic impression, urea breath test (URT), and histopathology. H. pylori infection was defined by positive test on histopathology and/or URT. RESULTS: Video recordings of 599 patients who received endoscopy were used to develop CADSS-HP. Subsequently, 456 patients participated in the prospective evaluation including 189 (41.4%) with H. pylori infection. With a threshold of 0.5, CADSS-HP achieved an area under the curve of 0.95 (95% confidence interval [CI], 0.93-0.97) with sensitivity and specificity of 91.5% (95% CI 86.4%-94.9%) and 88.8% (95% CI 84.2%-92.2%), respectively. CADSS-HP demonstrated higher sensitivity (91.5% vs 78.3%; mean difference = 13.2%, 95% CI 5.7%-20.7%) and accuracy (89.9% vs 83.8%, mean difference = 6.1%, 95% CI 1.6%-10.7%) compared with endoscopic diagnosis by endoscopists. Sensitivity of CADSS-HP in diagnosing H. pylori was comparable with URT (91.5% vs 95.2%; mean difference = 3.7%, 95% CI -1.8% to 9.4%), better than histopathology (91.5% vs 82.0%; mean difference = 9.5%, 95% CI 2.3%-16.8%). DISCUSSION: CADSS-HP achieved high sensitivity in the diagnosis of H. pylori infection in the real-time test, outperforming endoscopic diagnosis by endoscopists and comparable with URT. Clinicaltrials.gov ; ChiCTR2000030724.