Deep-Learning for the Diagnosis of Esophageal Cancers and Precursor Lesions in Endoscopic Images: A Model Establishment and Nationwide Multicenter Performance Verification Study.

BACKGROUND: Suspicion of lesions and prediction of the histology of esophageal cancers or premalignant lesions in endoscopic images are not yet accurate. The local feature selection and optimization functions of the model enabled an accurate analysis of images in deep learning. OBJECTIVES: To establish a deep-learning model to diagnose esophageal cancers, precursor lesions, and non-neoplasms using endoscopic images. Additionally, a nationwide prospective multicenter performance verification was conducted to confirm the possibility of real-clinic application. METHODS: A total of 5162 white-light endoscopic images were used for the training and internal test of the model classifying esophageal cancers, dysplasias, and non-neoplasms. A no-code deep-learning tool was used for the establishment of the deep-learning model. Prospective multicenter external tests using 836 novel images from five hospitals were conducted. The primary performance metric was the external-test accuracy. An attention map was generated and analyzed to gain the explainability. RESULTS: The established model reached 95.6% (95% confidence interval: 94.2-97.0%) internal-test accuracy (precision: 78.0%, recall: 93.9%, F1 score: 85.2%). Regarding the external tests, the accuracy ranged from 90.0% to 95.8% (overall accuracy: 93.9%). There was no statistical difference in the number of correctly identified the region of interest for the external tests between the expert endoscopist and the established model using attention map analysis (P = 0.11). In terms of the dysplasia subgroup, the number of correctly identified regions of interest was higher in the deep-learning model than in the endoscopist group, although statistically insignificant (P = 0.48). CONCLUSIONS: We established a deep-learning model that accurately classifies esophageal cancers, precursor lesions, and non-neoplasms. This model confirmed the potential for generalizability through multicenter external tests and explainability through the attention map analysis.

Construction of a bifunctional enzyme fusion for the combined determination of biogenic amines in foods.

Biogenic amines (BAs) are a group of low-molecular-mass organic bases derived from free amino acids. Due to the undesirable effects of BAs on human health, amine oxidase-based detection methods for BAs in foods have been developed. Here, we developed a bifunctional enzyme fusion (MAPO) using a Cu(2+)-containing monoamine oxidase (AMAO2) and a flavin adenine dinucleotide-containing putrescine oxidase (APUO) from Arthrobacter aurescens. It was necessary to activate MAPO with supplementary Cu(2+) ions, leading to a 6- to 12-fold improvement in catalytic efficiency (kcat/KM) for monoamines. The optimal temperatures of Cu(2+)-activated MAPO (cMAPO) for both tyramine and putrescine were 50 °C, and the optimal pH values for tyramine and putrescine were pH 7.0 and pH 8.0, respectively, consistent with those of AMAO2 and APUO, respectively. The cMAPO showed relative specific activities of 100, 99, 32, and 32 for 2-phenylethylamine, tyramine, histamine, and putrescine, respectively. The tyramine-equivalent BA contents of fermented soybean pastes by cMAPO were more than 90% of the total BA determined by HPLC. In conclusion, cMAPO is fully bifunctional toward biogenic monoamines and putrescine, allowing the combined determination of multiple BAs in foods. This colorimetric determination method could be useful for point-of-care testing to screen safety-guaranteed products prior to instrumental analyses.

Characterization of amine oxidases from Arthrobacter aurescens and application for determination of biogenic amines.

Biogenic amines (BAs) that are produced through naturally occurring decarboxylation of amino acids have toxicological effects on humans. Bacterial amine oxidases are useful tools for the rapid quantification of BAs in foods. To develop amine oxidases for the rapid detection of BAs, the genes for amine oxidases from Arthobacter aurescens TC-1, designated AMAO1, AMAO2, and AMAO3, respectively, were cloned and expressed in Escherichia coli. AMAO1 was catalytically inactive to BAs, and AMAO3 showed a narrow substrate spectrum. In contrast, AMAO2 exhibited activity with relative k cat/K M values of 100:49.6:7.6 for 2-phenylethylamine, tyramine, and histamine, respectively. AMAO2 also utilized putrescine and spermidine as substrates, with four or five orders of magnitude lower k cat/K M values than that of 2-phenylethylamine. AMAO2 and AMAO3 were seriously affected by substrate inhibition. Using BA mixtures (consisting of 2-phenylethylamine, tyramine, and histamine) as samples, the detection range of the enzymatic analysis of BA using AMAO2 was determined to be 2.5-120 µM, and its detection limit was 2.3 µM. Analysis of five commercial cheese products revealed that the BA contents determined by the enzymatic methods showed a good agreement with the sum of three monoamines and histamine by HPLC. Therefore, the enzymatic assay using AMAO2 can be used in quality control of food products through rapid, sensitive, and preliminary estimation of major BAs including the most important TyrN and HisN in foods.