Using deep learning and explainable artificial intelligence to assess the severity of gastroesophageal reflux disease according to the Los Angeles Classification System.

OBJECTIVES: Gastroesophageal reflux disease (GERD) is a complex disease with a high worldwide prevalence. The Los Angeles classification (LA-grade) system is meaningful for assessing the endoscopic severity of GERD. Deep learning (DL) methods have been widely used in the field of endoscopy. However, few DL-assisted researches have concentrated on the diagnosis of GERD. This study is the first to develop a five-category classification DL model based on the LA-grade using explainable artificial intelligence (XAI). MATERIALS AND METHODS: A total of 2081 endoscopic images were used for the development of a DL model, and the classification accuracy of the models and endoscopists with different levels of experience was compared. RESULTS: Some mainstream DL models were utilized, of which DenseNet-121 outperformed. The area under the curve (AUC) of the DenseNet-121 was 0.968, and its classification accuracy (86.7%) was significantly higher than that of junior (71.5%) and experienced (77.4%) endoscopists. An XAI evaluation was also performed to explore the perception consistency between the DL model and endoscopists, which showed meaningful results for real-world applications. CONCLUSIONS: The DL model showed a potential in improving the accuracy of endoscopists in LA-grading of GERD, and it has noticeable clinical application prospects and is worthy of further promotion.

Detecting the Critical States of Type 2 Diabetes Mellitus Based on Degree Matrix Network Entropy by Cross-Tissue Analysis.

Type 2 diabetes mellitus (T2DM) is a metabolic disease caused by multiple etiologies, the development of which can be divided into three states: normal state, critical state/pre-disease state, and disease state. To avoid irreversible development, it is important to detect the early warning signals before the onset of T2DM. However, detecting critical states of complex diseases based on high-throughput and strongly noisy data remains a challenging task. In this study, we developed a new method, i.e., degree matrix network entropy (DMNE), to detect the critical states of T2DM based on a sample-specific network (SSN). By applying the method to the datasets of three different tissues for experiments involving T2DM in rats, the critical states were detected, and the dynamic network biomarkers (DNBs) were successfully identified. Specifically, for liver and muscle, the critical transitions occur at 4 and 16 weeks. For adipose, the critical transition is at 8 weeks. In addition, we found some "dark genes" that did not exhibit differential expression but displayed sensitivity in terms of their DMNE score, which is closely related to the progression of T2DM. The information uncovered in our study not only provides further evidence regarding the molecular mechanisms of T2DM but may also assist in the development of strategies to prevent this disease.

Using deep learning to assess the function of gastroesophageal flap valve according to the Hill classification system.

BACKGROUND: The endoscopic Hill classification of the gastroesophageal flap valve (GEFV) is of great importance for understanding the functional status of the esophagogastric junction (EGJ). Deep learning (DL) methods have been extensively employed in the area of digestive endoscopy. To improve the efficiency and accuracy of the endoscopist's Hill classification and assist in incorporating it into routine endoscopy reports and GERD assessment examinations, this study first employed DL to establish a four-category model based on the Hill classification. MATERIALS AND METHODS: A dataset consisting of 3256 GEFV endoscopic images has been constructed for training and evaluation. Furthermore, a new attention mechanism module has been provided to improve the performance of the DL model. Combined with the attention mechanism module, numerous experiments were conducted on the GEFV endoscopic image dataset, and 12 mainstream DL models were tested and evaluated. The classification accuracy of the DL model and endoscopists with different experience levels was compared. RESULTS: 12 mainstream backbone networks were trained and tested, and four outstanding feature extraction backbone networks (ResNet-50, VGG-16, VGG-19, and Xception) were selected for further DL model development. The ResNet-50 showed the best Hill classification performance; its area under the curve (AUC) reached 0.989, and the classification accuracy (93.39%) was significantly higher than that of junior (74.83%) and senior (78.00%) endoscopists. CONCLUSIONS: The DL model combined with the attention mechanism module in this paper demonstrated outstanding classification performance based on the Hill grading and has great potential for improving the accuracy of the Hill classification by endoscopists.

A new attention mechanism module has been proposed and integrated into the DL model.According to our knowledge, this is the first study to establish a four-category DL model based on the Hill grading.The DL model demonstrated outstanding classification performance based on the Hill grading and has great potential for improving the accuracy of the Hill classification by endoscopists.

Home-Based Transcutaneous Neuromodulation Improved Constipation via Modulating Gastrointestinal Hormones and Bile Acids.

This study aims to investigate the role of transcutaneous neuromodulation (TN) on the regulation of gastrointestinal hormones and bile acids in patients with functional constipation (FC). Twenty FC patients were treated with TN for four weeks. The effects of TN on symptoms were evaluated by questionnaires. Plasma levels of serotonin (5-HT), motilin, somatostatin, and vasoactive intestinal peptide (VIP) were measured by ELISA and 12 individual bile acids assayed by liquid chromatography tandem mass spectrometry. Results were as follows. (1) TN treatment increased the frequency of spontaneous bowel movement, improved the Bristol Stool Score, and reduced Patient Assessment of Constipation Symptom score and Patient Assessment of Constipation Quality of Life score. (2) FC patients showed decreased plasma levels of 5-HT, motilin, and VIP and an increased plasma level of somatostatin (P < 0.05). Four-week TN treatment increased plasma levels of 5-HT and motilin and decreased the plasma level of somatostatin in the FC patients (P < 0.05). (3) Taurocholic deoxycholate, taurocholic acid, and taurocholic lithocholic acid were increased in the FC patients (P < 0.005) but reduced by TN treatment (P < 0.05). This study has suggested that the therapy may improve the symptoms of FC by alleviating the disorders of gastrointestinal hormones and bile acids.

[Effects of controlled alternate partial root-zone drip irrigation on apple seedling morphological characteristics and root hydraulic conductivity].

To investigate the effects of alternate partial root-zone drip irrigation (ADI) on the morphological characteristics and root hydraulic conductivity of apple seedlings, three irrigation modes, i.e., fixed partial root-zone drip irrigation (FDI, fixed watering on one side of the seedling root zone), controlled alternate partial root-zone drip irrigation (ADI, alternate watering on both sides of the seedling root zone), and conventional drip irrigation (CDI, watering cling to the seedling base), and three irrigation quotas, i. e., each irrigation amount of FDI and ADI was 10, 20 and 30 mm, and that of CDI was 20, 30 and 40 mm, respectively, were designed. In treatment ADI, the soil moisture content on the both sides of the root zone appeared a repeated alternation of dry and wet process; while in treatment CDI, the soil moisture content had less difference. At the same irrigation quotas, the soil moisture content at the watering sides had no significant difference under the three drip irrigation modes. At irrigation quota 30 mm, the root-shoot ratio, healthy index of seedlings, and root hydraulic conductivity in treatment ADI increased by 31.6% and 47.1%, 34.2% and 53.6%, and 9.0% and 11.0%, respectively, as compared with those in treatments CDI and FDI. The root dry mass and leaf area had a positive linear correlation with root hydraulic conductivity. It was suggested that controlled alternate partial root-zone drip irrigation had obvious compensatory effects on the root hydraulic conductivity of apple seedlings, improved the soil water use by the roots, benefited the equilibrated dry matter allocation in seedling organs, and markedly enhanced the root-shoot ratio and healthy index of the seedlings.