BackgroundLung adenocarcinoma (LUAD) is one of the most predominant subtypes of lung cancer. The gut microbiome plays a vital role in the pathophysiological processes of various diseases, including cancers.MethodsIn the study, 100 individuals were enrolled. In total 75 stool and blood samples were analyzed with 16s-rRNA gene sequencing and metabolomics (30 from healthy individuals (H); 45 from LUAD patients). In addition, 25 stool samples were analyzed with metagenomics (10 from H; 15 from LUAD). The linear discriminant analysis (LDA) effect size (LefSe) and logistic regression analysis were applied to identify biomarkers’ taxa and develop a diagnostic model. The diagnostic power of the model was estimated with the receiver operating characteristic curve (ROC) by comparing the area under the ROC (AUC). The correlation between biomarker’s taxa and metabolites was calculated using the Spearman analysis.ResultsThe α and β diversity demonstrated the composition and structure of the gut microbiome in LUAD patients were different from those in healthy people. The top three abundance of genera were Bacteroides (25.06%), Faecalibacterium (11.00%), and Prevotella (5.94%). The LefSe and logistic regression analysis identified three biomarker taxa (Bacteroides, Pseudomonas, and Ruminococcus gnavus group) and constructed a diagnostic model. The AUCs of the diagnostic model in 16s-rRNA gene sequencing and metagenomics were 0.852 and 0.841, respectively. A total of 102 plasma metabolites were highly related to those three biomarkers’ taxa. Seven metabolic pathways were enriched by 102 plasma metabolites, including the Pentose phosphate pathway, Glutathione metabolism.ConclusionsIn LUAD patients, the gut microbiome profile has significantly changed. We used three biomarkers taxa to develop a diagnostic model, which was accurate and suitable for the diagnosis of LUAD. Gut microbes, especially those three biomarkers’ taxa, may participate in regulating metabolism-related pathways in LUAD patients, such as the pentose phosphate pathway and glutathione metabolism.

This study aims to explore how the insurance business can combine artificial intelligence (AI), huge data technology and good design under the ecological background of AIoT to improve the efficiency of double-entry in the insurance business and improve the user experience. It is accomplished through observation, user interview and other methods to explore user needs and distress points, comprehensive use of image and video analysis, face recognition, behaviour detection and other AI technologies. aiCore intelligent dual-recording system is proposed, which integrated mobile terminal dual-recording application and backstage quality control system. The proposed system realised the design optimisation of the whole process of intelligent dual-recording scenario combining the three functions ‘front-end intelligent dual-recording, in-process real-time quality control and post-AI quality control.’ Using in-process detection, post-review and manual final audit, the aiCore intelligent dual-recording system has significantly improved the efficiency of dual-recording, quality inspection pass rate and user experience satisfaction.