Tuberculosis is one of the major infectious diseases that seriously endanger human health. Since 2014, it has surpassed human immunodeficiency virus/acquired immunodeficiency syndrome as the first infectious disease in patients with single pathogens. China is the third-largest country in the world in terms of high burden of tuberculosis. In 2016, there were about 900 000 new cases of tuberculosis in China. China is facing a severe tuberculosis epidemic, especially for the early diagnosis of tuberculosis and misdiagnosis of tuberculosis, which leads to delay in treatment and the spread of tuberculosis. With the application of artificial intelligence in the medical field, machine learning and deep learning methods have shown important value in the diagnosis of tuberculosis. This article will explain the application status and future development of machine learning and deep learning in the diagnosis of tuberculosis.
Objective To explore the relationships between the polymorphisms of inhibitor genes WIF1 and DKK1 in WNT signaling pathway and susceptibility to tuberculosis, clinical characteristics and laboratory indexes. Methods From December 2014 to November 2016, 475 tuberculosis patients and 370 healthy controls of West China Hospital of Sichuan University were enrolled in the study, and the clinical data of the subjects were collected. High-throughput genotyping technique was used to detect the polymorphisms of WIF1 rs58635985 and DKK1 rs11001548 in WNT signaling pathway. The allele frequency distribution, genotype, genetic model, clinical features and laboratory indexes of two single nucleotide polymorphisms were analyzed by χ2 test and logistic regression analysis. Results There was no significant difference in the allele frequency distribution (P=0.275, 0.949), genotype (P=0.214, 0.659) or genetic models: additive model (P=0.214, 0.659), dominant model (P=0.414, 0.827), recessive model (P=0.227, 0.658) of rs58635985 and rs11001548 between the tuberculosis group and the healthy control group. Subgroup analysis showed no significant difference in allele and genotype distribution between rs58635985 and rs11001548 (pulmonary tuberculosis group vs. healthy control group: P>0.05; pulmonary tuberculosis groupvs. extra-pulmonary tuberculosis group: P>0.05). There was no significant difference in the clinical features (fever, night sweat, fatigue,etc.) or laboratory indexes (complete blood count, erythrocyte sedimentation rate, TB-DNA, etc.) (P>0.05). Conclusions There is no association between rs58635985 of WIF1 gene or rs11001548 of DKK1 gene and genetic susceptibility, clinical characteristics and laboratory indexes in Han population in Western China. To expand the sample size for verification and analysis in different populations is necessary.
ObjectiveTo evaluate the expression level and diagnostic value of lnc-PAPSS2-2 (lnc-PA) in peripheral blood of active pulmonary tuberculosis (PTB) patients.MethodsFrom January 2011 to January 2018, 798 patients with active PTB and 1 650 healthy people undergoing health examination in West China Hospital of Sichuan University and their electronic health records (EHR) were collected. Peripheral blood lnc-PA levels were quantified by quantitative real-time polymerase chain reaction method. The data of lnc-PA and EHR were modeled using nomogram, and the receiver operating characteristic (ROC) curves of lnc-PA, EHR and the combination of lnc-PA and EHR were compared to evaluate the diagnostic value of lnc-PA for active PTB.ResultsThe level of lnc-PA was lower in active PTB patients than that in healthy controls (P<0.001). The areas under ROC curve of lnc-PA, EHR and their combination were 0.619, 0.962, and 0.964 in the training set and 0.626, 0.950, and 0.950 in the validation set, respectively.ConclusionThe diagnostic ability of lnc-PA is poor and that of EHR is good, which indicates that the clinical value of lnc-PA as a biomarker of active PTB remains to be further explored.