Toll-like receptor (TLR) is essential for the immune response to Mycobacterium tuberculosis (MTB) infection. However, the mechanism whereby TLR mediates the MTB-induced pleural mesothelial hyperpermeability in tuberculous pleural effusion (TBPE) remains unclear. Pleural effusion size and pleural fluid levels of vascular endothelial growth factor (VEGF) and soluble TLR2 (sTLR2) in patients with TBPE (n = 36) or transudative pleural effusion (TPE, n = 16) were measured. The effects of MTB H37Ra (MTBRa) on pleural mesothelial permeability and the expression of VEGF and zonula occludens (ZO)-1 in human pleural mesothelial cells (PMCs) were assessed. Levels of VEGF and sTLR2 were significantly elevated in TBPE compared to TPE. Moreover, effusion VEGF levels correlated positively, while sTLR2 values correlated negatively, with pleural effusion size in TBPE. In human PMCs, MTBRa substantially activated JNK/AP-1 signaling and upregulated VEGF expression, whereas knockdown of TLR2 remarkably inhibited MTBRa-induced JNK phosphorylation and VEGF overexpression. Additionally, both MTBRa and VEGF markedly reduced ZO-1 expression and induced pleural mesothelial permeability, while TLR2 silencing or pretreatment with anti-VEGF antibody significantly attenuated the MTBRa-triggered effects. Collectively, TLR2 mediates VEGF overproduction and downregulates ZO-1 expression in human PMCs, leading to mesothelial hyperpermeability in TBPE. Targeting TLR2/VEGF pathway may confer a potential treatment strategy for TBPE.

Endotracheal tubes (ETTs) provide a vital connection between the ventilator and patient; however, improper placement can hinder ventilation efficiency or injure the patient. Chest X-ray (CXR) is the most common approach to confirming ETT placement; however, technicians require considerable expertise in the interpretation of CXRs, and formal reports are often delayed. In this study, we developed an artificial intelligence-based triage system to enable the automated assessment of ETT placement in CXRs. Three intensivists performed a review of 4293 CXRs obtained from 2568 ICU patients. The CXRs were labeled “CORRECT” or “INCORRECT” in accordance with ETT placement. A region of interest (ROI) was also cropped out, including the bilateral head of the clavicle, the carina, and the tip of the ETT. Transfer learning was used to train four pre-trained models (VGG16, INCEPTION_V3, RESNET, and DENSENET169) and two models developed in the current study (VGG16_Tensor Projection Layer and CNN_Tensor Projection Layer) with the aim of differentiating the placement of ETTs. Only VGG16 based on ROI images presented acceptable performance (AUROC = 92%, F1 score = 0.87). The results obtained in this study demonstrate the feasibility of using the transfer learning method in the development of AI models by which to assess the placement of ETTs in CXRs.