Background:Pulmonary artery hypertension (PAH) is an incurable disease with a high mortality rate.Current medications ameliorate symptoms but cannot target adverse vascular remodeling. New therapeutic strategies for PAH need to be established. Methods:Using the weighted gene coexpression network analysis (WGCNA) algorithm, we constructed a coexpression network of dataset GSE117261 from the Gene Expression Omnibus (GEO) database. Key modules were identified by the relationship between module eigengenes and clinical traits. Hub genes were screened out based on gene significance (GS), module membership (MM), and mean pulmonary artery pressure (mPAP). External validations were conducted in GSE48149 and GSE113439. Functional enrichment and immune cell infiltration were analyzed using Metascape and CIBERSORTx. Results:The WGCNA analysis revealed 13 coexpression modules. The pink module had the highest correlation with PAH in terms of module eigengene (r=0.79; P=2e−18) and module significance (MS =0.43).Functional enrichment indicated genes in the pink module contributed to the immune system process and extracellular matrix (ECM). In the pink module, ECM2 (GS =0.65, MM =0.86, ρ=0.407, P=0.0019) and GLT8D2 (GS =0.63, MM =0.85, ρ=0.443, P=0.006) were identified as hub genes. For immune cells infiltration in PAH lung tissue, hub genes were positively correlated with M2 macrophages and resting mast cells, and were negatively correlated with monocytes, neutrophils, and CD4-naïve T cells. Conclusions:Our research identified 2 hub genes ECM2 and GLT8D2 related to PAH. The functions of these hub genes were involved in the immune process and ECM, indicating that they might serve as candidate therapeutic targets for PAH.

Keywords：lesion detection;| Computer-Aided Diagnosis CAD; convolutional neural network; deep learning Objective：We developed a universal lesion detector (ULDor) which showed good performance in inlab experiments. The study aims to evaluate the performance and its ability to generalize in clinical setting via both external and internal validation.Methods：The ULDor system consists of a convolutional neural network (CNN) trained on around 80K lesion annotations from about 12K CT studies in the DeepLesion dataset and 5 other public organspecific datasets. During the validation process, the test sets include two parts: the external validation datasets were comprised of 164 sets of non-contrasted chest-upper abdomen CT scans collected from a comprehensive hospital, and the internal validation datasets are comprised of 187 sets of low-dose helical CT scans from the National Lung Screening Trial (NLST). We run the model on the two test sets to output lesion detection. Three board-certified radiologists read the CT scans and verify the detection results by ULDor. We used positive predictive value (PPV) and sensitivity to measure the performance of the model in detecting space-occupying lesions at all extra-pulmonary organs, including the liver, kidney, pancreas, thyroid, lymph nodes, body wall, and thoracic spine at visualized CT images.Results：In the external validation, the lesion-level PPV and sensitivity of the model are 58% and 66%, respectively. On average, the model detected 2.1 findings per set, and among them, 0.9 were false positives. ULDor worked well for detecting liver lesions, with a PPV of 79% and a sensitivity of 93%, followed by kidney, with a PPV of 70% and a sensitivity of 58%. In internal validation with NLST test set, ULDor obtained a PPV of 75% and a sensitivity of 52% despite the relatively high noise level of soft tissue in CT images.