Abstract
Objectives: To develop and validate a transthoracic lung ultrasound (TLUS)-integrated clinical nomogram for predicting pulmonary arterial hypertension (PAH) in patients with connective tissue disease-associated interstitial lung disease (CTD-ILD). Methods: This multicenter retrospective study included 550 patients with CTD-ILD from the Second Affiliated Hospital of Fujian Medical University and 169 external cases from the Xijing Hospital, Fourth Military Medical University. Patients were randomly divided into a training cohort (n = 385) and an internal validation cohort (n = 165); the external dataset served as a testing cohort. Demographic, physiological, laboratory, pulmonary function, and TLUS data were collected. Univariate and multivariate logistic regression analyses identified independent predictors of PAH, which were used to construct a nomogram model. Discrimination was assessed using receiver operating characteristic (ROC) curves and area under the curve (AUC) values. Calibration, decision curve analysis (DCA), and clinical impact curves (CIC) were performed to evaluate model accuracy and clinical utility. Results: Five independent predictors were identified: respiratory rate, diffusing capacity of the lung for carbon monoxide (DLCO% predicted), TLUS score, red blood cell (RBC) count, and brain natriuretic peptide (BNP). The model achieved excellent discrimination with AUCs of 0.952 (95% confidence interval [CI]: 0.927–0.977) in the training cohort, 0.935 (95% CI: 0.885–0.985) in the validation cohort, and 0.874 (95% CI: 0.806–0.942) in the testing cohort, outperforming individual predictors. Calibration plots showed close agreement between predicted and observed probabilities, while DCA and CIC confirmed strong clinical benefit and applicability across all thresholds. Conclusions: This TLUS-integrated nomogram provides a noninvasive and reliable tool for individualized PAH risk assessment in CTD-ILD patients. By combining ultrasound findings with physiological and laboratory markers, the model enables accurate detection of high-risk cases and may assist clinicians in optimizing surveillance and management strategies.