Prediction of hemoptysis in patients with anomalous systemic arterial supply to normal basal segments of the lower lobe using a combination of CT features and clinical materials.

Abstract

Anomalous systemic arterial supply to the normal basal segments of the lower lobe (ASALL) is a rare anomaly with a common complication of hemoptysis. To estimate the risk of hemoptysis, this study aims to investigate the value of contrast-enhanced computed tomography (CT) and construct a risk-scoring model based on radiological features and clinical materials of patients with ASALL. Forty-three eligible individuals (17 women and 26 males), who underwent multiphase contrast-enhanced CT, were included in this study. Hemoptysis was predicted by combined systemic arterial features (CD-A) and combined demographic and radiological features (CD-R). Potential hemoptysis predictors were identified using multivariate regression analysis. A receiver operating characteristic (ROC) curve analysis was used to assess the prediction efficiency. The coefficient of regression model was used to build a combined risk scoring (CRS) model for hemoptysis. The decision curve analysis (DCA) was performed to evaluate the clinical usefulness of the risk-scoring model. Hemoptysis was present in 17 (39.5%) ASALL patients. The areas under the curve (AUCs) for the predicted performance of CD-A and CD-R were 0.869 and 0.890, respectively. Independent predictors generated a scoring model using the formula CRS = 3 × age + 3 × sex + 4 × [ground glass opacity (GGO)] + 3 × (CD-A >0.522). The prediction performance of this model was displayed with an AUC of 0.939. This scoring model was demonstrated to be significantly preferable to CD-A (P=0.046) and CD-R (P=0.02) by the Hanley and McNeil test. The DCA showed that the CRS model was more beneficial when the threshold probability was between 5% and 92%. The scoring model offers a viable method for evaluating the risk of hemoptysis in patients with ASALL by combining radiological and clinical data.