Detection of severe pulmonary hypertension based on computed tomography pulmonary angiography

Pulmonary hypertension (PH) is often diagnosed late in the disease course. As many patients may undergo computed tomography pulmonary angiography (CTPA) for exclusion of pulmonary embolism (PE), we aimed to create a model that can detect the existence of PH and grade its severity. Consecutive patients who underwent CTPA which was negative for PE, and echocardiography study within 24 h, were included. The CT parameters evaluated to assess PH were: the diameters of the main pulmonary artery (MPA), ascending aorta (AA), calculation of each heart chamber volume, and the severity of reflux of contrast material. Randomly, 70% of patients were included in the model creation group, and 30% were used to validate the model. The final study group included 740 patients, 268 male patients, median age 72 years. 374 patients (51%) had PH, of them 94 (13%) had severe PH on the echocardiography. Right atrium (RA) and Left atrium (LA) volume indices were the strongest parameter to indicate PH (area under the curve, AUC = 0.738 and 0.736, respectively), while Right ventricle (RV) and RA volume indices were the strongest parameter to identify severe PH (AUC = 0.735 and 0.715, respectively) with MPA diameter being the least influential indicator (AUC = 0.623). Using the patients age, gender, and multiple CTPA parameters, we created a model for predicting the existence of severe PH. After validation, the model demonstrated 91% sensitivity and a negative predictive value of 97%. Applying our models, CTPA can be used to identify severe PH immediately after the completion of CTPA exam.