Format

Send to

Choose Destination
Eur Radiol. 2019 Apr 8. doi: 10.1007/s00330-019-06188-7. [Epub ahead of print]

Combined automated 3D volumetry by pulmonary CT angiography and echocardiography for detection of pulmonary hypertension.

Author information

1
Department of Diagnostic and Interventional Radiology, Heidelberg University Hospital, Im Neuenheimer Feld 110, 69120, Heidelberg, Germany.
2
Translational Lung Research Center Heidelberg (TLRC), Member of the German Center for Lung Research (DZL), University of Heidelberg, Heidelberg, Germany.
3
Biomedical Computer Vision Group, BIOQUANT, IPMB and German Cancer Research Center (DKFZ), University of Heidelberg, Heidelberg, Germany.
4
Centre for Pulmonary Hypertension, Thoraxklinik at Heidelberg University Hospital, Heidelberg, Germany.
5
Department of Radiology, Section of Interventional Radiology, Cleveland Clinic Foundation, Cleveland, OH, USA.
6
Department of Radiology, Thoraxklinik at Heidelberg University Hospital, Heidelberg, Germany.
7
Department of Diagnostic and Interventional Radiology, Heidelberg University Hospital, Im Neuenheimer Feld 110, 69120, Heidelberg, Germany. fabian.rengier@web.de.
8
Translational Lung Research Center Heidelberg (TLRC), Member of the German Center for Lung Research (DZL), University of Heidelberg, Heidelberg, Germany. fabian.rengier@web.de.
9
Department of Radiology, German Cancer Research Center (DKFZ), Heidelberg, Germany. fabian.rengier@web.de.

Abstract

OBJECTIVES:

To assess the diagnostic accuracy of automated 3D volumetry of central pulmonary arteries using computed tomography pulmonary angiography (CTPA) for suspected pulmonary hypertension alone and in combination with echocardiography.

METHODS:

This retrospective diagnostic accuracy study included 70 patients (mean age 66.7, 48 female) assessed for pulmonary hypertension by CTPA and transthoracic echocardiography with estimation of the pulmonary arterial systolic pressure (PASP). Gold standard right heart catheterisation with measurement of the invasive mean pulmonary arterial pressure (invasive mPAP) served as the reference. Volumes of the main, right and left pulmonary arteries (MPA, RPA and LPA) were computed using automated 3D segmentation. For comparison, axial dimensions were manually measured. A linear regression model was established for prediction of mPAP (predicted mPAP).

RESULTS:

MPA, RPA and LPA volumes were significantly increased in patients with vs. without pulmonary hypertension (all p < 0.001). Of all measures, MPA volume demonstrated the strongest correlation with invasive mPAP (r = 0.76, p < 0.001). Predicted mPAP using MPA volume and echocardiographic PASP as covariates showed excellent correlation with invasive mPAP (r = 0.89, p < 0.001). Area under the curves for predicting pulmonary hypertension were 0.94 for predicted mPAP, compared to 0.90 for MPA volume and 0.92 for echocardiographic PASP alone. A predicted mPAP > 25.8 mmHg identified pulmonary hypertension with sensitivity, specificity, positive and negative predictive values of 86%, 93%, 95% and 81%, respectively.

CONCLUSIONS:

Automated 3D volumetry of central pulmonary arteries based on CTPA may be used in conjunction with echocardiographic pressure estimates to noninvasively predict mPAP and pulmonary hypertension as confirmed by gold standard right heart catheterisation with higher diagnostic accuracy than either test alone.

KEY POINTS:

• This diagnostic accuracy study derived a regression model for noninvasive prediction of invasively measured mean pulmonary arterial pressure as assessed by gold standard right heart catheterisation. • This regression model using automated 3D volumetry of the central pulmonary arteries based on CT pulmonary angiography in conjunction with the echocardiographic pressure estimate predicted pulmonary arterial pressure and the presence of pulmonary hypertension with good diagnostic accuracy. • The combination of automated 3D volumetry and echocardiographic pressure estimate in the regression model provided superior diagnostic accuracy compared to each parameter alone.

KEYWORDS:

Computed tomography angiography; Pulmonary arteries; Pulmonary artery pressure; Pulmonary artery volumetry; Pulmonary hypertension

PMID:
30963276
DOI:
10.1007/s00330-019-06188-7

Supplemental Content

Full text links

Icon for Springer
Loading ...
Support Center