The study was approved by the IRB and the informed consent was waived.
Subject selection
We divided the patients into 2 groups,
normal pulmonary artery pressure (< or = 25 mmHg) and pulmonary hypertension (>25 mmHg),
based on the mean pulmonary artery pressure (2) documented from right heart catheterization performed by the University of Maryland cardiology service between January 2010 and December 2012.
Patients with prior cardiac surgery,
congenital heart disease,
pericardial disease,
severe lung disease and critically ill patients documented on CT examinations were excluded from the study.
A total of 95 patients with available CT examinations,
45 patients with normal pulmonary artery pressure and 50 patients with PHTN were included and their CT examinations were retrospectively reviewed.
CT Examinations and Imaging Analysis
Axial CT examinations of the chest were performed using various CT parameters depending on the patient’s factors such as Body Surface Area (BSA)),
clinical indications and available CT scanners (40,
64 or 256 detector scanners).
CT scan of the chest without contrast enhancement includes routine CT chest without contrast and High Resolution CT Chest.
The following CT parameters were used [email protected] Thickness/Increment,
100-120 kVp,
200-250 mAs,
2-128 x 0.625 collimation,
0.824-0.993 Pitch,
0.5 s rotation time and 350-400 Scan FOV.
CT scan of the chest with contrast includes routine CT chest with intravenous contrast,
Computed Tomography Pulmonary Artery Angiography,
Computed Tomography Coronary Angiography and Cardiac Computed Tomography.
The following CT parameters were used 0.9-4@ 0.45-3 Thickness/Increment,
120 kVp,
250-400 mAs,
40-128 x 0.625 collimation,
0.703-0.993 Pitch,
0.27-0.5 s rotation time and 220-400 Scan FOV. The tube current,
voltage and scan FOV were also adjusted accordingly to the patient’s BMI who underwent CT Pulmonary Angiogram,
CT Coronary Angiogram and Cardiac CT.
Three radiologists (CTL,
SJK,
JJ) with 1,
5 and 8 years experience in cardiothoracic imaging measured the main pulmonary artery diameter on axial images using an electronic ruler on 3D work- station (Terarecon/Aquarius net). Radiologists chose the key images to measure MPA,
RPA and LPA independently but were not allowed to magnify the images or to view the images on sagittal or coronal reformats.
Adjusting the window and level was permitted.
The radiologists were also blinded to the patient clinical symptoms,
diagnosis and each other’s measurement to prevent bias.
The measurement then was documented for future data analysis.
Data Collection
We collected the patient’s demographic data such as age,
gender,
body weight (kg) and height (cm).
Body surface area (BSA,
m2) was calculated from the weight and height by the DuBois formula (BSA (m2)=0.007184 x Height (cm) 0.725 x Weight (Kg) 0.425).
Clinical parameters such as pulmonary artery systolic and diastolic pressure were also recorded.
Mean pulmonary artery pressure (MPAP) was calculated from systolic pressure (SP) and diastolic pulmonary pressure (DP).
MPAP (mmHg) is equivalent to DP + 1/3 (SP-DP).
This information was collected from right cardiac catherization.
The size and the distance from the pulmonary artery bifurcation of measured MPA were measured in centimeters.
This data was acquired from the 3D CT work- station using axial multiplanar reformatted images ( Fig. 1 ).
Statistical analysis
For patients with normal PAP and PHTN:
We tested the normality of the main pulmonary artery diameter (MPAD) by the Saphiro-Wilk test.
To analyze the statistical significance of MPAD and patients’ profiles such as age,
height,
weight and body surface area,
linear regression was used.
Student T-test was used for the gender.
The analysis was performed for both normal PAP and PHTH groups.
To evaluate inter-observer agreement of using 3.15 cm as a cut off value to diagnose PHTN,
we used Cohen’s kappa test.
We also evaluated the inter-rater variation of the measured MPAD in terms of size and its relationship to the bifurcation by one-way ANOVA test.
To determine whether the addition of RPA and LPA measurements improve the accuracy in diagnosing PHTN,
we added RPA and/or LPA diameter greater than 2.2 cm to the diagnostic criteria .
Studies showing either an enlarged MPAD (>/= 3.15 cm) or an enlarged RPA or LPA were considered positive and the differences between controls and PHTN patients computed using Chi-Square test.
To verify the accuracy of different cut off values of MPAD at 3.15 cm to diagnose PHTN,
we calculated the statistical values and demonstrated the data in the formats of histograms and ROI curves.