Pulmonary hypertension (PH) is defined as a mean pulmonary arterial pressure (mpap) of at least 25 mm Hg and is classified into five groups according

Transcription

1 Note: This copy is for your personal, non-commercial use only. To order presentation-ready copies for distribution to your colleagues or clients, contact us at Anand Devaraj, MD(Res), MRCP, FRCR Athol U. Wells, MD, FRCR, FRACP Mark G. Meister, MBCHB, MRCP, FRCR Tamera J. Corte, MBBS, FRACP Stephen J. Wort, MD, MRCP David M. Hansell, MD, FRCP, FRCR Detection of Pulmonary Hypertension with Multidetector CT and Echocardiography Alone and in Combination 1 Purpose: Materials and Methods: To test the reliability of potentially new computed tomographic (CT) indicators of pulmonary hypertension (PH) and to establish whether a combination of CT and echocardiographic measurements was more predictive of PH than either test alone. The institutional review board approved this retrospective study; patient consent was not required. Seventyseven patients undergoing right-sided heart catheterization were examined. CT diameters of the main pulmonary artery, ascending aorta, and thoracic vertebra and crosssectional area of the main pulmonary artery were measured. Segmental and subsegmental arterial diameters were recorded, and segmental artery size was compared with adjacent bronchus size by using a semiquantitative scoring system. The relationship between CT measurements and mean pulmonary arterial pressure (mpap) was tested with linear regression. Multivariate regression was used to establish a composite index of mpap by using CT markers of PH with echocardiography-derived right ventricular systolic pressure (RVSP). Post hoc logistic regression and receiver operating characteristic curve analysis were performed to test the diagnostic ability of the CT echocardiography composite. ORIGINAL RESEARCH n THORACIC IMAGING 1 From the Department of Radiology, Royal Brompton Hospital, Sydney Street, London, SW3 6NP, England. Received March 30, 2009; revision requested June 11; revision received July 13; accepted August 4; fi nal version accepted September 9. Address correspondence to D.M.H. ( davidhansell@rbht.nhs.uk ). Results: Conclusion: The ratios of the diameter of the main pulmonary artery to the diameter of the ascending aorta ( R 2 = 0.45; P,.001 ) and of the cross-sectional area of the pulmonary artery to the diameter of the ascending aorta ( R 2 = 0.45; P,.001) correlated equally with mpap. The ratio of the diameter of the main pulmonary artery to the diameter of the thoracic vertebra, the segmental arterial diameter, and the segmental artery-to-bronchus ratio were related to mpap but did not strengthen correlations compared with the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta alone. A composite index of the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta and echocardiographyderived RVSP was more strongly related ( R 2 = 0.55) to mpap and was more significantly predictive of PH than either measure alone. A combination of CT and echocardiographic markers of PH is more closely related to mpap than either test in isolation. q RSNA, 2010 q RSNA, 2010 Radiology: Volume 254: Number 2 February 2010 n radiology.rsna.org 609

2 Pulmonary hypertension (PH) is defined as a mean pulmonary arterial pressure (mpap) of at least 25 mm Hg and is classified into five groups according to cause ( 1 ). Regardless of the cause, the diagnosis of increased pulmonary arterial pressure is important because PH is associated with a poor prognosis ( 2 5 ). Rightsided heart catheterization remains the reference standard for the diagnosis of PH, although it is an invasive test with recognized complications ( 6 ). Echocardiography is a frequently used noninvasive screening tool that provides an estimate of systolic pulmonary arterial pressure. However, this measurement is sometimes impossible to make; moreover, it may be inaccurate, especially in the setting of diffuse lung disease ( 7 ). Computed tomography (CT) is commonly performed in patients suspected of having PH or in patients with an underlying diffuse lung disease who may be at risk for PH. In addition, the structure of the pulmonary vasculature at CT has been extensively studied as Advances in Knowledge n There is a moderately strong relationship between the size of the segmental pulmonary arteries and mean pulmonary arterial pressure (mpap). n A composite index of the pulmo- nary arterial diameter ascending aorta diameter ratio and echocardiography-derived right ventricular systolic pressure is more strongly predictive of pulmonary hypertension (PH) than either CT or echocardiographic measurements alone. n Measuring pulmonary arterial size by using coronal oblique reformats to calculate the ratio of the cross-sectional area of the main pulmonary artery to the diameter of the ascending aorta does not strengthen correlations with mpap compared with the ratio of the simple axial diameter of the main pulmonary artery to the diameter of the ascending aorta. a marker of increased mpap ( 8 13 ). In this regard, most investigators have concentrated on dilation of the main pulmonary artery (both in absolute terms and in relation to the size of the ascending aorta) at axial CT as a sign of PH; reported correlations between mpap and pulmonary arterial size have varied from weak to strong ( 8 13 ). It has also been shown that pulmonary arterial size (but not the ratio of the diameter of the pulmonary artery to the diameter of the ascending aorta) is an unreliable marker of mpap in patients with pulmonary fibrosis ( 13 ). Few studies have explored the more peripheral pulmonary vasculature, although an increase in the segmental artery to bronchus ratio in at least three lobes has been shown to be a reliable indicator of PH ( 12,14 ). The initial aim of this study was to establish whether it was possible to improve on the widely used CT sign of PH (namely, an increased ratio of the pulmonary arterial diameter to the diameter of the ascending aorta) by examining alternative CT measurements in a heterogeneous group of patients. The following measurements and ratios were evaluated: (a) the ratio of the cross-sectional area of the main pulmonary artery to the diameter of the ascending aorta, with multiplanar CT reformats used to calculate the pulmonary arterial area (because this measurement may more accurately reflect the size of the pulmonary artery); (b) the ratio of the diameter of the pulmonary artery to vertebral body diameter (given that the vertebral body is a more fixed structure representative of patient size [ 15 ]); (c) the dimensions of the segmental pulmonary arteries and a subjective assessment of the segmental artery bronchus ratio; and (d) the dimensions of the next generation Implication for Patient Care n A composite index of CT and echocardiography can be used to more accurately stratify patients for right-sided heart catheterization by increasing the likelihood of a confident diagnosis of PH. of pulmonary artery (ie, subsegmental artery). A further aim of the study was to establish whether morphologic CT markers of pulmonary arterial pressure could be strengthened by combining functional information obtained from echocardiography-derived right ventricular systolic pressure (RVSP) measurements. In summary, the purpose of the study was to test the reliability of potentially new CT indicators of PH and to establish whether a combination of CT and echocardiographic measurements was more predictive of PH than either test alone. Materials and Methods Patients This retrospective study was approved by the institutional review board. Informed consent was not required. We reviewed 77 patients (mean age, 55.6 years; median age, 59 years; age range, years ) at our institution who had undergone both volumetric thoracic CT and right-sided heart catheterization between March 2002 and March 2007 within 9 months of each other. Patients with congenital heart disease were excluded. The group consisted of 39 male patients (mean age, 58.3 years; median Published online /radiol Radiology 2010; 254: Abbreviations: AUC = area under the receiver operating characteristic curve CI = confi dence interval mpap = mean pulmonary arterial pressure PH = pulmonary hypertension RVSP = right ventricular systolic pressure Author contributions: Guarantor of integrity of entire study, D.M.H.; study concepts/study design or data acquisition or data analysis/ interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript fi nal version approval, all authors; literature research, A.D., S.J.W.; clinical studies, A.D., T.J.C., S.J.W., D.M.H.; statistical analysis, A.D., A.U.W.; and manuscript editing, A.D., A.U.W., T.J.C., S.J.W., D.M.H. Authors stated no fi nancial relationship to disclose. 610 radiology.rsna.org n Radiology: Volume 254: Number 2 February 2010

3 age, 62 years; age range, years) and 38 women (mean age, 52.9 years; median age, 52.5 years; age range, years). The patients had a spectrum of diseases associated with PH (groups 1, 3, 4, and 5 of the Venice PH classification system) ( 1 ) ( Table 1 ). Rightsided heart catheterization was performed by using standard techniques ( 16 ), with mpap obtained at rest in all patients. PH was defined as an mpap greater than 25 mm Hg. These patients had been examined in a previous study ( 13 ) that explored the validity of existing CT markers of PH in patients with pulmonary fibrosis. Echocardiography Transthoracic echocardiography was performed with an HP 7500 or 5500 model (Hewlett-Packard, Palo Alto, Calif) or a Vivid i model (GE Healthcare, Piscataway, NJ). Echocardiography-derived RVSP measurements were obtained by using standard techniques ( 17,18 ). Seventy-two of 77 patients (94%) underwent echocardiography, performed at a median time interval of 1.7 months from right-sided heart catheterization (range, 0 12 months). Tricuspid regurgitation was documented in 61 of 72 patients (85%). Table 1 Diagnoses in Patients Diagnosis No. of Patients Idiopathic pulmonary fi brosis 16 (21) Nonspecifi c interstitial pneumonia or 11 (14) fi brotic organizing pneumonia Chronic hypersensitivity 3 (4) pneumonitis Idiopathic pulmonary arterial 17 (22) hypertension Chronic thromboembolic disease 6 (8) Pulmonary venoocclusive disease 2 (3) Emphysema 5 (6) Sarcoidosis 11 (14) Lymphangioleiomyomatosis 1 (1) Lymphocytic interstitial pneumonia 1 (1) Bronchiectasis 2 (3) Non small-cell lung cancer 2 (3) Note. Numbers in parentheses are percentages of patients. CT Acquisition CT was performed with a four- or 64- section CT scanner (Siemens, Forchheim, Germany). All patients were imaged in the supine position, with breath-holding at full inspiration. Standard acquisition parameters were used: 90 ma, 120 kvp, 1.4 pitch, and 0.5-second rotation time. Thirty-one patients underwent unenhanced high-spatial-resolution volumetric CT of the thorax. Forty-six patients underwent CT pulmonary angiographic examination, with 90 ml of intravenous contrast agent (iopromide, Ultravist 300; Bayer Schering, Berlin, Germany) administered at 4 5 ml/sec; bolus tracking was used to trigger CT acquisition. CT scans were reconstructed at section widths of 1.25 mm (four-section scanner) or 1 mm (64-section scanner). CT was performed at a median time interval of 1 month from right-sided heart catheterization (range, 0 9 months). Image Interpretation CT scans were anonymized by an independent investigator not involved in subsequent scoring. Images were reviewed Figure 1 Figure 1: Axial CT image shows diameter of the main pulmonary artery measured at the level of the bifurcation. on a workstation (Aquarius Net; TeraRecon, San Mateo, Calif) by an observer who had no knowledge of the clinical information or pulmonary arterial pressure. The observer (A.D.), using electronic calipers, measured the widest short-axis diameter of the main pulmonary artery on axial sections at the level of the bifurcation ( Fig 1 ) (mediastinal window settings were used [width, 400 HU; level, 40 HU]). Electronic calipers were also used to measure the crosssectional area of the main pulmonary artery, just proximal to the bifurcation. This measurement was obtained with multiplanar reformats, such that the area to be measured was orthogonal to the long axis of the main pulmonary artery ( Fig 2 ). The widest short-axis diameters of the ascending aorta and thoracic vertebra were then measured on the same CT section used to measure the diameter of the main pulmonary artery. The mid anteroposterior diameter of the vertebra was chosen as a suitable area to evaluate because it was unlikely to be deformed by osteophytes or articulations with ribs ( 19 ). When osteophytes were present, they were not included in the measurement, and the vertebral body cortex was extrapolated from the existing definable margins of the rest of the vertebral body. Subsequently, the widest short-axis diameters were measured of the following four segmental arteries at their origins, according to established anatomic classification and nomenclature ( 20 ): apical segment of the right upper lobe, apicoposterior segment of the left upper lobe, posterior basal segment of the right lower lobe, and posterior basal segment of the left lower lobe. These arteries were selected because they run perpendicular to the scanning plane and are more reliably identified compared with more obliquely running vessels ( 21 ). Vascular structures were regarded as unmeasurable if they were distorted because of, for example, pulmonary fibrosis, or if there were anatomic variants. The average segmental arterial diameter was calculated (when four segmental arteries were identified). Radiology: Volume 254: Number 2 February 2010 n radiology.rsna.org 611

4 Figure 2 Figure 3 Figure 2: (a) Coronal-oblique reformat of the posterior main pulmonary artery, just anterior to the bifurcation. The pulmonary artery appears as an approximately circular structure. The image is created by aligning the ( b ) axial and ( c ) sagittal images along the long axis of the main pulmonary artery. Figure 3: CT image of posterobasal segmental artery of the right lower lobe (arrow) in a patient with nonspecifi c interstitial pneumonia and PH. The artery-bronchus ratio was scored as 3 (1.25 to two times the size of the adjacent bronchus). than 1.25 was considered abnormal and enlarged. Next, the widest short-axis dimensions of the following subsegmental arteries were measured at axial CT: the anterior branches of the apical segmental artery of the right upper lobe and the apicoposterior segmental artery of the left upper lobe and the mediobasal branches of the posterobasal segmental arteries of the right lower lobe and the left lower lobe. Identification of these small vessels was aided by tracing their course from the larger central pulmonary arteries on contiguous axial and coronal images, created by using the Aquarius Net software. The average subsegmental arterial diameter was calculated when four subsegmental arteries could be measured. A single observer measured these dimensions because it has previously been reported that the level of interand intraobserver variation with respect to CT caliper measurements is good ( 11,22 ). A semiquantitative assessment of the comparative size of the segmental arteries to the size of the segmental bronchi was performed independently by two experienced observers (A.D. and M.G.M., with 6 and 5 years of radiologic experience, respectively) by using the following scoring system: 1 indicated a ratio of short-axis diameter of artery to short-axis outer diameter of bronchus less than 0.75; 2, ratio between 0.75 and up to 1.25; 3, ratio between 1.25 and up to 2 ( Fig 3 ); 4, ratio between 2 and up to 3; 5, ratio greater than 3. An average of the two observers arterybronchus ratio scores was made. A consensus score was obtained in 59 of 308 instances (19%) in which there was a discrepancy between the two observers scores of greater than one grade or only one observer regarded the artery-bronchus ratio as unmeasurable. Lung window settings (width, 1500 HU; level, 500 HU) were used. Because segmental arteries may be slightly larger than the corresponding segmental bronchi in healthy individuals ( 23 ), an artery-bronchus ratio greater Statistical Analysis Data are expressed as medians with ranges (for nonnormally distributed data) or means with standard deviations (for normally distributed data) or proportions. Linear regression was used to assess the relationships between mpap derived from right-sided heart catheterization and all dimensions of the pulmonary vasculature, as well as ratios of dimensions. Observer variation of categorical (noncontinuous) data was expressed by using the weighted k coefficient. Multivariate regression analysis was conducted by using echocardiographyderived RVSP and the strongest individual CT correlate to identify whether a composite index of variables could be created to predict mpap. The assumptions of multiple linear regression were confirmed, as judged by testing for heteroscedasticity and examination of residual versus predictor plots. The ability of a CT echocardiography composite to distinguish between patients with and those without PH was assessed 612 radiology.rsna.org n Radiology: Volume 254: Number 2 February 2010

5 Table 2 Patient Characteristics Characteristic in a post hoc analysis by using logistic regression and the area under the receiver operating characteristic curve (AUC). Post hoc analyses were conducted to examine the validity of the CT echocardiography composite in patients in whom the time between CT and right-sided heart catheterization was restricted to 3 months and in patients with fibrotic lung disease. The accuracy of the CT echocardiography composite in patients with pulmonary fibrosis was compared with that in patients without pulmonary fibrosis by using the Fisher exact test. A P value less than.05 was considered to represent a statistically significant difference. Analyses were performed by using software (Stata, version 4; Stata, College Station, Tex). Results Datum Age (y)* 59.0 (17 79) Men 39 (51) Body surface area (m 2 ) Total lung capacity (%) Kco (%) mpap (mm Hg) Patients with PH 56 (73) Note. Unless otherwise noted, data are means 6 standard deviations. * Median age, with range in parentheses. Data are number of patients, with percentage in parentheses. Kco = gas transfer coeffi cient. Patient Characteristics and Observer Agreement Patient characteristics are outlined in Table 2. Mean mpap for the entire study group was 37.1 mm Hg (standard deviation); the median was 33.0 mm Hg (range, mm Hg). Fifty-six of 77 patients (73%) had PH (mpap. 25 mm Hg). The two observers artery-bronchus ratio scores showed good agreement. Weighted k values for those scores were 0.78 for the right upper lobe, 0.72 for the left upper lobe, 0.73 for the right Table 3 Relationships between CT and Echocardiographic Signs and mpap CT or Echocardiographic Variable lower lobe, and 0.79 for the left lower lobe. Main Pulmonary Artery: Correlation with mpap The ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta ( R 2 = 0.45; P,.001) and the ratio of the cross-sectional area of the main pulmonary artery to the diameter of the ascending aorta ( R 2 = 0.45; P,.001) correlated equally strongly with mpap ( Table 3 ). The ratio of the diameter of the main pulmonary artery to the diameter of the thoracic vertebra ( R 2 = 0.23; P,.001) weakened the association with mpap compared with the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta ( Table 3 ). Segmental and Subsegmental Pulmonary Arteries: Correlation with mpap In patients in whom all four segmental arteries could be measured by the single observer (50 of 77 patients [65%]), there was moderate correlation between the diameter of the segmental pulmonary artery and mpap ( R 2 = 0.24; P,.001). Evaluating the size of the segmental artery on the basis of a subjective comparison with the size of the accompanying bronchus showed a similar relationship between mean arterybronchus ratio scores and mpap ( R 2 = 0.19; P,.001) (in 43 of 77 [56%] patients for whom both observers agreed that all four segmental artery-bronchus ratios could be measured). Enlargement of at least three artery-bronchus R 2 Relationship with mpap P Value Diameter of pulmonary artery diameter of ascending aorta ratio 0.45,.001 Area of pulmonary artery diameter of ascending aorta ratio 0.45,.001 Diameter of pulmonary artery diameter of vertebral body ratio 0.23,.001 Diameter of segmental artery 0.24,.005 Average segmental artery bronchus ratio score 0.19,.005 Diameter of subsegmental artery Echocardiography-derived RVSP 0.44,.001 CT echocardiography composite 0.55,.001 ratios of four lobes was evident in 19 of 28 patients (68%) with PH and two of 15 patients (13%) without PH. The presence of at least three enlarged artery-bronchus ratios had a specificity of 87% and a sensitivity of 68% for the diagnosis of PH. The next generation of artery (the subsegmental artery) was measurable in all four lobes in 62 of 77 patients (81%). There was no correlation between the diameter of the subsegmental artery and mpap ( R 2 = 0.02; P =.3). CT and Echocardiography In 61 patients in whom tricuspid regurgitation was documented at echocardiography, there was a moderately strong correlation between echocardiographyderived RVSP and mpap ( R 2 = 0.44; P,.001). Multivariate regression analysis was used to determine whether echocardiography-derived RVSP and the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta (the strongest CT sign of PH) were independently related to mpap. This analysis was also used to generate a best-fit equation combining CT and echocardiographic measures that predicted mpap. The ratio of the diameter of the main pulmonary artery (dpa) to the diameter of the ascending aorta (daa) (regression coefficient, 23.6; P,.001) and echocardiography-derived RVSP (regression coefficient, 0.34; P,.001) were independently related to mpap. The best-fit equation was as follows: mpap = dpa/daa RVSP Radiology: Volume 254: Number 2 February 2010 n radiology.rsna.org 613

6 Figure 4 Figure 4: Scatterplots show relationship between mpap and (a) CT echocardiography composite index ( R 2 = 0.55) and (b) echocardiography-derived RVSP ( R 2 = 0.44). This equation (the CT echocardiography composite index) was more strongly related to mpap ( R 2 = 0.55; P,.001) (adjusted R 2 = 0.53) than CT or echocardiographic variables alone ( Table 3 ) ( Fig 4 ). Logistic regression analysis was used to compare the predictive abilities of echocardiography, CT, and the CT echocardiography composite index to distinguish between patients with and those without PH. A CT echocardiography composite index greater than 25 (equivalent to. 25 mm Hg) produced an odds ratio of 30.0 (95% confidence interval [CI]: 5.4, 166.9; P,.001). This was more than twice the values for widely used echocardiographic and CT signs of PH: The odds ratio of diagnosing PH with an estimated right ventricular systolic pressure greater than 35 mm Hg ( 24 ) was 13.7 (95% CI: 3.1, 60.8; P,.001), and the odds ratio for diagnosing PH with a ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta greater than 1 ( 11,14 ) was 9.0 (95% CI: 2.7, 30.6; P,.001). The sensitivity and specificity of a CT echocardiography composite value greater than 25 for detecting PH were 96% (42 of 44 patients with PH correctly identified) and 59% (10 of 17 without PH correctly identified), respectively. The AUC was 0.88 (95% CI: 0.77, 0.95), indicating excellent ability to discriminate between patients with and those without PH. Receiver operating characteristic analysis indicated that the best cutoff value for detecting PH was greater than 34 (sensitivity, 73%; specificity, 88%). By comparison, the AUC values for the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta and the values for echocardiography-derived RVSP were 0.82 (95% CI: 0.71, 0.9) and 0.83 (95% CI: 0.71, 0.91), respectively. Although the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta is a reliable marker of PH in patients with pulmonary fibrosis, it correlates less strongly with mpap than it does in patients without pulmonary fibrosis ( 13 ). Hence, we assessed the discriminative value of the CT echocardiography composite value greater than 25 in 30 patients with pulmonary fibrosis (of whom 24 had measurable tricuspid regurgitation at echocardiography). The sensitivity and specificity were 100% (14 of 14) and 50% (five of 10), respectively, and the AUC was 0.83 (95% CI: 0.62, 0.95) (compared with an AUC of 0.71 [95% CI: 0.51, 0.86] for the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta and an AUC of 0.75 [95% CI: 0.53, 0.9] for echocardiography-derived RVSP). There was no significant difference in the sensitivity and specificity of the CT echocardiography composite compared with those values in patients without pulmonary fibrosis (93% [28 of 30]) and 71% [five of seven], respectively; P..99 and P =.62, respectively). A subanalysis in patients in whom the time interval between CT and right-sided heart catheterization was 3 months or less (68 patients) also showed that the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta and echocardiography-derived RVSP were independently linked to mpap, and that the CT echocardiography composite index ( R 2 = 0.61; P,.001) was more strongly related to mpap than either of those two measures alone ( R 2 = 0.46, P,.001; R 2 = 0.49, P,.001, respectively). Discussion As a noninvasive test, CT is routinely performed in patients being investigated for a possible diagnosis of PH. It also has the potential to provide the first pointer toward the diagnosis of the condition ( 11,12,14 ). Echocardiography is more usually relied on to identify PH ( 17 ), although it also can substantially lead to an under- or overestimation of systolic pulmonary arterial pressure ( 25 ). This study has shown that although CT and echocardiographic measurements are both moderately strong correlates of mpap in patients with a spectrum of 614 radiology.rsna.org n Radiology: Volume 254: Number 2 February 2010

7 Limited studies have previously shown that enlargement of the segmental arteries, measured by making a subjective comparison with the size of the adjacent bronchus, occurs in patients with PH ( 12,14 ). A strength of our study was that we objectively measured absolute segmental arterial size by using electronic calipers in patients with PH. Although segmental arterial size was not more closely associated with PH than was the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta alone, the results from this study nevertheless suggest that segmental arterial size is a reliable marker of mpap. In practice, subjective assessment of segmental arterial size by comparison with the corresponding bronchus is more convenient. Our observations indicate not only that the degree of enlargement of the segmental artery bronchus ratio corresponds with the severity of PH but that this sign is highly reproducible. Multidetector CT enables the identification and evaluation of pulmonary arteries beyond the segmental level ( 26 ), although these smaller, more peripheral arteries have not previously been systematically examined in patients with PH. We found no statistically significant link between the diameter of the subsegmental pulmonary artery (increased or decreased) and mpap across a range of pulmonary arterial pressures. We speculate that this may, in part, be due to the reduction in peripheral vessel size that is seen only in patients with more severe PH ( 27 ). Limitations to this study included its retrospective nature and the fact that the study population was heterogeneous. The population was also biased toward patients with increased mpap and those in whom PH was suspected. Furthermore, although the numbers of patients with pulmonary vascular disease and pulmonary fibrosis were similar, certain conditions, such as chronic thromboembolic disease and emphysema, were underrepresented. Nonetheless, the population we examined is likely to reflect the type of patients who undergo right-sided heart catheterization in clinical practice. Moreover, hetunderlying disorders, using these tests in combination is considerably more powerful in determining increased pulmonary arterial pressure than either test in isolation. The implication is that CT and echocardiography are providing complementary information about increased pulmonary arterial pressure: CT offers anatomic information about the size of the pulmonary arterial tree, and echocardiography identifies the functional consequences of PH by evaluating secondary tricuspid regurgitation. Because right-sided heart catheterization is an invasive test, there is a need for accurate noninvasive markers that can be appropriately used to stratify patients for referral for right-sided heart catheterization. The composite CT echocardiography index, which was shown to reliably distinguish patients with and those without PH, has the potential to be used in such a manner. The ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta has previously been shown to correlate well with mpap ( 11 ). Our investigation has indicated that refining this sign with more precise methods to measure the main pulmonary artery (ie, obtaining the crosssectional area of the main pulmonary artery) is not helpful. It has been suggested that the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta is an accurate marker of mpap because confounding variables, such as patient size, influence the size of the pulmonary artery and ascending aorta equally, enabling a form of internal correction ( 11 ). Because the vertebral body is a fixed structure, the dimensions of which reflect a patient s overall body size ( 15 ), we postulated that the ratio of the diameter of the main pulmonary artery to the diameter of the thoracic vertebra might better reflect mpap. However, our results do not support this hypothesis. The size of the pulmonary artery is influenced by a variety of factors (only one of which is body size), and these factors are best reflected by the ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta. erogeneous populations are commonly used in studies examining the utility of noninvasive tests in PH ( 14,28 ). We also found the CT echocardiography composite to be useful in patients with pulmonary fibrosis, a scenario in which echocardiography and CT may be less reliable ( 7,13 ). Nevertheless, it would be useful to test the validity of the CT echocardiography composite index specifically in new groups of patients with various causes of PH. We also acknowledge that the time interval between CT and right-sided heart catheterization was long in some individuals. However, we found strong correlations despite this gap, and when the time interval between tests was restricted, the superior relationship between mpap and the CT echocardiography composite index was preserved. In conclusion, our study has shown that a combination of CT and echocardiographic markers of PH in the form of a composite index is more strongly related to mpap than either test in isolation. The study also highlights the usefulness of segmental arterial size as an indicator of PH, although across a spectrum of patients their relationship with mpap is not superior to that of the simple ratio of the diameter of the main pulmonary artery to the diameter of the ascending aorta alone. References 1. Hoeper MM. Definition, classification, and epidemiology of pulmonary arterial hypertension. Semin Respir Crit Care Med 2009 ; 30 ( 4 ): Coghlan JG, Handler C. Connective tissue associated pulmonary arterial hypertension. Lupus 2006 ; 15 ( 3 ): Lettieri CJ, Nathan SD, Barnett SD, Ahmad S, Shorr AF. Prevalence and outcomes of pulmonary arterial hypertension in advanced idiopathic pulmonary fibrosis. Chest 2006 ; 129 ( 3 ): Nunes H, Humbert M, Capron F, et al. Pulmonary hypertension associated with sarcoidosis: mechanisms, haemodynamics and prognosis. Thorax 2006 ; 61 ( 1 ): Rubin LJ. Diagnosis and management of pulmonary arterial hypertension: ACCP evidence-based clinical practice guidelines. Chest 2004 ; 126 ( 1 suppl ): 7S 10S. Radiology: Volume 254: Number 2 February 2010 n radiology.rsna.org 615

8 6. McGoon M, Gutterman D, Steen V, et al. Screening, early detection, and diagnosis of pulmonary arterial hypertension: ACCP evidence-based clinical practice guidelines. Chest 2004 ; 126 ( 1 suppl ): 14S 34S. 7. Arcasoy SM, Christie JD, Ferrari VA, et al. Echocardiographic assessment of pulmonary hypertension in patients with advanced lung disease. Am J Respir Crit Care Med 2003 ; 167 ( 5 ): Haimovici JB, Trotman-Dickenson B, Halpern EF, et al. Relationship between pulmonary artery diameter at computed tomography and pulmonary artery pressures at right-sided heart catheterization. Massachusetts General Hospital Lung Transplantation Program. Acad Radiol 1997 ; 4 ( 5 ): Kuriyama K, Gamsu G, Stern RG, Cann CE, Herfkens RJ, Brundage BH. CTdetermined pulmonary artery diameters in predicting pulmonary hypertension. Invest Radiol 1984 ; 19 ( 1 ): Moore NR, Scott JP, Flower CD, Higenbottam TW. The relationship between pulmonary artery pressure and pulmonary artery diameter in pulmonary hypertension. Clin Radiol 1988 ; 39 ( 5 ): Ng CS, Wells AU, Padley SPA. CT sign of chronic pulmonary arterial hypertension: the ratio of main pulmonary artery to aortic diameter. J Thorac Imaging 1999 ; 14 ( 4 ): Tan RT, Kuzo R, Goodman LR, Siegel R, Haasler GB, Presberg KW. Utility of CT scan evaluation for predicting pulmonary hypertension in patients with parenchymal lung disease. Medical College of Wisconsin Lung Transplant Group. Chest 1998 ; 113 ( 5 ): Devaraj A, Wells AU, Meister MG, Corte TJ, Hansell DM. The effect of diffuse pulmonary fibrosis on the reliability of CT signs of pulmonary hypertension. Radiology 2008 ; 249 ( 3 ): Perez-Enguix D, Morales P, Tomas JM, Vera F, Lloret RM. Computed tomographic screening of pulmonary arterial hypertension in candidates for lung transplantation. Transplant Proc 2007 ; 39 ( 7 ): Geraghty EM, Boone JM. Determination of height, weight, body mass index, and body surface area with a single abdominal CT image. Radiology 2003 ; 228 ( 3 ): Cook LB, Morgan M. Pulmonary artery catheterisation. Ann Acad Med Singapore 1994 ; 23 ( 4 ): Sciomer S, Magri D, Badagliacca R. Noninvasive assessment of pulmonary hypertension: Doppler-echocardiography. Pulm Pharmacol Ther 2007 ; 20 ( 2 ): Pepi M, Tamborini G, Galli C, et al. A new formula for echo-doppler estimation of right ventricular systolic pressure. J Am Soc Echocardiogr 1994 ; 7 ( 1 ): Nathan M, Pope MH, Grobler LJ. Osteophyte formation in the vertebral column: a review of the etiologic factors part II. Contemp Orthop 1994 ; 29 ( 2 ): Remy-Jardin M, Remy J, Artaud D, Deschildre F, Duhamel A. Peripheral pulmonary arteries: optimization of the spiral CT acquisition protocol. Radiology 1997 ; 204 ( 1 ): Coche E, Pawlak S, Dechambre S, Maldague B. Peripheral pulmonary arteries: identification at multi-slice spiral CT with 3D reconstruction. Eur Radiol 2003 ; 13 ( 4 ): Edwards PD, Bull RK, Coulden R. CT measurement of main pulmonary artery diameter. Br J Radiol 1998 ; 71 ( 850 ): Kim SJ, Im JG, Kim IO, et al. Normal bronchial and pulmonary arterial diameters measured by thin section CT. J Comput Assist Tomogr 1995 ; 19 ( 3 ): Barst RJ, McGoon M, Torbicki A, et al. Diagnosis and differential assessment of pulmonary arterial hypertension. J Am Coll Cardiol 2004 ; 43 ( 12 suppl S ): 40S 47S. 25. Homma A, Anzueto A, Peters JI, et al. Pulmonary artery systolic pressures estimated by echocardiogram vs cardiac catheterization in patients awaiting lung transplantation. J Heart Lung Transplant 2001 ; 20 ( 8 ): Ghaye B, Szapiro D, Mastora I, et al. Peripheral pulmonary arteries: how far in the lung does multi-detector row spiral CT allow analysis? Radiology 2001 ; 219 ( 3 ): Chrispin AR, Goodwin JF, Steiner RE. The radiology of obliterative pulmonary hypertension and thrombo-embolism. Br J Radiol 1963 ; 36 : Selimovic N, Andersson B, Bergh CH, et al. Pulmonary hemodynamics as predictors of mortality in patients awaiting lung transplantation. Transpl Int 2008 ; 21 ( 4 ): radiology.rsna.org n Radiology: Volume 254: Number 2 February 2010