Abstract
Pulmonary artery indexed for body surface area is the most accurate predictor for PH on CT in pulmonary sarcoidosis http://ow.ly/T863m
To the Editor:
Pulmonary hypertension (PH) is a known complication of pulmonary sarcoidosis with a prevalence ranging from 5% to 74% [1]. The aetiology of PH in sarcoidosis is not fully understood. Usually, it is attributed to the destruction of the distal capillary bed by lung fibrosis and/or chronic hypoxaemia. However, the severity of PH does not correlate consistently with the degree of pulmonary fibrosis, and PH exists in sarcoidosis patients without fibrosis, suggesting a multifactorial mechanism. The presence of PH is associated with a poor prognosis, and early diagnosis and treatment might improve outcome [1]. Echocardiography should always be performed when PH is suspected [2]. However, the accuracy of echocardiography in patients with interstitial lung diseases is often limited due to poor image quality and unreliable tricuspid regurgitation signal to measure the right ventricular systolic pressure (RVSP) [3]. Further invasive investigation with the gold standard, right heart catheterisation (RHC), is often required. In order to optimise the noninvasive diagnostic approach, there is a need for more accurate predictors of PH. Computed tomography (CT) may raise suspicion of PH in symptomatic patients or those examined for unrelated indications by showing an increased pulmonary artery (PA) diameter (≥29 mm) and PA diameter/ascending aorta diameter (AAD) ratio (≥1.0) [2]. Similarly, PA diameter indexed to body surface area (BSA) has been suggested as possible predictor of PH. However, these parameters have never been investigated in pulmonary sarcoidosis specifically.
In this study, PA diameter measurements on chest CT were retrospectively evaluated as predictors of PH. Patients suspected of PH and referred for analysis between November 2007 and May 2014 were included in cases with a consensus diagnosis of pulmonary sarcoidosis, aged ≥18 years, with availability of chest CT within 1 year of PH analysis. The analysis protocol was based on the European PH guideline [2], and consisted of an ECG, laboratory testing, and echocardiographic assessment of RVSP and secondary parameters. Subsequently, patients were classified as “PH likely” (RVSP >50 mmHg), “PH possible” (RVSP of 36–50 mmHg or presence of secondary signs with normal/absent RVSP signal) or “PH unlikely” (RVSP <36 mmHg or absence of signal without secondary signs). RHC was performed if PH was possible or likely. PH was defined as an invasively measured mean PA pressure (mPAP) ≥25 mmHg. Patients were divided into three groups: “PH confirmed by RHC”, “no PH confirmed by RHC” and “no PH based on echocardiography”.
A radiologist, specialised in interstitial lung diseases and blinded to haemodynamic data, reviewed the chest CT performed most recently to the echocardiographic analysis. The PA diameter was measured at the level of the bifurcation along the line from the centre of the adjacent ascending aorta perpendicular to the main PA axis [4]. As a derivative, the PA diameter was indexed to BSA [5] and calculated as a ratio with AAD. Furthermore, the percentage of fibrosis of the total lung area was described as “not significant” (<5%), “intermediate” (5–20%) or “severe” (>20%).
In total, 89 out of 103 pulmonary sarcoidosis patients referred for PH analysis were eligible for inclusion. The estimated prevalence of PH was 28.1%. Table 1 shows the baseline characteristics and outcomes. 16 patients had been referred as a result of a dilated PA. After excluding these patients, PA measurements remained highly significant. PA diameter indexed to BSA and PA diameter/AAD ratio correlated moderately with RHC-derived mPAP (r=0.51 and r=0.52, respectively).
Receiver operating curve analysis showed a good accuracy for PA diameter indexed to BSA and PA diameter/AAD ratio (area under the curve (AUC) of 0.88 and 0.81, respectively; p<0.001). Optimal cut-off values were calculated for PA diameter (30.6 mm), PA diameter indexed to BSA (16.02 mm·m−2) and PA diameter/AAD ratio (1.06). PA diameter indexed to BSA showed the best positive and negative predictive value for diagnosing PH (70% and 93.2%, respectively) followed by PA diameter/AAD ratio (64.3% and 88.5%, respectively).
In a subanalysis, we evaluated patients classified as “PH possible” (n=18). Only PA diameter indexed to BSA was able to significantly discriminate between the presence and absence of PH (17.0 versus 14.3 mm·m−2, respectively; p<0.006), with a high diagnostic accuracy (AUC 0.91) using a cut-off value of 15.2 mm·m−2. This was notably higher than PA diameter/AAD ratio (AUC 0.71).
Finally, we evaluated the correlation of PA measurements with the presence of fibrosis. PA diameter and PA diameter indexed to BSA had a weak but significant correlation with the extent of fibrosis (r=0.32 and r=0.42, respectively)
This is the first study to investigate chest CT parameters for predicting PH only in patients with pulmonary sarcoidosis. The results implicate that PA diameter indexed to BSA is the most reliable predictor of PH for both the general pulmonary sarcoidosis population and for the “PH possible” subgroup, classified by echocardiography. Therefore, PA measurement on chest CT might be valuable to further optimise PH analysis in pulmonary sarcoidosis.
Besides hypoxaemia, mechanisms for PH in sarcoidosis include specific vasculopathy of the vessel wall, perivascular fibrosis, external compression of the PA by severe lymphadenopathy, mediastinal fibrosis and left heart disease [1]. PA dilatation in PH can be due to chronically elevated PA pressure [6]. Similarly to aortic diameter, PA size may vary among individuals, depending on age, sex and BSA [7]. It has been suggested that indexing PA to BSA might lead to a higher accuracy that PA diameter/AAD ratio, since BSA influences both PA and aortic diameters [8].
Many studies describe positive correlations for mPAP with PA measurements in different subgroups of PH, especially for PA diameter indexed to BSA and PA diameter/AAD ratio [9]. Only few studies assessed these parameters in interstitial lung diseases, most reporting positive correlations [10–12]. Unfortunately, PA diameter indexed to BSA has not been evaluated in interstitial lung disease. Furthermore, two studies were unable to obtain positive correlations. Zisman et al. [13] studied 65 idiopathic pulmonary fibrosis (IPF) patients and found no significant correlation of PA diameters with mPAP, including PA diameter indexed to BSA. However, in that study, a mean mPAP of 26.1 mmHg might have been too low in order to cause dilatation. Similarly, Devaraj et al. [14] found no correlation of mPAP with PA diameter or PA diameter/AAD ratio in 30 patients with pulmonary fibrosis, of whom 11 had sarcoidosis. PA dilatation occurred in the absence of significant PH; however, this was unrelated to the extent of fibrosis. Our study involves the largest group of sarcoidosis patients studied and supports the hypothesis that the mechanism of PH in sarcoidosis is not fully explained by pulmonary fibrosis. Compared to patients with IPF, pulmonary fibrosis in sarcoidosis is characterised by a chronic course, often involving the upper lobe, and is predominantly of the nonhoneycombing type. Conversely, IPF progresses quickly, is mostly located peripherally or in the lower lobe, with a predominance of honeycombing [15]. This might lead to increased traction on the PA in IPF. For PH in sarcoidosis, the exact mechanism of PH and PA dilatation remains unknown, and requires further investigation.
Measurements of PA diameter and AAD are highly reproducible, with an excellent inter- and intraobserver variability [7]. Limitations of this study were mostly due to its retrospective nature, resulting in different types of CT and echocardiography equipment with variable time intervals. Second, not all patients underwent RHC, which might have led to underdetection of PH. Last, this study was conducted in a tertiary care centre.
In conclusion, our study demonstrates that measurement of PA diameter indexed to BSA is the most accurate predictor of PH on chest CT in patients with pulmonary sarcoidosis. Prospective research in a large, predefined group of sarcoidosis patients is warranted for further validation and clinical implementation.
Footnotes
Conflict of interest: None declared.
- Received March 25, 2015.
- Accepted September 16, 2015.
- Copyright ©ERS 2016