Introduction

Pulmonary hypertension (PH) is a severe condition with a historically poor prognosis (1). With certain disease-modifying therapies and traditional therapies, and with heart-lung transplantation, the outcome has been altered radically in certain subgroups of patients. Chronic thromboembolic pulmonary hypertension (CTEPH) is a cause of PH that is potentially treatable with anticoagulation and pulmonary endarterectomy (2). Therefore, the American College of Cardiology Foundation/American Heart Association consensus (3) and the European Society of Cardiology guideline (2) both recommend ventilation/perfusion (V/Q) scan as a pivotal test in the initial assessment to screen for CTEPH. Further investigations such as computed tomography (CT) or conventional pulmonary angiography are then indicated if V/Q scan result is positive (2). In PH, planar V/Q scintigraphy has 97.4% sensitivity compared with 51% for CT pulmonary angiography (CTPA) for detection of chronic thromboembolic disease (4).

The advent of V/Q single-photon emission computed tomography (SPECT) has overcome some of the limitations of planar V/Q imaging, by offering even higher sensitivity (80% to 100%), specificity (93% to 100%), and accuracy (94%) in thromboembolic diseases (5). Compared with planar V/Q scans, the 3-dimensional tomography images from SPECT allow more accurate diagnosis of thromboembolic disease while differentiating other lung disorders such as chronic obstructive pulmonary disease, interstitial lung disease, and veno-occlusive disease (6). The European Association of Nuclear Medicine (EANM) established diagnostic criteria for interpreting thromboembolic disease on the basis of focal V/Q mismatches in segmental and subsegmental pulmonary arteries (7).

It is, however, recognized that patients with abnormal V/Q SPECT findings sometimes may not have significant thromboembolic disease demonstrated on confirmatory pulmonary angiography (8). Some patients also have abnormal patchy and heterogeneous perfusion defects on V/Q SPECT images that are difficult to classify (6,8,9). The functional significance of these abnormal V/Q SPECT results in pulmonary arterial hypertension (PAH) remains unclear. We sought to evaluate the incidence and clinical significance of abnormal V/Q SPECT findings in the context of patients with PAH who had no detectable thromboembolism observed during corroborating investigations.

Methods

Patient cohort and characterization of PH

Patients referred to the National Pulmonary Hypertension service at the Royal Free NHS Foundation Trust in London, United Kingdom typically undergo V/Q scanning at the Royal Free Hospital if this has not been previously performed. Patients who underwent V/Q scan subsequent to 2013 were studied because SPECT V/Q was introduced at that date. We included all patients who had SPECT V/Q scanning between 2013 and January 2016 at the Royal Free Hospital within 3 months of their initial diagnostic right-sided heart catheterization (RHC) in this study. The study complied with the Declaration of Helsinki and was approved by the Institutional Review Board.

RHC was performed according to European Society of Cardiology guidelines (2). PH is defined as mean pulmonary arterial pressure ≥25 mm Hg on RHC (2). The cause of PH was defined according to World Health Organization classifications (10). Echocardiography was used to assess left-sided heart disease. Spirometry and CT scans of the chest were performed in all patients. We excluded patients clinically diagnosed with significant left-sided heart and lung disease.

V/Q SPECT were interpreted according to EANM criteria, and these patients were then screened for thromboembolic disease (Figure 1). The 6-min walking distance and functional class were assessed in all patients. We obtained data on whether therapy with pulmonary vasodilators was initiated and on whether the patient was treated with anticoagulation. The patients were followed up after V/Q SPECT, and RHC and all-cause mortality were recorded.

V/Q SPECT

V/Q SPECT scans were performed as part of the initial assessment. The scans were acquired using triple-head gamma camera (Philips Irix, Eindhoven, the Netherlands) equipped with low-energy, high-resolution, and parallel-hole collimators. Perfusion studies were performed after injection of 100 MBq of 99m-technetium (^99mTc)-macroaggregated albumin (^99mTc-MAA, MAA, CIS Bio International, Saclay, France; and 99mTc, Mallinckrodt Medical, Staines-upon-Thames, United Kingdom) during 2 respiratory cycles. Ventilation scans were performed during quiet tidal inhalation of 81 mKr (190 keV, Sandwell and West Birmingham NHS Trust, United Kingdom) or ^99mTc-Technegas (140 keV, Cyclomedica, Kingsgrove, Australia). All scans were clinically reported by experienced nuclear medicine physicians at the Royal Free Hospital on the basis of clinical indications provided by the treating physicians at the time of the scan. “Positive scans” are those that meet the EANM reporting criteria for pulmonary thromboembolic disease (7). All scans were subsequently reviewed for the purposes of this study by 1 of the authors (M.H.) blinded to all clinical information. The scans were then reanalyzed according to the pattern of perfusion defect into: 1) normal scans with no perfusion defects; 2) focal perfusion defects that indicate segmental or subsegmental obstruction but not fulfilling EANM criteria for thromboembolism; and 3) global perfusion defects. Where there was a discrepancy between the 2 reports, the blinded report was used.

Statistical analysis

Continuous variables are summarized using the mean ± SE for normally distributed data and medians for non-normally distributed data. Categorical variables are described as a percentage and analyzed using the chi-square test or the Fisher exact test where appropriate. Linear regression model was used to compare between patients with abnormal V/Q SPECT but no thromboembolism and those with normal V/Q SPECT scans. For comparisons among 3 groups (normal, focal defect, and global defects on V/Q SPECT), 1-way analysis of variance test was used with Bonferroni post hoc analysis on significant associations. Analysis of covariance test was used for adjusted analyses. All analyses were performed with SPSS software version 17.0 (SPSS, Inc., Chicago, Illinois), and GraphPad Prism 6.0 (La Jolla, California).

Results

High incidence of positive V/Q SPECT in nonthromboembolic PAH

A total of 244 patients underwent V/Q SPECT and RHC for PH assessment between 2013 and 2016; of these, 136 patients had PH confirmed on RHC. Suspected thromboembolism on V/Q SPECT was reported in 44 patients according to EANM criteria. After further corroborating investigations including CTPA-confirmed positive thromboembolism in 10 patients, and 10 patients had previous established diagnosis of CTEPH, 24 patients with no or negative CTPA results further underwent conventional pulmonary angiogram, which identified thromboembolism in 5 patients. This gives a total of 19 patients with positive V/Q SPECT but no thromboembolic disease detectable on conventional invasive pulmonary angiography, and 15 (78.9%) of these patients had group I PAH (Figure 1 and Table 1). Among those patients with negative V/Q SPECT, 60 (65.2%) patients had group I PAH. A total of 27 (29.3%) patients had anticoagulation at baseline for other comorbidities such as atrial fibrillation. There were no significant demographic differences or differences in pulmonary hemodynamics or baseline function (functional class or 6-min walk distance) between those patients with positive V/Q SPECT findings and those without (Figure 2).

Perfusion defects on V/Q SPECT are associated with hemodynamic severity

The V/Q SPECT imaging of the 75 patients with group I PAH were reanalyzed according to the pattern of perfusion defects (Figure 3). There were minimal discrepancies with an interclass correlation coefficient of 0.98 (95% confidence interval: 0.97 to 0.99). Most patients with focal perfusion defects had subsegmental mismatches, which are nondiagnostic for thromboembolism. The remaining patients with focal or global perfusion defects on V/Q SPECT all had CT pulmonary angiograms and targeted invasive pulmonary angiography in selected cases to exclude distal thromboembolisms. Compared with those patients with normal V/Q SPECT findings, there was more group I PAH in patients with global perfusion defects (chi-square test; p = 0.022), and in particular there were more patients with systemic sclerosis (Fisher exact test; p = 0.019). Among patients with group I PAH, analysis of variance testing among the 3 groups of perfusion defects showed significant differences in mPAP (F = 9.93; p = 0.0002), and pulmonary vascular resistance (PVR) (F = 5.9; p = 0.004) (Table 2 and Figure 4). The associations persisted after adjustment for multiple comparison with Bonferroni corrections. Post hoc analysis showed no significant differences between patients with normal V/Q and focal V/Q defects for both mPAP (p = 0.18) and PVR (p = 0.60). The associations remained significant after adjusting for vasodilator and baseline anticoagulation therapy, with mPAP (F = 7.48; p = 0.0011) and PVR (F = 3.80; p = 0.027), as well as after Bonferroni corrections. The associations were similar in a subgroup analysis focusing on patients with systemic sclerosis, mPAP (F = 8.27; p = 0.001), and PVR (F = 4.12; p = 0.023).

High mortality rate in group I PAH with global perfusion defects on V/Q SPECT

Patients were followed up for a median of 372.7 ± 200.0 days. Death of any cause (all-cause mortality) occurred in 6 patients with group I PAH, and all had global perfusion defects on V/Q SPECT. The underlying diagnosis of these patients included 3 with idiopathic PAH, 2 with systemic sclerosis, and 1 with another connective tissue disorder. Among patients with group I PAH, those with global perfusion defects on V/Q SPECT had significantly worse survival rate compared with those with normal or focal perfusion defects on V/Q SPECT (log-rank p = 0.037), with a hazard ratio of 5.63 (95% confidence interval: 1.11 to 28.50) calculated with the Mantel Haenszel approach described by Bernstein et al. (11) (Online Figures 1 to 3).

Discussion

There was a high incidence (43.1%) of patients with PAH in our cohort with positive V/Q SPECT scan results but no thromboembolism demonstrable in corroborating investigations. After diagnostic screening in the 19 patients with positive V/Q SPECT findings, all had no detectable thromboembolism on conventional invasive pulmonary angiography, the current gold standard (12,13). None of the patients with negative V/Q SPECT results subsequently had thromboembolism detected with other modalities of investigations, a finding that demonstrates that the high negative predictive value of a normal V/Q scan also pertains to patients with PAH.

We found a false-positive rate of 43% (specificity 67%) among patients with PAH. A previous study looked at a mixed group of causes of PAH and found a lower false-positive rate of 5% to 10% when using planar V/Q scanning (4). Our study used V/Q SPECT, which offers better accuracy for detecting segmental and subsegmental pulmonary embolisms (5).

PAH is associated with pruning of pulmonary vessels, and in early nuclear medicine studies patchy perfusion defects were described (8,9). Attempts were also made to quantify the heterogeneity of perfusion defects, and this was shown to be predictive of pulmonary systolic pressures (14). We have found that these perfusion defects are often reported as pulmonary emboli by using current, sensitive V/Q SPECT scanning and contemporary reporting criteria.

Global perfusion defects on V/Q SPECT scanning were associated with a severe PAH phenotype, with more adverse pulmonary hemodynamics and higher mortality rates. This finding suggests that the V/Q scan may have utility beyond the exclusion of pulmonary thromboembolic disease. In particular, such global perfusion defects could reflect obliteration of distal pulmonary vessels, a pathological feature that anticipates progression to right ventricular failure (13,15). This could offer a noninvasive assessment to identify patients with more severe PAH. Prospective studies would be helpful to establish whether perfusion defects on V/Q scanning provide independent prognostic information and whether they may identify patients who would benefit from specific therapies.

Study limitations

The limitations of this study follow from its observational single-center design. Although invasive pulmonary angiograms were not routinely performed in all patients, we have described as false positive only those patients with negative results on conventional pulmonary angiograms and as true positive only patients whose CTEPH cases were confirmed with either CTPA or conventional pulmonary angiograms. The subgroup analysis included patients with group I PAH, who are still a heterogeneous group but nevertheless all have arterial stiffness or vasoconstriction. The correlations between V/Q SPECT and severity of PAH were also consistent when only patients with systemic sclerosis were included. The associations with PVR gave further mechanistic insight that V/Q SPECT images could be related to arterial disorders. Further studies of V/Q SPECT in subtypes of group I PAH are needed to investigate the correlation with the underlying pulmonary arterial disease. A larger series of patients studied prospectively will also be helpful to confirm the worse outcome of patients with global perfusion defects on V/Q SPECT.

Conclusions

Taken together, PAH is associated with frequent perfusion abnormalities detectable on V/Q SPECT scanning that are not the result of thromboembolic disease and that may reflect the severity of the disease as evidenced by worse pulmonary hemodynamics and poorer prognosis. Further prospective study of these associations is merited.

Footnotes