Introduction

Cardiac magnetic resonance (CMR) imaging allows for noninvasive assessment of myocardial edema, fibrosis, or infiltration.^1,2 Presence of late gadolinium enhancement (LGE) in myocardial walls has been found in a majority of patients with cardiac diseases, and is often indicative of fibrosis induced by chronic overpressure and / or overload of the heart.^3-6 In comparison with typical LGE localization in ventricular free walls after an ischemic event, LGE in pulmonary arterial hypertension (PAH) and hypertrophic cardiomyopathy is mostly found at junctions of the interventricular septum and right ventricular (RV) free walls, referred to as RV insertion points (RVIPs).^7,8

Inflammatory processes occurring in pulmonary vessels play an important role in PAH pathophysiology.^9,10 We have previously confirmed that in the case of RV failure and hemodynamic impairment due to PAH, changes in the levels of circulating cytokines (eg, interleukin 6 or stromal cell–derived factor 1α) are related to myocardial wall glucose metabolism alterations observed on hybrid positron emission tomography / magnetic resonance imaging (PET/MRI).¹¹

We have also introduced a new parameter called LGE mass index (quantitative assessment of LGE mass at RVIPs divided by a patient’s body surface area), which strongly correlated with RV dysfunction and predicted prognosis in individuals with PAH,⁴ as opposed to the sole qualitative assessment (presence of LGE at RVIPs) performed a few years earlier by Swift et al¹² in a group of PAH patients. Qualitative and / or quantitative assessment of LGE on cardiac imaging varies depending on natural PAH progression or specific therapy. Therefore, we hypothesized that if the presence of LGE at RVIPs is equivalent to an abundance of fibrotic tissue (ie, tissue with low metabolic activity), it should affect local glucose uptake at RVIPs. To simultaneously evaluate myocardial fibrosis and glucose metabolism at RVIPs and to avoid time shift in the scans we used hybrid imaging—CMR and ¹⁸F‑fluorodeoxyglucose (FDG) PET.

The main aims of the study were 1) to evaluate glucose metabolism in the RVIPs at which LGE was qualitatively described on CMR imaging, and 2) to assess whether the LGE pattern, reflective of local cardiac tissue changes occurring at RVIPs, may alter in the course of the disease and following initiation of advanced targeted PAH treatment.

Patients and methods

Ethics

The study was approved by the local Bioethics Committee of the Medical University of Bialystok, Poland (R‑I‑002/140/2017), and all patients provided written informed consent to participate.

Results

General patient characteristics at baseline and follow‑up, including hemodynamic parameters derived from RHC and CMR results, are presented in Table 1 (modified from our previous publication⁴). Most of the PAH patients (n = 12; 60%) were at an intermediate risk of 1‑year mortality according to the 2015 European Society of Cardiology guidelines.¹⁵ Individuals with idiopathic / heritable PAH accounted for 60% of the group (n = 12). Baseline mean (SD) pulmonary artery pressure (mPAP) was 50.5 (18.3) mm Hg, and mean (SD) right ventricular ejection fraction (RVEF) was 45.1% (9.6%). LGE at RVIPs was present in the entire study group. LGE did not involve the interventricular septum in any patient. Median (IQR) baseline LGE mass at RVIPs was 5.4 (2.3–9.4) g and median (IQR) SUV_RVIPS was 6.33 (2.5–9.9).

Follow‑up visits were scheduled 24 months after the baseline visit. A total of 20 PAH patients completed the follow‑up and their data were compared with data of the corresponding baseline group (20 matched pairs). At follow‑up, we observed a significant improvement of MRI‑derived RVEF (mean [SD], 45.1% (9.6%) vs 52.4% (12.9%) at baseline; P = 0.01) and of hemodynamic parameters obtained from RHC, such as mPAP (mean [SD], 50.5 [18.3] mm Hg vs 42.8 [18.6] mm Hg at baseline; P = 0.03) and pulmonary vascular resistance (PVR; mean [SD], 8.9 [5.7] Wood units vs 7.3 [4.7] Wood units at baseline; P = 0.04) (Table 1).

During the follow‑up period, 16 patients (80%) experienced a clinical end point (death, n = 4; clinical symptoms of PAH progression, n = 12). There was a clinical trend toward a decrease in SUV_RVIPS on follow‑up PET scans (median [IQR], 5.18 [3.3–7.7] vs 6.33 [2.5–9.9] at baseline; mean change, –1.48; P = 0.16). Follow‑up scans also revealed a clinical tendency for median LGE mass to increase (median [IQR] 6.3 [3.4–11.4] g vs 5.4 [2.3–9.4] g at baseline; mean change, 1.26 g; P = 0.27).

No significant correlation was found between LGE mass and SUV_RVIPS neither on baseline nor on follow‑up scans (R = –0.05; P = 0.76 and R = –0.08; P = 0.72, respectively). There was no association between the change in LGE between the baseline and follow‑up and change in SUV_RVIPS during the follow‑up. These data suggest that the extent of LGE is not associated with lower metabolism in RVIPs, undermining the hypothesis that LGE at RVIPs reflects fibrosis in PAH patients.

Discussion

This is the first study involving a quantitative analysis of myocardial fibrosis and active inflammation (and their relationship) with a use of hybrid imaging. We have previously confirmed that early determination of RV dysfunction in PAH, especially before clinical worsening, is possible noninvasively with the use of PET/MRI imaging.¹⁶ This hybrid technique also allows for assessment of glucose metabolism in the RV free wall, which was previously linked to the prognosis of PAH patients.¹⁴ In this study, we found that PET/MRI may be a helpful tool to verify previous assumptions about cardiac LGE in PAH, which changes the perspective of routine assessment of LGE presence.

LGE at RVIPs in severely ill PAH patients seems to be a nearly universal finding (irrespective of PAH etiology), and was often interpreted as replacement fibrosis.¹⁷ Importantly, it may also be found in apparently healthy individuals.⁴ Inflammation is also a major regulator of reparative response in tissue damage, but prolonged and uncontrolled remodeling of extracellular matrix often leads to deterioration of cardiac function and poor prognosis.¹⁸ In this study, we confirmed that local cardiac tissue lesions in PAH visualized as LGE on CMR may be indeed still metabolically active and susceptible to alterations resulting from the applied PAH therapy or disease progression.

We hypothesize that what is considered LGE may in fact be associated with local inflammation. This underlines the importance of research into the role of inflammation in PAH pathogenesis, which may result in development of new immunomodulatory therapies targeting not only affected pulmonary vessels but also the cardiac muscle.

Inflammation may become no longer active when the myocardium shows advanced fibrosis and severely dyskinetic, thinned wall. Thus, areas with fibrotic tissue tend to have poor metabolism (with decreased or minimal FDG uptake) on PET imaging; however, some discrepancies in this regard were also observed.¹⁸ Mismatch between a visual assessment of fibrotic tissue in the form of CMR LGE and sustained or even increased metabolism is probably due to dominance of the diffuse (interstitial) fibrosis type over the typical replacement fibrosis seen in the damaged tissue. Thus, the prognostic value of LGE presence on CMR may be limited for the assessment of diffuse interstitial fibrosis, which is often found in patients with hypertrophic or dilated cardiomyopathy.

There is still little evidence to confirm which type of fibrosis may be found at RVIPs of patients with PAH. In our study population, all patients had an LGE pattern at the insertion points. In other studies, performed on bigger populations with a greater spectrum of disease severity, LGE was not so often found at RVIPs of individuals with mild or early‑stage PAH.⁷ It seems that local tissue alterations occurring in RVIPs are strongly related to progression of the disease. Thus, the exact cause of LGE presence at RVIPs in patients with early‑stage PAH or without RV dysfunction is still to be elucidated.¹⁹

In this study, we observed a variability of LGE mass at RVIPs of individual PAH patients. If LGE truly indicated the presence of dead / fibrotic tissue, we should have obtained a strong negative correlation with glucose uptake at this site, which would reflect poor metabolism of fibrotic tissue. The lack of such an association suggests the presence of diffuse interstitial fibrosis with potential local inflammation, rather than the typical replacement type.

Anderson et al²⁰ described 3 different forms of cardiac fibrosis: reactive (interstitial; preserving cardiac structure and function), perivascular, and replacement fibrosis. They suggested that after a period of reactive interstitial fibrosis (during which the heart adapts to hemodynamic changes), replacement fibrosis may initiate upon cardiomyocyte death.²⁰ It seems that in treated PAH patients with high PVR and mPAP, LGE on CMR imaging indicates the dynamic, interstitial type of cardiac fibrosis. Thus, presence of LGE at RVIPs should not be simply interpreted as necrotic tissue in this group of patients.