Long-term impact of myocardial inflammation on quantitative myocardial perfusion-a descriptive PET/MR myocarditis study

Abstract

Purpose: Whether myocardial inflammation causes long-term sequelae potentially affecting myocardial blood flow (MBF) is unknown. We aimed to assess the effect of myocardial inflammation on quantitative MBF parameters, as assessed by 13N-ammonia positron emission tomography myocardial perfusion imaging (PET-MPI) late after myocarditis.

Methods: Fifty patients with a history of myocarditis underwent cardiac magnetic resonance (CMR) imaging at diagnosis and PET/MR imaging at follow-up at least 6 months later. Segmental MBF, myocardial flow reserve (MFR), and 13N-ammonia washout were obtained from PET, and segments with reduced 13N-ammonia retention, resembling scar, were recorded. Based on CMR, segments were classified as remote (n = 469), healed (inflammation at baseline but no late gadolinium enhancement [LGE] at follow-up, n = 118), and scarred (LGE at follow-up, n = 72). Additionally, apparently healed segments but with scar at PET were classified as PET discordant (n = 18).

Results: Compared to remote segments, healed segments showed higher stress MBF (2.71 mL*min^-1*g^-1 [IQR 2.18-3.08] vs. 2.20 mL*min^-1*g^-1 [1.75-2.68], p < 0.0001), MFR (3.78 [2.83-4.79] vs. 3.36 [2.60-4.03], p < 0.0001), and washout (rest 0.24/min [0.18-0.31] and stress 0.53/min [0.40-0.67] vs. 0.22/min [0.16-0.27] and 0.46/min [0.32-0.63], p = 0.010 and p = 0.021, respectively). While PET discordant segments did not differ from healed segments regarding MBF and MFR, washout was higher by ~ 30% (p < 0.014). Finally, 10 (20%) patients were diagnosed by PET-MPI as presenting with a myocardial scar but without a corresponding LGE.

Conclusion: In patients with a history of myocarditis, quantitative measurements of myocardial perfusion as obtained from PET-MPI remain altered in areas initially affected by inflammation. CMR = cardiac magnetic resonance; PET = positron emission tomography; LGE = late gadolinium enhancement.

Keywords: Cardiac magnetic resonance; Inflammation; Myocardial blood flow; Myocarditis; Perfusion; Positron emission tomography.

Fig. 1

Hybrid PET/MR imaging acquisition protocol employed at follow-up

Fig. 2

An imaging study of a 22-year-old male with a history of myocarditis. The dotted line in (A) depicts the cross-section of the selected slices provided in the top row. CMR at baseline revealed patchy and subepicardial LGE in the anterolateral, lateral, inferolateral, and inferior wall of the basal to mid-ventricular LV myocardium (B) with areas of T2 hyperintensity in the same locations (C). Polar plots depicting segmental percent enhancement of LGE (D) and T2 (E) provide an overview of the affected segments. PET/MR was performed 32 months later, with CMR revealing no residual LGE in any segments (F, G). By contrast, 13N-ammonia PET showed reduced radionuclide retention at rest in the inferolateral basal and in the lateral LV myocardium (H, I). In parallel, the 13N-ammonia washout rate (J) was increased in the lateral and inferolaterobasal LV walls

Fig. 3

Boxplots of MBF values from rest (gray boxes) and stress (blue boxes) 13N-ammonia PET imaging stratified according to the different groups of segments. Boxes depict the interquartile range (25th to 75th quartile) and contain a horizontal line depicting median values. Statistically significant differences between values from different groups are indicated by horizontal lines with whiskers and the corresponding p-values

Fig. 4

Boxplots of MFR values from 13N-ammonia PET imaging stratified according to the different groups of segments

Fig. 5

Boxplots of K2 values from rest (gray boxes) and stress (blue boxes) 13N-ammonia PET imaging stratified according to the different groups of segments

Fig. 6

An imaging study of a 48-year-old male patient with a history of myocarditis diagnosed 7 years prior. CMR at baseline depicted a subtle area with mid-wall LGE in the anteroapical LV myocardium, as shown in the short-axis view (A). At follow-up, no LGE was depictable in this region (B), indicating the initial LGE to be due to inflammation. However, PET retention images showed reduced radionuclide retention in the same anteroapical region (C). Additional areas with subepicardial LGE at baseline, as shown in the three-chamber (D) and two-chamber (E) views were persistent at follow-up in the inferolaterobasal wall but not in the lateral wall (F, G). As depicted by the polar plots of PET retention data at follow-up (H), both areas showed reduced radionuclide retention. Finally, the polar plot depicting the 13N-ammonia washout rate (I) reveals increased values in the entire lateral and inferolateral wall extending to the anteroapical region