Right ventricular strain analysis with positron emission tomography-derived feature tracking: Direct comparison with...

Positron emission tomography (PET) provides robust assessment of cardiac function, whereas evaluation of the right ventricle (RV) remains challenging. Myocardial strain analysis has emerged as a sensitive marker for subtle wall dysfunction. We aimed to evaluate the feasibility of RV strain quantification using a newly developed PET feature tracking technique, with magnetic resonance (MR) strain as the reference standard.

Consecutive 123 patients (mean age: 68 years; males: 101) who underwent rest ¹³N ammonia PET/MR were retrospectively enrolled. Semiautomatic PET feature tracking was performed using previously validated in-house MATLAB-based software. RV global longitudinal strain (GLS %) and circumferential strain (GCS %) were derived from 4-chamber and short-axis gated PET cine images, respectively. For comparison, reference values were quantified from simultaneously acquired breath-hold cine MR using commercially available software.

Mean GLS and GCS were -24.1 ± 6.9% vs -23.4 ± 5.6% and -16.2 ± 5.4% vs -16.6 ± 4.2% for PET and cardiovascular magnetic resonance imaging, respectively. Both parameters showed strong correlations between the two modalities (r = 0.77 and 0.74 for GLS and GCS, respectively; P < 0.0001). Bland-Altman analysis demonstrated acceptable agreement, despite the presence of significant proportional bias (GLS: bias 0.1 ± 4.4% [limits of agreement: -8.6 to 8.8], r = 0.2, and P = 0.01; GCS: bias of -0.7 ± 3.6% [limits of agreement: -7.8 to 6.4], r = -0.4, and P < 0.0001, respectively).

PET-derived RV wall strain analysis demonstrated a significant correlation with MR strain. PET feature tracking has the potential as a useful tool to detect RV wall dysfunction in routine clinical practice.