Advanced cardiac magnetic resonance techniques for the detection of severe myocardial fibrosis in severe aortic stenosis
R Kockova, T Adla, D Sedmera, M Kocka, A Praveckova, P Kacer, L Sukupova, E HonsovaAbstract
Background
Myocardial fibrosis is a key determinant of adverse outcomes in patients with severe aortic stenosis (1); however, its reliable noninvasive detection remains challenging.
Purpose
To evaluate the diagnostic performance of cardiac magnetic resonance (CMR)–derived myocardial strain and tissue characterization parameters for identifying severe myocardial fibrosis using histology as the reference standard.
Methods
Thirty-one patients with severe aortic stenosis meeting guideline criteria for surgical valve replacement were prospectively enrolled. All patients underwent comprehensive CMR, including feature-tracking–derived global longitudinal (GLS), circumferential (GCS), and radial strain, native T1 mapping, and extracellular volume (ECV) quantification. During surgery, deep needle biopsies were obtained from the basal interventricular septum. Diffuse myocardial fibrosis (Hx-FIB) was quantified on Picrosirius Red–stained sections as collagen volume fraction using digital image analysis.
Results
Mean age was 64 ± 10 years; 25% of patients were female. Mean transvalvular gradient was 57 ± 19 mmHg, and aortic valve area was 0.80 ± 0.24 cm². Mean GLS was 10 ± 4%, GCS 18 ± 8%, native T1 1007 ± 43 ms, ECV 28.3 ± 6.1%, and Hx-FIB 19.6 ± 5.2% (Table 1). Seven patients (23%) exhibited extensive fibrosis (Hx-FIB > 30%). Heart-rate–adjusted GLS and GCS, as well as left ventricular mass, showed moderate correlations with Hx-FIB (r > 0.43). Native T1 and ECV were also significantly correlated with Hx-FIB (r = 0.38 and r = 0.31, respectively). Combinations of GLS with ECV or native T1 demonstrated good diagnostic performance (AUC 0.851 and 0.839, respectively; both P < 0.001), with sensitivities of 86% and 100% and specificities of 75% and 62% using cutoff values of ≤11% for GLS, >30% for ECV, and >1009 ms for native T1. (Fig 1) Similarly, combinations of GCS (cutoff ≤17%) with ECV or native T1 showed strong diagnostic performance for detecting Hx-FIB > 30% (AUC 0.821 and 0.881, respectively; both P < 0.01), with sensitivities of 86% and 100% and specificities of 75% and 79%.(Fig 1)
Conclusions
Combined CMR-derived myocardial strain and tissue characterization parameters enable accurate noninvasive identification of histologically confirmed severe myocardial fibrosis in patients with aortic stenosis.Figure 1Table 1