Magnetic resonance imaging with 3-dimensional analysis of left ventricular remodeling in isolated mitral regurgitation: implications beyond dimensions

Abstract

Background: Although surgery is indicated in patients with mitral regurgitation (MR) when left ventricular (LV) end-systolic (LVES) dimension is >40 mm, LV ejection fraction may decrease after mitral valve surgery. We hypothesize that significant LV remodeling before surgery is not reflected by standard echocardiographic parameters measured at the base of the heart.

Methods and results: Ninety-four patients (age, 54 ± 11 years; 38% female) with degenerative isolated MR underwent cine magnetic resonance imaging with tissue tagging and 3-dimensional analysis. In 51 control subjects (age, 44 ± 14 years; 53% female), the relation between LVES volume (LVESV) and LVES dimension was quadratic, whereas in 94 MR patients, this relation was cubic, indicating a greater increase in LVESV per LVES dimension among MR patients. Moreover, magnetic resonance imaging LVESV from summated serial short-axis slices was significantly greater than LVESV assessed with the Bullet formula in MR patients, attributed to a more spherical remodeling distal to the tips of the papillary muscles (P<0.001). Thirty-five patients underwent mitral valve repair per current guideline recommendations. LV ejection fraction decreased from 61 ± 7% to 54 ± 8% (P<0.0001) and maximum shortening decreased significantly below normal at 1 year postoperatively (P<0.0001). Despite normalization of LV stroke volume and LV end-diastolic volume/mass ratio, there was a persistent significant increase in distal LVES 3-dimensional radius/wall thickness ratio and LVESV index after surgery.

Conclusions: Despite apparently preserved LVES dimension, MR patients demonstrate significant spherical mid to apical LVES remodeling that contributes to higher LVESV than predicted by standard geometry-based calculations. Decreased LV strain after surgery suggests that a volumetric analysis of LV remodeling and function may be preferred to evaluate disease progression in isolated MR.

Figure 1

(A) Representative LV end-systolic (ES) with two chamber view (top row) and 3-dimensional surface representations (bottom row) using color scales of LVES endocardial surface circumferential curvature (1/mm) from a control subject and a MR patient; (B) Systematic simulation of MR LV remodeling with respect to control. The MR heart has the same LVES dimension and similar long axis length as control. However, there is lesser curvature from the mid to distal LV segments represented by a dimmer red color in the MR patient vs. control (brigher yellow color). These changes in the MR patient contribute to a more spherical LV remodeling and a larger LVES volume. Blue dot: mid-septum; black triangle: apex.

Figure 2

Correlation between LVESV and LVESD in MR patients (A) and control subjects (B). The solid lines represent the fitted model for the LVESD vs. LVESV relation with 95% confidence intervals (dash lines), which is cubic in MR patients (n=94) and quadratic in controls (n =51). Data of calculated LVESV using the Bullet formula are shown as dark cyan points in (A) and (B). The difference between the measured LVESV from summated short axis images and calculated LVESV from Bullet formula in MR were plotted in (C) vs. LVES circumferential curvature at distal LV. The solid back line in (C) represents the fitted model.

Figure 3

Three-dimensional LVED (left column) and LVES (right column) geometry with maximum shortening in the controls (n=51) and MR patients (n=94) divided into those with LVESD <37 mm and LVESD ≥37 mm. These data demonstrate progressive global LVED and LVES remodeling in both groups of MR patients compared to controls. However, LV maximum shortening remains normal or even supra-normal in both MR groups. *: P<0.05 vs. controls; †: P<0.05 vs. MR patients with pre-operative LVESD<37 mm.

Figure 4

Comparison of LV end-diastolic (ED) geometric remodeling in controls and in surgical MR patients with pre-operative LVESD < and ≥ 37mm before and after surgery. These data demonstrate progressive LV remodeling at ED in the two MR groups and their recovery after surgery. LVED R/T ratio is normalized after surgery in both MR groups. Circumferential curvatures in MR LVESD<37mm are normalized after surgery while in MR LVESD≥37mm, circumferential curvatures are increased yet not normalized. *: P<0.05 vs. controls; †: P<0.05 vs. pre-operative MR.

Figure 5

Comparison of LV end-systolic (ES) geometric remodeling in controls and in surgical MR patients with pre-operative LVESD < and ≥ 37mm before and after surgery. These data demonstrate progressive LV remodeling at ES that is not normalized after surgery in the MR patients with pre-operative LVESD ≥37mm, while it is normalized after surgery in MR patients with pre-operative LVESD <37mnm. *: P<0.05 vs. controls; †: P<0.05 vs. pre-operative MR.

Figure 6

Comparison of LV end-systolic (ES) maximum shortening in controls and in surgical MR patients with pre-operative LVESD < and ≥ 37mm before and after surgery. These data demonstrate that maximum shortening is decreased below normal in both groups of MR patients except that LVES maximum shortening is preserved at the base in patients with pre-operative LVESD <37 mm. *: P<0.05 vs. controls; †: P<0.05 vs. pre-operative MR.