Magnetic resonance imaging-based finite element stress analysis after linear repair of left ventricular aneurysm

Abstract

Objective: Linear repair of left ventricular aneurysm has been performed with mixed clinical results. By using finite element analysis, this study evaluated the effect of this procedure on end-systolic stress.

Methods: Nine sheep underwent myocardial infarction and aneurysm repair with a linear repair (13.4 +/- 2.3 weeks postmyocardial infarction). Satisfactory magnetic resonance imaging examinations were obtained in 6 sheep (6.6 +/- 0.5 weeks postrepair). Finite element models were constructed from in vivo magnetic resonance imaging-based cardiac geometry and postmortem measurement of myofiber helix angles using diffusion tensor magnetic resonance imaging. Material properties were iteratively determined by comparing the finite element model output with systolic tagged magnetic resonance imaging strain measurements.

Results: At the mid-wall, fiber stress in the border zone decreased by 39% (sham = 32.5 +/- 2.5 kPa, repair = 19.7 +/- 3.6 kPa, P = .001) to the level of remote regions after repair. In the septum, however, border zone fiber stress remained high (sham = 31.3 +/- 5.4 kPa, repair = 23.8 +/- 5.8 kPa, P = .29). Cross-fiber stress at the mid-wall decreased by 41% (sham = 13.0 +/- 1.5 kPa, repair = 7.7 +/- 2.1 kPa, P = .01), but cross-fiber stress in the un-excluded septal infarct was 75% higher in the border zone than remote regions (remote = 5.9 +/- 1.9 kPa, border zone = 10.3 +/- 3.6 kPa, P < .01). However, end-diastolic fiber and cross-fiber stress were not reduced in the remote myocardium after plication.

Conclusion: With the exception of the retained septal infarct, end-systolic stress is reduced in all areas of the left ventricle after infarct plication. Consequently, we expect the primary positive effect of infarct plication to be in the infarct border zone. However, the amount of stress reduction necessary to halt or reverse nonischemic infarct extension in the infarct border zone and eccentric hypertrophy in the remote myocardium is unknown.

Figure 1

Iterative process for determining material parameters. Initial parameters were taken from previously published studies. C’s from equation 1 in aneurysmal and nonaneurysmal regions, and T_max from equation 3 were iteratively scaled to achieve correct volumes. Remaining parameters were then scaled to achieve good agreement with tagged MRI strains as measured by rms error. EDV, End-diastolic volume; ESV, end-systolic volume; rms, Root-mean-square.

Figure 2

rms error of model predictions relative to strain measurements made with tagged MRI. rms error in the plication models was the same as rms error in the aneurysm models, indicating both FE models reproduced strain measurements with the same certainty. rms, Root-mean-square.

Figure 3

End-systolic midwall fiber stress in a representative aneurysmal heart (A) and postplication heart (B), septal-posterior view. Averaged across all models (C), fiber stress decreased in the border zone and infarct but did not significantly decrease in remote regions. After plication, border zone stress was at the same level as remote regions. *P <.05 plication versus sham. BZ, Border zone.

Figure 4

End-systolic midwall cross-fiber stress in a representative aneurysmal heart (A) and postplication heart (B), septal-posterior view. Cross-fiber stress concentration exists in the infarct of the aneurysmal heart (A) and remains in the nonexcluded septal infarct after plication (B). Averaged across all models (C), cross-fiber stress decreased in the border zone and infarct but not in remote regions. After plication, border zone cross-fiber stress was still significantly elevated above remote regions. *P < .05 plication versus sham. †P < .05 plication remote versus border zone. BZ, Border zone.