Noninvasive quantification of principal strains in normal canine hearts using tagged MRI images in 3-D

• Published in: The American journal of physiology Vol. 264; no. 1 Pt 2; pp. H205 - H216
• Main Authors: Azhari, H, Weiss, J L, Rogers, W J, Siu, C O, Zerhouni, E A, Shapiro, E P
• Format: Journal Article
• Language: English
• Published: United States 01-01-1993
• Subjects: Animals; Dogs; Evaluation Studies as Topic; Heart - anatomy & histology; Heart - physiology; Image Processing, Computer-Assisted; Magnetic Resonance Imaging - methods; Reference Values; Stress, Mechanical
• ISSN: 0002-9513
• DOI: 10.1152/ajpheart.1993.264.1.h205

Previous studies of myocardial strains have been largely limited to invasive procedures in isolated regions. Utilizing images from tagged magnetic resonance imaging (MRI) in three dimensions (3-D), we noninvasively quantified in vivo endocardial (endo) and epicardial (epi) principal strains (PS) throughout the left ventricle (LV) and tested their uniformity. Seven normal hearts in paced reclosed-chest dogs were studied. Combining long- and short-axis images, 24 myocardial cuboids were tagged at end diastole (ED), imaged at end systole and ED, and reconstructed in 3-D. The cuboids were circumferentially arranged in three parallel approximately 1-cm-thick short-axis slices (8 cuboids/slice). By application of finite strain analysis to each cuboid, the major PS (i.e., magnitude and angle with respect to the circumferential direction of maximal shortening) was calculated for each face. when strains are averaged globally (i.e., using all regions), the average magnitude of endo PS exceeded epi PS approximately twofold: -0.24 +/- 0.07 (SD) vs. -0.11 +/- 0.05 (P < 0.05). On the other hand, the PS angle with respect to the circumferential direction at epi exceeded the angle at endo (P < 0.05) by 33 +/- 36 degrees. Average PS direction at epi, 59 +/- 32 degrees, aligned with the expected fiber direction but was perpendicular to it at endo, 26 +/- 30 degrees. However, significant regional variations were observed from wall to wall and from apex to base. When variations in location were accounted for, the pooled SD for the PS magnitudes dropped to 0.032 and for the direction to only 15 degrees. Furthermore, it was found that the anterior-posterior pair of walls contracted differently (P < 0.05) from the septal-lateral pair of walls. These observations indicate that regional myocardial function is heterogeneous and is location dependent.