Development of a cardiovascular magnetic resonance‐compatible large animal isolated heart model for direct comparison of beating and arrested hearts

Figures

• 

  FIGURE 1. A schematic of the protocol design. A full description of the experimental protocol is provided in the text. See subsequent figures for colour‐bars, units and the description of histology

• 

  FIGURE 2. A schematic of the experimental setup. Long tubing lines connected the heat exchanger/oxygenator (which also served as a reservoir) located in the control room to the heart within the chamber at the magnet. A second heat exchanger, located on the bed of the scanner, accounted for heat loss in the perfusion tubing. Pacing, ECG and temperature sensors entered the chamber via ports and perfusion pressure was measured in the aorta cannula near the coronary ostia

• 

  FIGURE 3. The custom 3D‐printed cardiovascular magnetic resonance (CMR)‐compatible perfusion chamber. This equipment was used for imaging the beating and arrested heart within the scanner (overview, (a)). The heart was supported at the isocentre of the scanner on a 3D‐printed mesh‐design tray (shown in (b)), which could be attached to the removable lid. The lid (b) included the aortic perfusion cannula (i). The left ventricle cavity vent is clearly visible in (c) (ii). Oil‐filled tubing (iii) on the bottom of the tray provided a fiducial reference for post‐CMR histology (c)

• 

  FIGURE 4. Still frames from cine cardiovascular magnetic resonance (CMR) in the beating Langendorff‐perfused heart. Long (a) and short axis (b) still frames from cine acquisitions in heart #10. Both images clearly show the right (i) and left (ii) ventricular blood pools. The long axis, approximately equivalent to a four‐chamber view (a), shows the perfusion through the aorta ((iii), the red arrow indicates the perfusate flow direction) and the location of the left ventricle cavity vent (iv). The short axis image (b) also shows effluent perfusate in the bottom of the chamber (vi) before it is drained with the secondary pump circuit and the oil‐filled tubing (v). Video S3 shows the equivalent beating heart images

• 

  FIGURE 5. Diffusion tensor cardiovascular magnetic resonance (DT‐CMR) results from heart #10 acquired in the most and least contracted cardiac phases in the beating heart and in both slack and contracted arrested states. There is good agreement between the equivalent arrested and beating states, but little difference between the two contractile states (most contracted beating vs. least contracted beating) and (most contracted arrest vs. least contracted arrest) in both cases. The mean (median for E2A) and standard deviation (interquartile range for E2A) over the left ventricular myocardium is shown below each map, except for the helix angle (HA), where the mean transmural helix angle gradient is shown

• 

  FIGURE 6. Example T1 and T2 maps from heart #10. T1 data were acquired using the Modified Look Locker Imaging method and T2 data using a T2 preparation‐based method. Data were acquired ~1 h after the heart began contracting after reperfusion

• 

  FIGURE 7. A comparison of T1, T2 and parametric mapping between the beating and arrested hearts with comparison with literature values. Reference T1 and T2 values were obtained in the study described in Scott et al. 22 Langendorff values were obtained in hearts #9 (most contracted, least contracted and arrested slack) and #10 (beating most contracted and least contracted and arrested contracture). T2 mapping beating heart data were acquired in the most contracted phase, and T1 mapping data were acquired in least contracted for heart #9 and most contracted for heart #10, because of triggering problems in the least contracted phase. Literature values are plotted as mean ± standard deviation or median ± interquartile range over the cohort as available, 2 , 3 , 4 , 9 , 22 , 23 , 24 , 25 , 26 , 27 , 28 , 29 whereas the Langendorff values are plotted as individual points for the mean left ventricular value (median for E2A) in each pig heart. Langendorff beating heart values are shown as contracted (referred to as ‘most contracted’ in the text) or slack (referred to as ‘least contracted’ in the text)

• 

  FIGURE 8. E2A distribution within the left ventricular myocardium in corresponding beating and arrested hearts ((a) contracture arrest/beating most contracted and (b) slack arrest/beating least contracted) with comparison to equivalent data acquired in vivo in a set of different porcine hearts (N = 6 systole (c), N = 5 diastole (d)). The Langendorff‐perfused heart does not achieve a diastolic‐like E2A distribution in the least contracted state. Reference data (c and d) replotted as median ± interquartile ranges are from Ferreira et al. 12

• 

  FIGURE 9. Example bright field histology. Masson‐stained 5‐μm sections in the longitudinal–transmural plane with a magnified region (right). Cytoplasm stains red/pink, collagen blue and nuclei black. The full stack of images is shown in Video S5

• 

  FIGURE 10. Example fluorescent confocal histology images. Images are from 3D fluorescent confocal acquisition of a thick section ~100‐μm stained with wheat germ agglutinin (red, cell membranes) and DAPI (green, nuclei). Sections were cut parallel to the epicardial surface. The 3D acquisition allows resectioning in any plane. The images shown were acquired using a tile scan (four columns, six rows) method and the in‐plane field of view is ~1 x 1.5 mm2 after combination (as shown). A movie of the same data is shown in Video S6