Study Design

The biomechanical properties of four versions of the SIS-ECM material were evaluated through tensile testing to determine factors such as stiffness, maximum stress, and maximum load. The version of SIS-ECM with the most favorable biomechanical performance was chosen for in vitro posterior mitral valve reconstruction. A customized patch design for posterior mitral valve reconstruction was proposed based on geometric measurements from cardiac MRI scans of healthy pigs. Subsequently, in vitro posterior mitral valve reconstruction was conducted using a modified left heart simulator to compare the geometric parameters of the valve before and after the reconstruction.

Uniaxial Tensile Testing of Small Intestinal Submucosal Extracellular Matrix

Four versions of the SIS-ECM material (CorMatrix®, Cardiovascular Inc., Alpharetta, GA, USA), including 2- and 4-ply lyophilized (0.14- and 0.28-mm thickness) and 2- and 4-ply vacuum-pressed (0.07- and 0.14-mm thickness) SIS-ECM, underwent mechanical uniaxial tensile testing. The SIS-ECM samples were cut into hourglass-shaped specimens using a dedicated die-cutter to ensure uniformity and reproducibility of the sample sizes. The hourglass shape is designed to prevent the samples from rupturing at the clamped areas, ensuring that any breakage occurs in the gauge area. After being cut, the dimensions of the gauge area were 4 × 3.5 mm, as shown in Fig. 1. The width and length of the samples were measured using calipers, while the thickness of the samples was measured using a thickness gauge with a constant pressure of 18 kPa (Model 7301, Mitutoyo Corporation, Kawasaki, Japan).

Fig. 1

Dimensions of the SIS-ECM samples according to the American Society for Testing and Material tensile test standard

Before tensile testing, the SIS-ECM samples were hydrated in saline (Ringer-acetate, Fresenius Kabi, Fresenius SE & Co. KGaA, Bad Homburg vor der Höhe, Germany) for 10–15 min. The samples were tested in a Bose ElectroForce 3200 (Bose Corporation, ElectroForce Systems Group, Minnesota, USA) with a displacement sensor (+—0.001 mm) and a 225N load cell (+—0.002 N). Two custom gripping clamps were used for securing the samples during the experiments (G341-10–32 Needlenose Screw Vice Grip Rigid Mount, TestResources Inc., USA). The distance between the two clamps was incrementally increased at a rate of 0.5 mm/s and a sampling frequency of 10 Hz, leading to the eventual rupture of the samples. The samples underwent a 10 mm displacement, with displacement and force data recorded using a load cell in the tension machine.

Definition of Stress–Strain Relationship

The samples were examined by a stress–strain relationship. Engineering strain (\(\varepsilon )\) is defined as the change in length of a sample divided by its original gauge length [12]:

$$\varepsilon =\frac_}_}$$

Engineering stress (\(\sigma\)) is defined as the force divided per unit area [12]:

where \(_\) is the original cross-sectional area.

The stress–strain curve shows fundamental mechanical properties of a material. Initially, the samples exhibit elastic behavior with a linear stress–strain relationship, characterized by Young’s modulus (\(E=\sigma /\varepsilon )\) [12].

Beyond the yield strength, permanent deformation occurs. The curve peaks at the ultimate tensile strength before the material weakens and ruptures, with fracture strain measured at the final data point.

Posterior Mitral Valve Patch Design

A 10 × 7 cm sheet of 2-ply vacuum-pressed SIS-ECM was used to create the posterior MV patch for reconstruction of the posterior MV leaflet and associated chordae tendineae. To accurately size the patch, we obtained MRI scans of five healthy six-month-old, 80 kg female pigs of mixed Yorkshire and Danish Landrace breeds from the University of Aarhus Experimental Animal Farm in Aarhus, Denmark. These scans were conducted to acquire geometric measurements of the mitral valve and sub-valvular apparatus (Fig. 2). During MRI, the animals were anesthetized by continuous intravenous administration of 4.375 mg/kg/h Propofol, and 6.2 µg/kg/h Fentanyl.

Fig. 2

Geometrical measurements based on magnetic resonance imaging scans in a long-axis view of the heart in diastole and b short-axis view of the heart in diastole of five healthy, female 80-kg pigs used to propose the posterior mitral valve patch design. AP: annular plane; APMA: anterolateral papillary muscle to annulus length; APMP2: anterolateral papillary muscle to P2 length; P1, P2, and P3: posterior mitral valve scallops; PPMA: posteromedial papillary muscle to annulus length; PPMP2: posteromedial papillary muscle to P2 length

MRI scans were performed with a 1.5 Tesla Philips MRI scanner using a spine coil under the pig and an 18 elements anterior coil. The pigs were placed in a supine position. 2-dimensional cine images oriented around the MV were acquired in the short- and long-axis orientations. Furthermore, cine images oriented perpendicular to the MV were obtained (STACK1 and STACK2). For each stack, 8 slices with 6 mm thickness, a pixel size of 1.2 × 1.2 mm and 30 frames with a temporal solution of 32 ms were acquired. Finally, near-isotropic static 3-dimensional images at systole and diastole were acquired with a voxel size of 1.5 × 1.25 × 1.25 mm. The MRI data were stored, and the analysis was performed offline using software developed by MRI physicists at the Magnetic Resonance Research Centre, Aarhus University Hospital.

In a long-axis view, the distance from the fibrous head of the anterolateral papillary muscle to the middle segment of the P1 scallop and the distance from the posteromedial papillary muscle to the middle of the P3 scallop was measured. Similarly, the distances from each of the two papillary muscle tips to the middle of the P2 scallop was measured. In a short-axis view, the circumferential dimensions of each of the three posterior scallops (P1, P2, and P3) was measured. All distances were measured from inner edge to inner edge.

Experimental In Vitro Setup

A porcine heart was chosen for this in vitro model, due to its similarity to the human heart physiology and anatomy [13]. The MV apparatus, encompassing the annulus, anterior and posterior leaflets, chordae tendineae, and anterior and posterior papillary muscles, was excised from the hearts of seven healthy six-month old, 80 kg female Danish Landrace pigs sourced from a local abattoir. Only MV apparatuses with intact chordae tendineae and matching dimensions of a Physio I™ annuloplasty ring size M36-M38 (Edwards Lifesciences, Irvine, CA) were included.

A power calculation based on a previous study with a similar setup was performed [8] using the equation below

$$n=2_+_\right)\frac\right)}^$$

where zα is the α-error of 0.05 and zβ is the probability of 85% (power) of which a change of 17% (δ) would be detected with an assumed standard deviation (σ) of no more than 8%. Using this formula, it can be found that n = 7.

The Modified Left Heart Simulator

A modified left heart simulator, specifically designed for studying atrioventricular valves [8, 14, 15], was used for the functional geometric measurements. The model utilized a static flow and did not incorporate a pump or a compliance chamber. Tap water was used as the operating fluid. The simulator consists of an acrylic cylinder, an annular holding plate, and two papillary muscle rods (Fig. 3). The annulus holding plate, 3D printed to fit the annular proportions of a standard human MV, partitioned the acrylic cylinder into a ventricular and an atrial section. The inner circumference of the annulus holding plate was reinforced with Dacron® (Dupont, Wilmington, DE). The MV apparatus annulus was secured to the Dacron lining on the annulus holding plate with 5–0 Optilene® (B Braun, Melsungen, Germany) via a continuous interlocking suturing technique. Each papillary muscle was fitted with a fixture, secured with a 5–0 Optilene® suture. Subsequently, the fixture was securely fastened to its associated papillary muscle rod in the left heart simulator. For geometric measurements, a high-resolution camera was used to capture photographs of the MV apparatus from an atrial perspective.

Fig. 3

Schematic drawing of the in vitro left heart simulator

A Mikro-Tip pressure catheter (SPR- 305S, Millar Instruments, Houston, TX) was utilized to measure left ventricular pressure in the left heart simulator. Since we have previously shown that a gradually increase in left ventricular pressure does not result in a significant geometric difference between the native and reconstructed MV, all geometric data were extracted at 120 mmHg [15]. Furthermore, maximum pressure tests were performed on all reconstructions to evaluate the short-term maximum resistance under static pressure. These tests involved gradually increasing the pressure until the valves ruptured. Finally, leaflet coaptation height was visually observed through the ventricular section chamber.

Reconstruction of the Posterior Mitral Valve Leaflet

After collecting all the necessary data from the native MV apparatus, the posterior leaflet was surgically removed along with the corresponding chordae tendineae. Afterwards, the annular end of the leaflet patch was connected to the MV annulus with a 5–0 Optilene® continuous interlocking suture (Fig. 4). The patch was securely attached to the anterior leaflet at each commissure with a plication suture.

Fig. 4

Schematic drawings of the posterior mitral valve patch design and implantation technique. a The specific dimensions of the patch design. b The measured A1, A2, A3 and P1, P2, P3 segments seen from a superior view. The big dots represent the coaptation point. c The native mitral valve implanted in the in vitro model. d The reconstruction, illustrating the 3-suture loop technique used to attach the four anchoring points of the patch to their respective papillary muscle heads, placed approximately 1 cm apart on each papillary muscle. e The patch being implanted in the in vitro model. f The patch implanted in the in vitro model. The red arrows indicate the applied ventricular pressure closing the reconstructed valve during systole

The ventricular extremity of the patch had four anchoring points. Both anchoring points were affixed to the apex of each papillary muscle, resulting in two connection points per papillary muscle spaced 1 cm apart. Three suture loops with 5–0 Optilene were employed to attach each anchoring point to the papillary muscle tips.

A crease was formed on the patch between the two attachment points on the individual papillary muscle. These two folds split the leaflet patch into three sections: P1, P2, and P3. The P1 and P3 segments served the purpose of replacing the P1 and P3 scallops as well and the chordae tendineae of the original posterior leaflet. This was achievable because the height of the leaflet patch was commensurate with the length between the native MV annulus and the tip of the native papillary muscle.

Data Acquisition and Data Analysis

Utilizing the imaging software ImageJ (US National Institutes of Health, Bethesda, MD), the geometric data was extracted from all digital images manually. All images were assessed employing a 5 mm scale. This standard was established based on a shared aperture in every image. The software was utilized to identify the annulus, positioning it slightly within the sutures for both the native and reconstructed valves. This area was identified as the total leaflet area. The annulus and coaptation line demarcated the anterior and posterior leaflets. The A2 segment was delineated from the midpoint of the anterior annulus to the coaptation point and the P2 segment from the coaptation point to the posterior annulus. Combined, these two segments formed the A2P2 segment. Two lines delineated 12 mm apart from the A2P2 line, were designated as A1P1 and A3P3 and measured accordingly. A1, A3, P1, and P3 were gauged similarly. The same approach was used for both the native and the reconstructed valves, and the data was subsequently compared.

Statistical Analysis

Normality of stress–strain and geometric data was assessed using histograms and quantile plots and tested using the Shapiro–Wilk test. Even though the stress–strain results of the lyophilized SIS-ECM was normally distributed, the stress–strain results of the vacuum-pressed SIS-ECM did not follow a Gaussian distribution and stress-stain data is expressed as median with interquartile range for easy comparison of the different SIS-ECM types. Comparison of 2- and 4-ply vacuum-pressed SIS-ECM was performed using Mann–Whitney U-test, while 2- and 4-ply lyophilized SIS-ECM was compared using independent t-test. The geometric parameters were normally distributed in the native and the reconstruction group, expressed as mean with standard deviation, and the two groups were analyzed correspondingly using dependent t-test. All tests were two-tailed and interpreted at a statistical significance level of 0.05. The statistical analyses were performed using SAS® Enterprise Guide® software, version 7.1 (SAS Institute Inc., Cary, NC).