A novel 3D-Printed preferential posterior mitral annular dilation device delineates regurgitation onset threshold in an ex vivo heart simulator

Abstract

Mitral regurgitation (MR) due to annular dilation occurs in a variety of mitral valve diseases and is observed in many patients with heart failure due to mitral regurgitation. To understand the biomechanics of MR and ultimately design an optimized annuloplasty ring, a representative disease model with asymmetric dilation of the mitral annulus is needed. This work shows the design and implementation of a 3D-printed valve dilation device to preferentially dilate the posterior mitral valve annulus. Porcine mitral valves (n = 3) were sewn into the device and mounted within a left heart simulator that generates physiologic pressures and flows through the valves, while chordal forces were measured. The valves were incrementally dilated, inducing MR, while hemodynamic and force data were collected. Flow analysis demonstrated that MR increased linearly with respect to percent annular dilation when dilation was greater than a 25.6% dilation threshold (p < 0.01). Pre-threshold, dilation did not cause significant increases in regurgitant fraction. Forces on the chordae tendineae increased as dilation increased prior to the identified threshold (p < 0.01); post-threshold, the MR resulted in highly variable forces. Ultimately, this novel dilation device can be used to more accurately model a wide range of MR disease states and their corresponding repair techniques using ex vivo experimentation. In particular, this annular dilation device provides the means to investigate the design and optimization of novel annuloplasty rings.

1. Introduction

Valvular heart disease is a common cause of morbidity and mortality globally and affects approximately 2.5% of the United States population. [1] The most prevalent cause of valvular heart disease is mitral regurgitation (MR). There are several causes of MR, including primary disease of the mitral valve apparatus (e.g. rupture of the chordae tendineae, damage to the leaflets, stretching of the annulus, papillary muscle dysfunction, and endocarditis) and secondary regurgitation due to underlying diseases, such as ischemic cardiomyopathy, nonischemic cardiomyopathy, or any process that results in left ventricular dilation. [2] In both types of MR, to compensate for reduced forward stroke volume, the left ventricle (LV) dilates. This enlargement of the LV leads to further annular dilation along with papillary displacement, further worsening MR in a positive feedback loop. Thus, regardless of etiology, most patients with chronic MR exhibit some degree of mitral annular dilation, though due to the structure of the fibrous skeleton of the heart, the annulus preferentially dilates posteriorly [2] Ring annuloplasty is an essential component of surgical repair operations for MR. [3,4] By using an annuloplasty ring to reduce the dilated annulus size, leaflet coaptation is improved and further annular dilation is prevented. [2,5] Although numerous annuloplasty ring variations exist, there is little data on the optimal geometry and material composition. [6] It has been shown that flexible rings, in comparison with rigid rings, result in improved mitral valve function after implantation [7], and previous studies indicate that saddle-shaped conformations result in improved coaptation and lower forces on the ring itself [8,9]. Moreover, some reports have observed that complete rings may prevent MR recurrence more effectively in comparison to partial rings. [10,11] Nevertheless, a recent literature review states that, due to the underpowered nature of published datasets, it is impossible to determine a statistically significant difference between any of the 37 different annuloplasty rings that were studied. [12]

In order to understand the biomechanics of MR and ultimately design the optimal annuloplasty ring, we can turn to ex vivo cardiac simulators which represent a valuable method of quantitatively analyzing valvular biomechanics. These simulators have been successfully implemented to analyze and optimize surgical techniques for MV repair. [13–19] However, ex vivo experimentation is limited by its ability to accurately mimic physiologic conditions for the valve, and thus much of the research is devoted to finding methods to more accurately simulate in vivo disease states. [20–23] Additionally, because many cardiac diseases involve multifaceted dysfunction as well as geometric changes to the valve, one approach to avoid confounding variables is to simulate each component individually, and thus more precisely analyze the specific effect of a given facet of a disease. [24] Here, we present a means of more accurately simulating the selective posterior dilation of the annulus that typically occurs with chronic MR. Though there has been one previous ex vivo dilation model utilizing adjustable segments to dilate the annulus, [25] the goal of this study is to develop a device capable of mimicking the asymmetric annular dilation state ex vivo with a continuous dilation profile and the ability to test annuloplasty ring designs.

2. Materials and methods

2.1. Dilation device design

A novel 3D-printed valve dilation device, inspired by an iris, was designed to preferentially dilate the posterior mitral valve annulus to more accurately recreate physiologic mitral annular dilation. Fig. 1 A and B show labeled renderings of the final design. A high-resolution 3D printer (Carbon M2 Printer; Redwood City, CA) enabled the manufacture of parts with a large range of material properties, complex geometries, and features as small as 0.2 mm. To accurately model mitral annular dilation, the device needed to mimic the d-shape of a natural annulus while dilating the posterior leaflet with negligible dilation of the anterior leaflet. The final design incorporates five overlapping aperture blades. Fig. 1 C shows the design with a porcine mitral valve in its native state; Fig. 1 D shows the same valve in a dilated state. Note that, for image clarity, the photographs were taken without water filling the valve, so the valve leaflets slightly obscure the orifice size. To achieve a d-shaped aperture that preferentially dilates the posterior leaflet, one of the five blades features a straight inner edge. The anterior section of the annulus is sewn to the straight inner edge of this anterior blade and thus the anterior annulus does not dilate. The posterior annulus is sewn to the remaining four blades, which then expand radially outward to dilate the posterior annulus.

Fig. 1.

(A) Rendering of the dilation device. (B) Labeled exploded view rendering of the dilation device. (C) A porcine valve sewn into the dilation device in its native state. (D) The same valve dilated by the device. Dashed blue line denotes the outline connecting each suture attachment point, used to quantify the annular circumference for each dilation state. As the pivot plate is turned, the overlapping blades rotate outward and open the aperture to stretch the annulus into a dilated configuration. The posterior section of the annulus is preferentially dilated to mimic in vivo mitral valve dilation.

The blades of the device were printed using cyanate ester, a strong and rigid plastic that allows for the extremely thin (0.62 mm) blades required for the device to fit in the left heart simulator. Additionally, small perforations were included along the inner border of the blades to facilitate suturing to the valve annulus. The blades also feature pins on both ends, one pointed up and one down, that fit into a stationary base plate and pivot plate respectively. The base plate is elastic polyurethane, a flexible material that created a hemostatic seal around the valve and also facilitated smooth sewing. A cyanate ester ring spacer separates the base plate of the device from an additional stationary top plate which surrounds the raised section of the pivot plate; this spacer allows for pressure to be applied to the top plate (transmitted to the spacer and to the bottom plate) to establish a seal between the heart simulator left ventricular chamber and the bottom plate, while not impeding the pivot plate’s freedom to turn.

2.2. Valve preparation and mounting

Porcine hearts (n = 3) were locally obtained and the mitral valves were carefully excised to preserve the annulus, leaflets, and chordae. Only valves with intercommissural distances between 30–34 mm were used to allow for sufficient dilation with the dilation device, and hearts with aberrant papillary muscle anatomy were excluded. To attach the valve to the device, eight interrupted 2–0 braided polyester sutures were used to tack the valve to the device in its native, undilated state. These tacking sutures were placed within a cuff of left atrium and the outermost section of the more fibrous peri–annular tissue and attached to the dilation device. By using this attachment method and not rigidly anchoring the annulus to the device, we were able to keep the annulus partially independent from the rigid dilation device. This atrium attachment provided more normal annular kinematics and preserved the natural elasticity and compliance found in the annular structures, which then allowed for testing of annuloplasty repair. Care was taken to ensure that the anterior section of the annulus was sewn to the straight anterior leaf, and that no suture ties restricted movement of the overlapping leaves. To seal the valve to the ring, and thus seal the juncture between the left atrium and LV sections of the left heart simulator (described below), a continuous 2–0 polypropylene suture tacked the elastic polyurethane backing of the ring to a cuff of left atrium tissue. Valves were dilated only after being tacked and attached by the left atrial cuff and peri–annular fibrous tissue to the dilation device. Annuloplasty repair was performed on select specimens after the valves were attached to the dilation device and dilated, as described in further detail below. For proper papillary muscle positioning, each muscle was sewn to custom molded silicone holders using four interrupted, pledgeted, 2–0 braided polyester horizontal mattress sutures. These holders were then affixed to carbon fiber rods within the left ventricular chamber of the heart simulator. The rods were threaded through a custom gasket setup which provided three degrees of movement, thus permitting precise positioning control to mimic in vivo placement based on geometry measurements taken of the heart prior to explanation.

2.3. Left heart simulator

The left heart simulator has been previously described in depth. [18,19,26,27] In brief, a custom 3D-printed left heart chamber (Fig. 2A) was mounted to a pulsatile linear piston pump (ViVitro Superpump, ViVitro Labs, Victoria, BC, Canada) with the ability to generate physiologic conditions. Ventricular, aortic, and left atrial pressures were recorded using pressure transducers (Utah Medical Products Inc., Midvale, Utah), and flow through the aortic and mitral locations was recorded using electromagnetic flow probes (Carolina Medical Electronics, East Bend, North Carolina). To ensure transduction of the flow meters, 0.9% normal saline was used as the test fluid and maintained at 37 °C physiologic testing condition. The pump was set to generate an effective stroke volume of 70 mL/beat at 70 bpm, and was programmed to comply with ISO 5840 standards for in vitro valve testing. Cardiac output was held at 5 L/min while peripheral resistance and compliance were titrated to produce a physiologic pressure waveform (systolic 120 mmHg, diastolic 80 mmHg).

Fig. 2.

(A) Diagram of the left heart simulator. (B) Porcine mitral valve chorda tendinea instrumented with an FBG strain gauge sensor. Red arrows denote the suture attachment points.

The valve was dilated in small increments—approximately 5 percent dilation per increment—while 10 cardiac cycles of hemodynamic data were collected at each stage of dilation. Video was also obtained at each stage with a high-speed digital camera (Chronos 1.4, Kron Technologies, Burnaby, BC, Canada) at an en face perspective of the mitral valve, and then processed using ImageJ (Bethesda, Maryland) to quantify the annular circumference. Annular circumference was measured as the total distance along the closed loop connecting each suture attachment point as shown in Figs. 1 C and 1 D. Percent annular dilation was then calculated as the annular circumference for a given dilation state divided by the baseline (native) annular circumference (and multiplied by 100%). Aortic, ventricular, and atrial pressure tracings as well as flow tracings were recorded using the ViVitest software and data acquisition system (ViVitro Labs). MR at each stage was calculated using the flow measurements. The metric for regurgitant fraction includes normal backflow at the beginning of systole, so low levels of regurgitation would be measured even in healthy valves; thus, the change in MR with respect to the baseline (undilated) measurement was calculated at each stage of dilation. A preliminary annuloplasty ring test was performed to explore the use of the dilation device as a means of testing annuloplasty ring design; an additional porcine valve was tested in a native state, dilated state, and then repaired with a full, rigid annuloplasty ring (Carpentier-Edwards Physio II, Edwards, Irvine, CA). To perform annuloplasty repair, 2–0 Ethibond annular sutures were placed circumferentially around the annulus. The sutures were then passed through the sewing cuff of the annuloplasty ring and tied down in a similar fashion to annuloplasty ring repair performed in the operating room. Due to the attachment method using a cuff of left atrium, some elasticity and compliance in the dilated annulus was preserved, which allowed for annuloplasty repair.

2.4. Chordae tendineae force measurement

Real-time force measurements on the chordae tendineae were also recorded at 1000 Hz over an entire cardiac cycle for each stage of dilation. Chordal forces were measured using customized force sensors designed and built by our laboratory [29] based on Fiber Bragg Grating (FBG) optical strain gauge sensors (DTG-LBL-1550 125 μm FBGS International, Belgium) with a sensitivity of approximately 0.1 μstrain [28]; this corresponds to relative length changes of 10^−7, which would correlate to 5e-7 mm for the 5 mm strain gauges used in this study. The force measurement method and strain gauge calibration has been described previously. [19, 29] FBG sensors have a low mass and a thin profile of 0.75 mm after encasing in a wire spring and water-resistant tubing. Thus, multiple chordae could be instrumented without disrupting the hemodynamics and structural integrity of the valve. The strain gauges were manufactured by threading the FBG sensor through a protective coil sheath that served as a mechanism for suture attachment and increased the durability of the sensor by supporting flexibility without cracking. Each sensor was individually calibrated using an Instron Microtester (Norwood, MA) with a 10 N load cell. The sensor was subjected to known tensile loads (0 N to 2N—representative of the forces during experimentation) while strain data was collected using an optical interrogator with dedicated software (Micron Optics si255 with ENLIGHT, Micron Optics, Atlanta, Georgia) at a sampling frequency of 1000 Hz. A least squares regression was used to correlate the FBG strain data to the Instron load data, specific to each sensor. For each valve, multiple native chordae (n = 4–5) were instrumented with FBG sensors. Care was taken to capture a variety of locations—including primary and secondary chordae in both posterior and anterior positions. The sensors were fixed to native chordae using double-armed polytetrafluoroethylene (PTFE) suture (Gore-Tex® Suture, WL Gore & Associates Inc., Flagstaff, Arizona) sewn proximally and distally to the FBG gratings, as shown in Fig. 2 B. Finally, the chord segment between the suture attachment points was cut, allowing the force on the chord to transmit entirely through the FBG sensor while the length of the instrumented chord remained constant. Maximum chordal forces were calculated as the average force during systole.

2.5. Statistical analysis

Statistical significance was set at P < 0.05 for all tests. Continuous variables are reported as mean ± standard deviation unless otherwise specified. For the hemodynamic data, non-parametric Friedman tests were used to compare continuous variables between groups. This test accounts for non-normally distributed forces and for the fact that the experimental method included multiple data collection stages for each valve. The non-parametric Friedman test reports one significance value for multiple treatment groups to identify significance in results correlated with changes in annular dilation, and then individual pairwise comparisons were performed between datasets of interests. Operating under the assumption that prior to this threshold the slope of MR versus percent annular dilation would be negligible, a piece-wise linear regression model was used to fit the data with the slope of the first section set to 0. Additionally, a least-squares regression model was used for fitting chordal force data. In both models, p -values were calculated using the F -test.

3. Results

Fig. 3 A shows the change in mitral regurgitant fraction versus percent annular dilation for a compilation of all valves tested. To account for variations between valves, the change in regurgitant fraction is in reference to the regurgitation measured in the baseline state (corresponding to 0% annular dilation). The data showed a threshold for the onset of MR as the percent annular dilation increased, and the resulting piecewise model was: G = 0 when d < 25.6; G = 2.5*d – 64 when d ≥ 25.6 (R² = 0.90, p < 0.01), where d is the percent annular dilation and G is the regurgitant fraction increase from baseline. This represents a threshold value for the onset of MR at 25.6% annular dilation. For annular dilation greater than 25.6%, a 2.5% increase in regurgitant fraction was observed for every 1% increase in annular dilation.

Fig. 3.

(A) Piece-wise least-squares linear regression fit of regurgitant fraction (with respect to baseline regurgitation) versus percent annular dilation: G = 0 when d < 25.6; G = 2.5*d – 64 when d ≥ 25.6 (R² = 0.90), where d is the percent dilation and G is the regurgitant fraction increase from baseline. (B) Composites of primary and secondary chordal forces (with respect to baseline forces) versus percent annular dilation. Least-squares linear regression fit with annular dilation lower than the 25.6% dilation threshold: equation.

A plot of the change in forces on the chordae, measured with respect to baseline levels, versus percent annular dilation is shown in Fig. 3 B. For each valve, the primary chordae forces were averaged, and the secondary chordae were averaged. A least-squares linear regression model was applied to the data with annular dilation less than the 25.6% threshold identified in the regurgitant fraction versus dilation analysis, resulting in the following equation: ΔF = 0.011*d (R² = 0.75, p = 0.002), where ΔF is the change in chordal force from baseline and d is the percent annular dilation. Additionally, a least-squares linear regression model applied to the post-threshold data gave the following equation, ΔF = −0.028*d + 0.94 (R² = 0.16, p = 0.073). Though the data shows a general downward trend post-threshold, this slope is not significant due to the high variation in forces with increased MR. Three representative force tracings for one of the mitral valves tested are shown in Fig. 4: baseline forces (4A), forces at the maximum dilation point before the annular dilation threshold (4B), and forces at the maximum dilation post-threshold (4C). The force tracings correspond to the forces over the course of one cardiac cycle.

Fig. 4.

Representative force tracings for composites of each class of chordae for a single valve at baseline (A), at the maximum dilation point prior to the dilation threshold (B), and at maximum dilation (C).

There was no significant difference between the measured pressures and flows when annular dilation remained below the identified threshold value. However, as the annulus was dilated above the threshold and MR increases, the hemodynamics change. Table 1 contains the hemodynamic data for the following states: baseline, pre-threshold dilation, and maximum dilation post-threshold. No significant regurgitant fraction difference between baseline and pre-threshold states was found (p = 0.519), but a there was a significantly higher regurgitant fraction in the maximum dilation state compared to baseline (p = 0.024). The hemodynamic results are also illustrated in Fig. 5, which shows the flow (5A) and pressure (5B) data for the same three states with the shaded region representing standard deviation. Both flow and pressure remain relatively unchanged between baseline and pre-threshold states, but were altered between the pre-threshold and maximum dilation states. Additionally, the supplemental annuloplasty ring experiment results are shown in Fig. 6. The baseline measured flows (6A) and pressures (6B) corresponded to expected in vivo values with an maximum aortic pressure of 126.0 mmHg, a minimum aortic pressure of 88.6 mmHg, and a cardiac output of 4.6 L/min [30,31]; both flow and pressure appear to be largely restored after the annuloplasty. The regurgitant fraction was measured to be 12% at baseline, 34% when dilated, and 14% post-repair.

Table 1.

Hemodynamic Parameters.

	Baseline	Pre-Threshold dilation	Max dilation
Heart rate (bpm)	70.00 ± 0.00	70.00 ± 0.00	70.00 ± 0.00
Mean arterial pressure (mmHg)	106.69 ± 0.22	100.35 ± 4.38	68.10 ± 9.66
Diastolic pressure (mmHg)	89.07 ± 1.62	83.71 ± 5.00	55.51 ± 9.57
Systolic pressure (mmHg)	125.01 ± 1.36	117.90 ± 3.71	83.09 ± 10.11
Mean atrial pressure (mmHg)	6.94 ± 1.18	5.93 ± 0.87	5.89 ± 0.10
Mean ventricular pressure (mmHg)	47.66 ± 0.95	45.80 ± 1.66	32.26 ± 3.62
Cardiac output (liters/min)	3.52 ± 0.21	3.92 ± 0.54	3.22 ± 0.71
Effective stroke volume (ml)	50.26 ± 3.04	56.00 ± 7.77	46.02 ± 10.10
Pump stroke volume (ml)	109.83 ± 0.03	109.81 ± 0.18	109.84 ± 0.15
Mitral valve mean gradient (mmHg)	−2.27 ± 1.10	−2.95 ± 0.99	−0.86 ± 1.21
Mitral valve mean back pressure (mmHg)	105.38 ± 2.82	100.20 ± 2.62	64.98 ± 9.18
Mitral forward flow time (sec)	0.53 ± 0.00	0.52 ± 0.01	0.53 ± 0.00
Mitral forward volume (ml)	66.09 ± 2.71	74.29 ± 7.45	86.42 ± 7.95
Mitral valve RMS forward flow (ml/sec)	216.28 ± 66.40	301.76 ± 77.74	366.21 ± 50.73
Mitral regurge fraction (%)	23.91 ± 4.27	24.83 ± 4.31	47.15 ± 7.80
Mitral leakage rate (ml/sec)	−32.77 ± 13.75	−40.27 ± 9.44	−108.73 ± 13.13
Mitral leakage volume (ml)	−8.77 ± 3.47	−11.04 ± 2.55	−27.85 ± 2.67
Mitral closing volume (ml)	−7.05 ± 1.04	−7.27 ± 0.31	−12.57 ± 2.50
Ventricular energy (mJ)	1012.10 ± 68.10	1001.63 ± 80.72	797.92 ± 105.88
TransMitral forward energy loss (mJ)	16.02 ± 9.00	1.51 ± 5.98	10.17 ± 17.88
TransMitral closing energy loss (mJ)	34.58 ± 5.88	35.70 ± 3.07	62.37 ± 8.95
TransMitral leakage energy loss (mJ)	124.48 ± 45.79	150.09 ± 33.54	243.97 ± 12.31
TransMitral total energy loss (mJ)	175.08 ± 55.35	187.31 ± 41.34	316.50 ± 20.02

Data presented as mean ± standard deviation. p-values calculated using non-parametric Friedman tests.

Fig. 5.

(A) Mean flow tracings showed no significant difference between the regurgitant fraction at baseline and prior to the 25% annular dilation threshold. However, the regurgitant fraction was significantly higher at maximum dilation compared to baseline (p = 0.024). (B) Mean pressure tracings also show the aortic, left ventricular (LV), and left atrial (LA) pressures remaining near baseline levels prior to the threshold, but mean aortic and LV pressures were lowering significantly at maximum dilation (p < 0.001). Shaded area represents standard deviation.

Fig. 6.

(A) Mean flow tracings at baseline, after maximum dilation, and after repair with an annuloplasty ring. Flow was restored to baseline levels after repair. (B) Mean pressure tracings illustrate that the aortic, left ventricular, and left atrial pressures were partially restored to baseline levels after repair.

4. Discussion

Much of the previous literature devoted to studying MR focused on the relationship between regurgitant fraction and annular size (rather than percent annular dilation), and thus no threshold for the onset of MR could be easily identified. [32–34] The dearth of sufficient in vivo literature on this comparison is likely due to the lack of data featuring annular measurements with respect to a baseline state over the progression of a disease. However, with our dilation device to selectively dilate the posterior annulus of a porcine mitral valve, this comparison can be made using an ex vivo model, thus enabling more physiologically relevant simulations of dilation-induced MR. The MR onset threshold was clearly present in this study in the regurgitant fraction (with respect to baseline) versus percent annular dilation data. The piecewise function used to determine the 25.6% threshold value as well as the slope for the post-threshold data produced a better fit with a higher R² and lower p-value than both a linear regression model and a quadratic regression model. This threshold is likely due to a relatively large coaptation surface in healthy mitral valves, [35] which allows for a moderate degree of annular dilation before leaflet coaptation is severely impaired.

Diseased valves, especially those with excess myxomatous leaflet tissue, may have a different threshold value than that determined for healthy valves that are mechanically dilated. Future studies could examine the effect of annular dilation on a variety of diseased valves. Additionally, with an increased sample size and the inclusion of human valves, identifying the percent annular dilation threshold could be used as a predictor for patients who are at risk for severe MR. Annular size has been proposed as a metric to predict progression from moderate to severe MR, [34] but this threshold value of percent annular dilation that results in severe MR could provide a more accurate predictor. With data from multiple surveillance echocardiography studies, clinicians could potentially even predict when a patient is likely to develop severe mitral regurgitation and identify the ideal temporal window to undergo surgical repair, which at present is oftentimes unclear.

The results of the chordal force data revealed that the force profile on the chordae tendineae was also affected by the degree of dilation. As dilation increased prior to the threshold, the forces on both primary and secondary chordae tendineae increased. This increase in force was as expected. [13] In cases of chronic functional MR, these chordal forces would likely continue to increase after MR onset as the heart remodels to maintain pressure despite the increasing regurgitant fraction. However, the chordal forces did not increase post MR onset in this study because the ex vivo dilation device modeled acute MR after the threshold was reached, where the increasing regurgitant fraction resulted in a sharp decrease in LV and aortic pressures. Importantly, the wide range of forces present post-threshold may illustrate the destabilizing effect of MR on the mitral valve apparatus—when the valve is not coapting properly, the force distribution across chordae is altered. In a properly coapting valve, once the leaflets are in apposition, forces are redistributed throughout the mitral valve apparatus, primarily in a lateral direction, much like a roman arch. [36] In a valve with regurgitation and improper coaptation, the forces cannot be distributed laterally, and instead the leaflet acts as a sail with wide swings in force distribution and magnitude. These chordal forces can be minimized with an optimal annuloplasty ring. Thus, using hemodynamics and chordal force measurements as metrics, a multitude of annuloplasty devices can be tested and analyzed with this dilation device and experimental protocol.

While we have created a novel and quantifiable mechanism for inducing annular dilation in previously healthy porcine valves for ex vivo analysis, our study is not without limitations. One limitation lies in the dynamic movement of the mitral valve annulus over the course of a cardiac cycle. While the dilation device gives a close approximation of the mitral annulus and the posterior dilation, the mitral valve in vivo fluctuates between a flat and saddle conformation, which plays a role in the stresses on valvular apparatus. [9,37] To more accurately study mitral regurgitation, the device will be modified to support these conformational changes. In particular, preliminary work to manufacture an annulus mount with flexible 3D-printed materials has shown promise in supporting the physiologic conformational changes during ex vivo experimentation. The device design can be optimized to utilize rigid features to hold the valve in a dilated configuration and flexible material to support dynamic movement, thus advancing the simulation technology towards a more accurate simulation of MR. Furthermore, force sensors can be utilized to monitor the forces on the mitral valve annulus, which will be an important metric in analyzing annuloplasty rings, especially those designed to support in vivo conformational changes. A second limitation of our study is that we chose to focus only on annular dilation—only one of a plethora of pathophysiologic findings in mitral valve disease. Degenerative mitral disease, for example, is characterized not only by annular dilation, but often chordal elongation and/or rupture, as well as excessive and thickened leaflet tissue. Ischemic or functional mitral regurgitation is typically associated with ventricular dilation, papillary muscle displacement, and chordal teathering of otherwise normal leaflets, with annular dilation occurring late in the disease process. Our heart simulator has the capability to mimic these parameters through the use of various papillary muscle anchoring positions, chordal transection, cross-species models utilizing oversized valves to mimic redundant leaflet tissue, and altered ventricular mechanics via pathophysiologic pump waveforms. Future studies will incorporate a multitude of these factors into a single model to more accurately mimic degenerative or functional mitral regurgitation. In this study, however, our primary goal was to design, develop, test, and validate a mechanism to induce pathologic annular dilation in previously healthy valves in a quantifiable way; by focusing on validating the annular dilation mechanism, we could reduce the effects of confounding factors. All ex vivo models using animal tissue are limited by the rarity of naturally diseased specimens, thus one must instead model the pathology—annular dilation in this case. Ultimately, through the development and validation of ex vivo models for each element of valvular disease, we can provide comprehensive biomechanics analysis of the disease pathology and optimize repair techniques for the benefit of patients.

5. Conclusion

A novel annular dilation device was developed as a reliable means for reproducing mitral valve dilation ex vivo, supporting direct analysis of the relationship between annular dilation and valve biomechanics, as well as providing the foundation for repair technique examination. The device was shown to be successful in preferentially dilating the posterior annulus and in inducing MR. Hemodynamic data demonstrated that MR increased linearly with respect to percent annular dilation when dilation was greater than a 25.6% dilation threshold, while the annular dilation did not cause significant increases in regurgitant fraction prior to the threshold. This threshold represents a key facet of the complex biomechanics of MR, which can improve our understanding of how a patient’s individual mitral valve geometry and dilation state dictate progression of MR. With modifications to the material properties of the device components, this device can capture the complexities of modeling a dilated mitral valve. In turn, this model can be used not only for biomechanics analysis of the diseased state, but also to investigate the design and optimization of annuloplasty rings and surgical repair techniques.