50, and a single posterior control point 52 centered on a minor axis of the D-shape. The control points are again illustrated as tubular crimps, though as will be explained below other configurations are possible. FIG. 2B shows the core member 40 in elevational view illustrating an anterior bow 54. The core member 40 when in its relaxed, unstressed state desirably has the same shape as the Carpentier-Edwards® Physio® Annuloplasty Ring available from Edwards Lifesciences.

A still further embodiment of the present invention is shown in FIGS. 3A and 3B. A core member 60 for a flexible mitral annuloplasty ring defines a posterior portion 62 and an anterior portion 64. The core member 60 has a modified D-shape with the outwardly convex posterior portion 62 being pulled in on the right side so as to be asymmetric. As with FIGS. 2A-2B the core member 60 has a closed peripheral shape, but in this embodiment in its unstressed state mimics the shape of the Carpentier-McCarthy-Adams IMR ETlogix™ Annuloplasty Ring, also available from Edwards Lifesciences.

The core member 60 includes four discrete control points or members 68, 70, 72, 74 around the periphery at strategic locations. A first anterior control point 68 is located, when implanted, at one of the commissures of the mitral annulus, and a second anterior control point 70 is at the other commissure. As before, the anterior control points 68, 70 provide some rigidity for the core member 60 and also bend the flexible cable 66 at the opposite anterior corners. A first posterior control point 72 provides rigidity and curves the cable 66 on the left side in plan view, while a second posterior control point 74 is located on the right side in a pulled-in region. FIG. 3B shows the right side of the posterior portion dipping downward at 76, and the control point 74 desirably shapes the cable 66 in this area.

Now with reference to FIG. 4A, an annuloplasty ring 80 comprises a core member that resembles the core member 40 of FIG. 2A, and includes a closed length of braided cable 82 and a plurality, in this case five, discrete control points or members 84. This annuloplasty ring 80 in its relaxed, unstressed state is shaped to mimic the Carpentier-Edwards® Physio II™ Annuloplasty Ring available from Edwards Lifesciences. Although not shown in elevation, the Physio II™ ring has more pronounced upward bows on both the anterior and posterior sides. Also, the larger ring sizes of the Physio II™ ring become less D-shaped and more circular to better correct for pathological changes in mitral annular dimensions seen in larger patients.

FIG. 4B shows a short length of the braided cable 82, which includes seven strands of wire including a central wire and six strands 83 wound helically therearound. This construction is also known in the art as a simple 1×7 cable, having a single winding of seven wires. Other cable constructions are also possible, such as 1×3 or 1×19 simple braids. Preferably, however, the core members will include flexible cables having multi strand braids, such as 7×7, 7×19, 19×7 or even 7×7×7 braided cables. Each of these possible braid constructions is seen in FIGS. 17A-17G, and will be described in greater detail below.

The left side of FIG. 4A shows an outer fabric cover 86 which has been cut away to illustrate a portion of the inner core member. FIG. 4C shows a preferred cross-sectional layout, with the fabric cover 86 surrounding a suture-permeable interface 88, such as a silicone rubber tube. The interface 88 closely surrounds the control point 84, which in the illustrated version is a crimped tube. Inside the crimp 84 is the braided cable 82.

FIG. 5 schematically illustrates the core member of the annuloplasty ring 80 squeezed into an elongated shape to fit within a tubular access device 90. The flexible cable 82 facilitates the conversion from D-shaped to linear so that the ring 80 may be introduced to an implant site through the access device 90. The access device 80 may be a cannula or introducer tube, or other similar expedient.

This delivery method is enabled by the multi-stranded cable 82 which has the flexibility to accommodate large amounts of bending without permanent deformation. However, the disadvantage of cable is that it is not as easy to permanently shape into a ring. This issue is addressed by introducing the “control points” 84 at discrete locations on the cable 82 where a defined bend is desired. Eventually, these control points might be precise spot-welds on the cable ring, but in the illustrated embodiment small steel pins or tubes are crimped or wrapped around a section of cable 82 and bent to the desired curvature.

FIGS. 6A and 6B show a still further core member 100 in the shape of a tricuspid annuloplasty ring. As in the earlier embodiments, exterior components such as a silicone interface and fabric cover are not shown to better illustrate the flexible core member 100. The core member 100 when in its relaxed, unstressed configuration is the same shape as an Edwards MC³Annuloplasty System available from Edwards Lifesciences.

The core member 100 includes a flexible braided cable 102 having two free ends 104a, 104b. A series of discrete control points or members 106, 108, 110, 112, 114 provide rigidity and shape the cable 102. The core member 100 has the classic tricuspid shape in plan view, starting at the first free end 104a and extending in a clockwise direction around a first segment corresponding to the aortic part of the anterior leaflet in which two control members 106, 108 are located. Adjacent to the first segment is a second segment corresponding to the remaining part of the anterior leaflet in which is located a third control member 110, the second segment ending at the postero septal commissure and a fourth control member 112. Finally, a third segment extends from about the fourth control member 112 to the second free end 56b, which is mid-way along the septal leaflet, and includes a fifth control member 114. The nomenclature for these segments is taken from the standard anatomical nomenclature around the tricuspid annulus.

As before, each of the control members 106, 108, 110, 112, 114 provides both rigidity and shape to the core member 100. For instance, the control members 106, 108, 110, 112, 114 all provide the convex curvature in plan view, and also induce the vertical deflections seen in elevational view in FIG. 6A. In the illustrated embodiment, the control members are tubular metallic crimps, but as mentioned above may be provided in different configurations.

FIGS. 7A and 7B schematically illustrate a technique for delivering an annuloplasty ring having the core member 100 in a minimally-invasive manner. Because of the open nature of the core member 100, with the two free ends 104a, 104b, the ring may be opened up or stretched out relatively straight in a stressed state as seen in FIG. 7A and inserted within a tubular access device 120. The access device 120 may be inserted through an access port in the patient's chest, for example, so that its distal end is positioned at the tricuspid annulus. The core member 100 is seen being expelled from one end of the access device 120 in FIG. 7B and immediately assuming its relaxed unstressed state. In practice, the ring will be expelled from the distal end of the access device 120 so as to assume the unstressed ring shape in approximately the proper implant location, at which time sutures or staples may be used to attach the ring to the annulus. Additional systems for delivering the annuloplasty rings described herein will be presented below.

Now with reference to FIGS. 8A-8C, a slightly different core member 130 for a tricuspid annuloplasty ring is shown. The core member 130 includes a braided cable 132 extending from a first free end 134a to a second free end 134b. A number of discrete control points or members 136, 138, 140, 142, 144 are spaced apart along the cable 132. In its relaxed state as shown, the cable 132 is in the shape of a Physio II™ Tricuspid Annuloplasty Ring soon available from Edwards Lifesciences, and includes a waveform shape with up and down regions and two upturned free ends 134a, 134b.

Instead of the tubular crimps for control points as shown above, each control member 136, 138, 140, 142, 144 includes a length of wire or cable wrapped helically around the cable 132. The wrapped wires perform the same function as the crimped metallic tube and provide both rigidity and shape to the core member 130.

The control points or members may be formed in a number of ways other than the crimped tubes and wrapped wires shown above. It is important to understand that the terms “control point” or “control member” refer to short rigid regions (regions of high modulus) on the otherwise relatively flexible (low modulus) ring. The goal of providing a number of discrete rigid regions is to add rigidity and control the final ring shape, which would be difficult with a purely flexible cable. These control points might, for example, be precise spot-welds on the cable ring, or small steel pins crimped or wrapped around a section of cable and bent to the desired curvature. In general, “control points” may be provided by tubular crimps, wound wires, welds, splices, silver solder, heat fused areas, or spot welded regions. Other possibilities include a polymer overmolded around the cable or even certain adhesives that are durable enough to withstand the repetitive flexing motion of the annuloplasty rings.

The concept of a flexible (low modulus) cable combined with carefully selected control points (regions of high modulus) allows designers to “tune” the overall effective modulus of the cable. For example, very flexible cables (e.g. Elgiloy with a moderate strand count and cable diameter of ~0.05 in), could be modified into less flexible ring geometries using careful placement of control points. Once a “target modulus” is predicted for a cable such that appropriate amounts of local displacement will occur along the ring, a variety of cable materials can be selected. Since the use of control points will dictate what the effective modulus is of a particular cable type, material selection need not be constrained by the inherent stiffness of the cable material. A flexible cable, stiffened by control points, provides the ring with sufficient flexibility to compress for delivery through a catheter, while maintaining rigidity in the deployed state. This gives designers valuable freedom, in that materials and cross section can be selected based on cost/familiarity; cable strand count and control points, rather than inherent material properties, are the key design variables.

Furthermore, and as mentioned previously, control points serve to both create the permanent 3D geometry in an otherwise flexible cable, and to locally modify the flexibility of the ring within a given region, allowing asymmetric deflection under the cardiac cycle to be designed into the product. One example of materials is a cable from FWM 1058 Elgiloy, 7×19 strand arrangement, 0.062″ diameter, with short tubular Elgiloy crimps.

FIGS. 9A-12B and 13A-16B illustrate the results of computer simulations of both closed and open rings when certain out-of-plane forces are applied with different control points.

In developing the idea of controlled bending in cables, a number of different computer models have been created and evaluated to simulate the types of forces that these rings will experience inside the heart. In particular, the simulations include a D-ring “control point” model where control points are added and changes in overall displacement are observed, and a C-ring “control point” model where control points are added and changes in overall displacement are observed. It is important to note that these models merely shed light on the concept of “control point-based cable rings” and are not completely representative of what would be seen experimentally. The major goal of these models is to show that cable rings can be manipulated to function similarly to solid-core rings, but still maintain enough flexibility to make minimally invasive (MIS) procedures possible. Also, these models demonstrate that the appropriate placement and number of control points can control both the amount and discrete location of cable displacement.

Parametric Study: Ring Bending Modulus Versus Maximum Displacement

In order to explore the potential of a cable+control points design for MIS annuloplasty rings, we have performed a parametric study of maximum displacement within a ring over a range of ring material modulus values. This model was created using the finite element analysis package COMSOL™ along with a Pro-E geometry of the Edwards generic 196869 “D” ring (mitral valve). Cardiac loads were assumed to be consistent with the forces in the z-axis, described in Table 1.

TABLE 1

CARDIAC FORCES EXERTED BY MITRAL VALVE ON D RING

Location
Force Magnitude

Anterior
0.83N

Posterior
0.73N

Posterior Commissure
−2.35N

Anterior Commissure
−2.64N

Even though the mitral valve exerts a force in the x-y plane of about 1.88 lbf, this loading condition was neglected in order to simplify the model and focus on the main displacement of the ring in the z-plane. In addition to the four loading conditions seen in Table 1, four locations on the ring were defined as constraints, or areas of zero displacement.

For the parametric model, several modulus values were evaluated for the ring under the same loading conditions. The displacement of the ring was computed for each modulus value and used to create a curve that compares the maximum displacement with the modulus value. A common metric that is useful in describing the elastic behavior of a material is the Elastic Modulus (or Young's Modulus). This value relates the stress applied to a material to the strain that it experiences through the relationship described in Hooke's law. When materials are tested in tension, a material with a lower elastic modulus will experience greater deformation than a material with a higher elastic modulus. However, since these simulations are dealing with bending forces and not tensile forces, we are instead concerned with the bending modulus (also referred to as the flexural modulus) of these cables. Similar to the trend seen with elastic moduli, materials with a lower bending modulus will bend or deflect more than a material with a higher bending modulus. Though there are ways of calculating the bending modulus of a material as a function of its elastic modulus, there is no substitute for experimental measurements of a material's bending modulus. Generally, the bending modulus of a solid-core wire is greater than its elastic modulus, whereas the bending modulus of multi-stranded cable is significantly lower than its elastic modulus.

The graph of FIG. 26 was created by tracking the displacement of the posterior commissure (found to deflect the most) over a range of modulus values. The relationship between the observable modulus and the maximum displacement can be broken down into three functionally different zones:

Zone 1, referred to as the “pure cable” zone, represents the region of low modulus values characteristic of cable. The specific modulus used in this simulation is the Bending Modulus, which is different than the tensile modulus (known as the Elastic Modulus or Young's Modulus). Though cable and solid-core wire have similar Elastic Modulus values, the Bending Modulus for cable is significantly less than for solid-core wire, (hence its greater flexibility). Under the same applied loads, a cable will deflect more than a solid-core wire, due to its lower bending modulus. In this region, one can change the allowable maximum displacement by selecting cables with different alloys, diameter, or strand count to achieve the desired modulus value. By knowing that lower modulus values correspond to greater maximum displacements, one can select an appropriate cable for a given application.

Zone 3, referred to as the “pure solid-core” zone, represents the region of high modulus values that are characteristic of solid-core wire. When given the same loading conditions as a ring made of cable, a solid-core ring will experience much less overall displacement. In addition, since solid-core wire does not have the inherent flexibility of cable, deformation that occurs will likely be permanent (when compared to cable).

Zone 2, referred to as the “hybrid” zone, represents high potential interest as the intermediate region where rings can be manufactured to take advantage of the overall flexibility of pure cable, but maintain areas of structural rigidity seen in solid-core wire. In this region, low-modulus cables can be “adjusted” to an effective modulus which is greater than their native modulus by introducing control points-point-welds along the ring that can be assumed to have a local modulus that resembles a solid-core wire. Since areas of “pure cable” remain between these control points, the ring will still exhibit much of the same flexibility as pure cable. As more control points are introduced, the ring will exhibit a higher effective modulus until it eventually approximates the modulus of a solid-core wire (this would be the case with an infinite number of control points).

This hybrid region represents the “tunable” range one can utilize by introducing point welds into the cable ring rather than selecting a different material, different thickness, or different strand count. By choosing appropriate locations for these control points, the deformation allowed in each plane can be controlled in addition to the maximum limit.

Control Point Study: D Ring, FIGS. 9A-12B

In this study, we examined the effects of adding control points on localized displacements, paying attention to the areas of displacement as well as the maximum values. For this simulation, the same geometry and loading conditions described previously for the parametric study were used. Instead of adjusting modulus values throughout the simulation, we selected values representative of a semi-flexible cable and control points and used these values throughout. The cable bending modulus used was 6E8 Pa (about 8.7E4 psi), taken from literature values as a typical modulus near the lower end of the cable range. We used a control point modulus of 2E22 Pa in order to approximate a region with a “near-infinite” bending modulus, as bending within the weld would not be expected if the weld was centered at a distinct point. We also compared the control point model to a similar ring model representing solid-core wire with no control points with a bending modulus of 1.027E10, an order of magnitude less than the elastic modulus for commercially pure titanium (FWM product info).

So, for example, FIG. 9A shows the relaxed shape of a flexible ring 150 having no control points, and FIG. 9B shows the ring shape 12 after having been subjected to the three vertical force arrows shown. FIG. 10A is a ring 154 with two control points 156, and FIG. 10B is the shape 158 after loading with the three vertical forces. FIGS. 11 and 12 continue the progression with more control points 162, 168, and the resulting shapes under load are seen decreasing in FIGS. 11B and 12B. The most obvious trend throughout this study is that as more control points are added, the overall displacement of the ring decreases. Localized displacement tends to decrease the most around areas where control points are added as seen between FIGS. 10B and 11B. Since adding more control points will inherently form a ring that is more representative of a solid-core ring, we expect that overall displacement will decrease for each additional control point added. The important message to take away here is that, by controlling the placement and amount of control points, one can design a cable ring that has regions of controlled displacement. The control points are analogous to points on a spline curve, where each point controls how the line curves.

Control Point Study: C Ring, FIGS. 13A-16B

FIGS. 13A-16B show open or C-shaped tricuspid rings having none, one 186, two 192, and three 198 control points. The corresponding simulated loaded shapes are seen in FIGS. 13B, 14B, 15B and 16B.

The C ring displacement model was very similar to the D model previously described, except that a different loading scheme was used. Instead of 4 independent forces acting on the ring, as seen in the previous model, the C ring model only used one force in the z plane. In reality, one would expect to see the two free ends of the C ring exhibit some displacement since they are sutured to the aortic root and thus part of the contracting heart. However, these ends were modeled as constraints to simplify the model and focus primarily on the effects of adding control points to the C ring as it is pulled down on the anterior end, as seen in FIGS. 13B, 14B, 15B and 16B. The force created by the cardiac cycle was represented by a single force pulling the ring down in the negative z-axis from the anterior end. The force magnitude used was 0.6 N, a little more than half of the anterior force created by the mitral valve. The same modulus values described for the D model, for pure cable and for the control point regions, were used for the C model.

The largest different between the D and C ring results is that the C ring approximated zero displacement with only 3 control points whereas the D ring required about 6. The main cause of this difference is the geometry of the two rings, namely that the C ring is constrained near its midpoint and only has one load throughout the entire geometry. Since the D ring model is less constrained than the C ring model, it has more opportunities to distribute the applied loads intro corresponding displacements. However, we still see the same trend, where adding more control points decreases not only the local z-displacements but the overall displacements as well.

FIGS. 17A-17G show a number of different braided wire configurations that may be used. These include: a simple 1×3 cable in FIG. 17A, a simple 1×7 cable in FIG. 17B, and a simple 1×19 cable in FIG. 17C. Multi-stranded cables include multiple braided cables braided with one another, and include: a 7×7 cable in FIG. 17D, a 7×19 cable in FIG. 17E, a 19×7 cable in FIG. 17F, and a 7×7×7 cable in FIG. 17G. Each of these cables comprises many individual strands that are twisted around each other whereas solid-core wire is composed of a single strand. Even though wide ranges of materials and alloys can be used for both, cable is much more versatile than solid-core wire since different alloys can be used for different strands, different strand counts and geometric placements can be used, and different amounts of coiling can be used. This contrasts the basic nature of solid-core wire where only a single alloy can be used. Because of this unique geometry, cables are typically stronger than wire and yet are also more flexible. When pulled in tension from both ends, cable acts similarly to wire since the different strands are all being pulled in the same direction. However, when a cable is bent, the different strands are allowed to slide past each other slightly, which creates spaces for other strands to occupy and thus is much more flexible than a solid-core wire with the same overall diameter. It is this unique property of cable that makes it an attractive alternative to solid-core wire with respect to annuloplasty rings for minimally invasive surgery. More information on medical grade cables is available from Fort Wayne Metals headquartered in Fort Wayne, IN. In particular, some cables may be coated with inert polymers for greater biocompatibility.

Although the present application contemplates using both simple (i.e., single braided) and multi-stranded (i.e., multiple braids intertwined) cables, multi-stranded cables are believed better suited for the MIS delivery approach. For open rings, simple cables may be easily stretched linearly for passage through an access tube, but once permitted to relax and resume the annuloplasty ring shape, these simple cables may not have the requisite stiffness for annulus remodeling. As such, a greater number of control points would have to be used, which may place undesirable limitations on overall ring performance. Furthermore, simple cables formed into closed rings may not be able to be squeezed into a linear shape without kinking into permanent bends. On the other hand, multi-stranded cables are more flexible in bending due to their generally smaller individual strands and the ability of those strands to slide with respect to one another. Moreover, in open rings multi-stranded cables retain larger stiffness in the plane of the ring to provide good remodeling without an excessive number of control points.

Preliminary Evaluation of Fort Wayne Metals Cable Samples

A. Semi-Quantitative Analysis of Cable Samples

A series of cable samples, representing typical standard products for biomedical applications, was provided by Fort Wayne Metals (FWM). Table 2 summarizes physical properties of the samples. It should be noted that these are not the only materials contemplated, and the list of suitable materials includes alloys of stainless steel, Titanium, Cobalt Chromium, Nitinol (NiTi) and Platinum-Iridium. Further, blends or combinations of these various materials could be utilized to obtain particular performance characteristics. The number of permutations is essentially limitless.

TABLE 2

CABLE SAMPLES PROVIDED BY FWM

Sample
Material
Diameter (in)
Strand Count

1
Ti 6Al 4V ELI
0.0375
19 × 7

2
Ti 6Al 4V ELI
0.0423
7 × 7

3
L-605
0.0625
19 × 7

4
L-605
0.080
7 × 7

5
FWM-1058
0.062
7 × 19

6
316 LVM
0.078
7 × 7

7
316 LVM
0.0475
1 × 19

8
316 LVM
0.0425
1 × 7

9
MP35N
0.063
7 × 7

10
FWM-1058
0.125
7 × 19

A preliminary, semi-quantitative analysis was performed on these samples to determine issues with cable material, diameter, and strand count with respect to the control point concept. FIG. 11 illustrates the experimental setup. A minimum bending diameter was determined visually, by bending the cable sample back upon itself until either permanent deformation occurred or cable strands began to separate. At this orientation, measurements were taken by a caliper. The force required to hold this minimum bending diameter was estimated by manually applying the necessary load while the cable was resting on a laboratory scale. Additionally, the cable samples were evaluated for minimum bending diameter with moderate deformation (defined as a ~10 degree bend remaining in the cable after removing load), as well as “robustness”, which was based on qualitative observation of how much bending/deformation cables could withstand without suffering permanent damage (kinking, strand separation, or permanent deformation). The results of this preliminary analysis are presented in Table 3.

TABLE 3

RESULTS OF SEMI-QUANTITATIVE ANALYSIS ON C

ABLE SAMPLES PROVIDED BY FWM.

Min Dia

Def. Dia

Sample
(mm)
Force (g)
Robustness
(mm)

1
6.9
48
F
4.8

2
9.5
130
G
6.5

3
14.9
228
G
9.4

4
25.4
460
G
13.7

5
12.1
185
G
8

6
20.4
560
G
12

7
16.2
480
F
10.7

8
22.8
580
P
20

9
17.6
385
G
9.9

10
16.5
410
G
10.5

Results in Table 3 may be sorted to identify good (G), acceptable or fair (F), and poor (P) values with respect to the features necessary for use in MIS Annuloplasty Rings. As discussed previously, the ideal characteristic is for a cable to be sufficiently flexible to compress for delivery through a catheter, yet maintain rigidity in the deployed state. Given this, samples that had a minimum bending diameter of <10 mm were considered good, while those with a minimum bending diameter of >20 mm were considered poor. While force to maintain this bending diameter is not a direct measure of cable bending modulus, it is a reasonable indirect measure; for this reason, an arbitrary value of >400 g was considered good, while <200 g was considered poor. One noticeable result was that low-strand-count cables (#7 & #8), were considerably less robust compared to the higher strand count cables.

Among these cable samples, samples 2, 3, 9, & 10 had the best overall relative combination of stiffness, compressibility, and robustness. While it is premature to form specific cable selection recommendations, qualitative observations and this data suggest that a cable diameter of less than 0.08 in, combined with a strand count of 7×7, 7×19, or 19×7, is best suited for the control point concept. Material type is a secondary consideration.

B. Cable Selection Considerations

Preliminary evaluation of FWM samples are consistent with the results of computer simulations, with both indicating that a wide variety of cable materials could be used for annuloplasty ring applications. Section I.D. discussed “tuning” the overall effective modulus of the cable through carefully selected control points. Since the use of control points will dictate the effective modulus of a given cable type, material selection is not constrained by the inherent stiffness of the cable material. A likely cable selection strategy is to:

Select material based on availability/familiarity.

Select cable diameter to be similar in diameter to current “solid-core” rings.

Select a standard, off-the-shelf cable, with moderate strand count and low bending modulus, to achieve maximum compression for delivery through catheter.

Add control points necessary to form cable into required three-dimensional geometry.

Add additional control points and/or increase length of control points to achieve required effective modulus and desired local maximum displacements along ring.

Iterate with greater strand count if local maximum displacements are too great.

Thus a flexible cable, stiffened by control points, provides the ring with sufficient flexibility to compress for delivery through a catheter, while maintaining rigidity in the deployed state. Prototypes have been constructed employing this strategy (low modulus+sufficient control points to stiffen the ring). It is also possible to combine multiple cable types to achieve the combination of high bending for deployment as well as high post-deployed stiffness.

FIGS. 18A-18C are side, posterior, and top plan views, respectively, of an alternative flexible open mitral annuloplasty ring 220 with control points. The annuloplasty ring 220 includes a flexible multi-stranded cable 222 having two free ends 224. In the illustrated embodiment the free ends