In vivo documentation of shape and position changes of MRI-visible mesh placed in rectovaginal septum
Abstract
Background and objective: Large deformations in synthetic meshes used in pelvic organ prolapse surgery may lead to suboptimal support for the underlying tissue, graft-related complications as well as recurrence. Our aim was to quantify in vivo longitudinal changes in mesh shape and geometry in a large animal model.
We compare two commonly used mesh shapes, armed and flat, that are differently affixed. The secondary outcomes were active and passive biomechanical properties.
Methods: A total of 18 animals were used. Six each were implanted With either an arm mesh, a flat mesh or underwent a sham surgery.
PVDF meshes loaded with Fe2O3 were used to facilitate their visualization in vivo. MR images were taken at 2, 14 and 60 days after implantation and 3D models of the meshes were created at each time point.
We calculate the Effective Surface Area (ESA), i.e. the support that the mesh provides to the underlying tissue using custom developed techniques. Longitudinal changes in the mesh shape were studied by comparing the respective 3D models using part comparison analyses.
The root means square difference (RMSD) and the modified Hausdorff distance (MHD) were calculated to obtain an objective value for the part comparisons. Wall thickness maps were produced on 3D models.
Mesh arm length and their ellipticity profiles were also evaluated. Active and passive biomechanical tests on vaginal tissue overlaying the mesh were conducted using a contractility assay and a uniaxial loading protocol.
Results: MR images of 5 animals in each group were used for longitudinal comparison. Compared to the initial implant size, there was an immediate drop in the ESA measurement at day 2 of almost 32.22 [7.06] % (median [IQR]) for flat meshes, and by 17.59 [6.50] % for arm meshes.
After 14 days, the reduction in area was 41.84 [14.89] % and 27.18 [20.44] %, and at day 60 it was 36.61 [6.64] % and 26.43 [14.56] % for the flat and armed meshes respectively. The reduction in area in the two groups was different between the two groups only day 14 (p = 0.046).
The ellipticity of the arms was 0.81 [0.08] (median [IQR]) and there was no significant change in the ellipticity profiles over time. The mesh arm length did not change significantly over time.
The part comparison showed a maximum difference of 4.26 [3.29] mm in 3D models according to the MHD measure, which is clinically not relevant. Comparison of high thickness areas on the thickness maps correlated well with the areas of mesh folding in the arm mesh group observed during postmortem dissection.
Thickness maps did not help us understand why the flat meshes had a reduction in support area. The comfort zone stiffness of the flat mesh and of the central part of the arm mesh were 2.4 fold and 4.5 times stiffer compared to sham groups, respectively.
The arms were 36% stiffer than the central part of the mesh. The comfort zone length of the sham group was 46% longer than the flat mesh group (p = 0.027) and 59% longer than that of the central part of the arm mesh (p = 0.005).
There was no significant difference in vaginal contractile forces generated in samples from the ann, flat mesh, and sham groups. Conclusions: This is a first longitudinal study observing deformations in vaginally implanted synthetic meshes in a large animal model.
A novel methodology is presented to calculate the area of the vaginal tissue effectively supported by the mesh implant. Immediately post-operatively, a reduction in 32% and 17% was noted, which remained stable over the 60 following days of observation.
We use thickness maps to analyze the cause of this dramatic immediate reduction. In the armed mesh we found it to be mesh folding at the interface between the arms and central part.
For the flat mesh we suggest that pore aggregation during suturing.