An image-based framework for the biomechanical analysis of mitral valve prolapse repair

Background. Isolated posterior leaflet prolapse is the most frequent pathologic dysfunction of degenerative mitral valve (MV) prolapse; well-established and conservative surgical techniques, based on artificial neochordae implantation, are nowadays available as well as minimally invasive and percutaneous strategies can be adopted. On the one hand, neochordoplasty is less invasive than MV resection and allows for “respecting rather than resecting” the diseased MV leaflet; however, a long-standing experience is required to intraoperatively select the appropriate neochordal configuration and precisely assess neochordal length. Hence, the identification of sound criteria for the selection of the appropriate solution is mandatory to achieve repeatable and standardized results. On the other hand, the Mitraclip® device is the only percutaneous system for edge-to-edge currently approved for clinical use in humans and feasible as a standalone approach in some clinical scenarios. Despite its conceptual ease, the procedure of implantation is still complex requiring skilled operators, a high-level overall clinical staff and a demanding learning curve to efficiently perform the procedure. Image-based framework. A patient-specific modeling strategy was developed for the assessment of MV prolapse from in vivo cardiac magnetic resonance imaging (cMRI). Subsequently, finite element (FE) models were run to quantitatively compare, for the first time and on a patient-specific basis, the biomechanical effects of a broad spectrum of different neochordal implantation techniques for the repair of isolated posterior MV prolapse, focusing on the impact of surgical precision in setting neochordal length on postoperative MV biomechanics. At this aim, segmentation of rotational cine long-axis cMRI planes was performed on 15 patients with degenerative MV prolapse and on 10 healthy subjects and ad hoc post-processing allowed to assess several morphological and functional MV parameters. The motion of annulus and papillary muscles (PM) was quantified, and integrated with intraoperative surgical details to assess location and extent of the prolapsing region. Four patient-specific and FE structural MV models were derived, reproducing the end-diastolic MV configuration, realistic annular and PMs kinematic boundary conditions, and including hyperelastic and anisotropic mechanical properties of tissues and non-linear response of artificial neochordae. Simulations were run with the commercial solver ABAQUS/Explicit. Preliminary FE results reproduced both the preoperative systolic prolapsing MV configuration and the physiological model with an intact chordal apparatus, i.e. providing the physiological “ideal” distance between the insertion of the ruptured chordae and the corresponding papillary tip. An “ideal” setting and several “sub-optimal” though realistic settings of neochordal lengths were then simulated on different neochordal procedures ranging from single neochorda to “premeasured” loop configurations. All “sub-optimal” configurations re-established adequate valvular competence but differed in millimetric neochordal lengths on a range of ± 2mm, compared to the corresponding “ideal” repair. Finally, a novel computational approach to the most recent Mitraclip® percutaneous technique was proposed to simulate MV biomechanics, throughout the entire cardiac cycle, before and after the device implantation, assessing the biomechanical impact of the procedure on MV apparatus. Finite element evidences. Significant changes (p<0.05) were noticed in prolapsing patients, with respect to healthy subjects, in terms of mitral annular perimeter and area, MV billowing height and volume. PMs kinematics and MV longitudinal apically displacement remained comparable. Image-based FE models reliably reproduced MV prolapse alterations and well compared to ground truth from cMRI images. Also, FE modelling strategy proved able to quantify MV biomechanics associated to post-surgical conditions. Tested “ideal” neochordal configurations effectively fixed MV prolapse with satisfactory levels of coaptation area although the outcomes of MV repair proved to be technique-dependent. Indeed, implantation of multiple neochordae improved the repositioning of the prolapsing region: chordal tension of the prolapsing region was partially transferred from the intact native chordae to the artificial neochordae, thus improving mechanical stress distribution on the posterior MV leaflet. In “sub-optimal” MV plasty, despite absence of residual regurgitation, a millimetric sub-optimal suture length markedly impacted on MV biomechanics. A 2 mm-shorter single neochorda dramatically increased mechanical leaflet stress and neochordal tension. In double neochordal plasty, if a neochorda was 2mm-longer, the other neochorda was overloaded whereas opposite results occurred with a 2mm-shorter neochorda, both resulting at the end in asymmetric load redistribution. In neochordal loop configuration, when only the length of one neochorda was altered, a notable increase in leaflet stresses was obtained with load entirely settled on a single suture. Mitraclip® implantation significantly improved systolic leaflets coaptation, without inducing major alterations in systolic peak stresses. However, diastolic orifice area markedly decreased and leaflets diastolic stresses became comparable, although lower, to systolic ones. Conclusions. The present work clearly showed that mitral neochordoplasty largely depends on the intraoperative assessment of the appropriate length of artificial sutures. For this purpose, patient-specific numerical models can: i) explain the biomechanical drawbacks due to a “sub-optimal” millimetric tuning of the suture length, ii) elucidate their relationship with patient-specific features and neochordal suture lengths, and iii) pave the way to a more reproducible and effective surgery. According to FE results, it is possible to define in the single clinical scenario a theoretical patient-specific “gold standard” neochordal technique assessing the relief of tension on the native chordae, the degree of stress redistribution on MV leaflets, and peak mechanical stresses observed along the leaflet free margin. However, since these biomechanical variables are not directly measureable by surgeons, postoperative changes in MV biomechanics assessed with cMRI-derived FE models can represent a valuable background to deepen the biomechanical implications of MV repair. Moreover, clinically relevant indications were obtained applying the developed modeling FE approach to the analysis of Mitraclip® device: this allowed for computing biomechanical variables of clinical interest with reference to a real world clinical scenario and providing a rather exhaustive analysis of the biomechanical effects of the Mitraclip® device on MV apparatus, combining for the first time the morphological information provided by cMRI with a reasonably detailed description of tissues mechanical properties and boundary kinematic constraints. In conclusion, the present work represents the first effort towards a virtual patient-specific assessment of MV repair: the FE approach herein proposed may be useful in order to quantitatively assess the efficacy of MV repair strategies and, if further tested with longitudinal studies, provide a deeper insight into the potential mechanisms of recurrent mitral regurgitation, thus promoting an on-going patient-specific optimization of neochordal techniques. The modeling FE approach may also be expanded to analyze clinical scenarios that are currently critical for Mitraclip® implantation, e.g. functional mitral regurgitation and annular dilation, helping the search for possible solutions.