The goal of health care systems Primary goal of health care policy - PDF document

Computational biodesign “cardiovascular devices” Prof. Pascal Verdonck IBiTech - Institute Biomedical Technology Ghent University The goal of health care systems • Primary goal of health care policy = to maximize the health of the population within the limits of the available resources, and within an ethical framework built on equity and solidarity principles. Report of the Belgian EU Presidency, adopted by the EU Council of Ministers of Health in Dec 2010 2 1

A great day: market approval ) Testing Full -- Human Post Market Develop Verification Tests Concept Feasibility Market Release ment -- -- Surveillance Process (Validation) (Europe) optimization A new dimension in the cycle: The customer Who is the customer ? • Health care system • Physician • Reimbursement • Patient • Insurance company • Regulatory bodies • …. The wisdom of the customer • Risk Management • User friendliness • Quality control • Development of the device: start from the end ! The Wisdom of the Crowd: “ A group of (divers) individuals has always more intellect than one expert” th Aristoteles, 4 c BC 2

High quality design requires ongoing user input ! ) User Testing Full -- Human Post Market Develop Verification Tests Feedback / Concept Feasibility Market Release ment -- -- Surveillance Process (Validation) (Europe) Interactions optimization Risk Management RV LV RA AO LA 3

Computational Fluid Dynamics CFD • Computational Fluid Dynamics • Transport phenomena: fundamental principles of mass, momentum and energy conservation expressed in algebraic, differential or integral representation • Numerical description in space and time • Design tool in aerospace, automotive and process industry 4

The role of computational fluid dynamics for heart valve design Kris Dumont, Benjamin Vandersmissen, Sebastiaan Annerel, Patrick Segers, Jan Vierendeels, Pascal Verdonck 5

Experimental model Experimental model Torsie Hoek draden 26° 45° 22° 55° 17° 70° Draden ifv torsie 80 Draden ifv torsie 70 60 50 Lineair (Draden ifv 40 torsie) 30 y = -2,7869x + 117,05 20 10 0 10 15 20 25 30 6

Experimental model Model 1 Model 2 Model 3 7

Experimental model 1 2 3 4 5 6 a b 7 8 9 Experimental model 8

CFD & Heart Valves FSI Hemodynamic Results 9

Aim of the Study • Comparison of ATS Open Pivot Valve in mitral versus aortic position • Applying a dedicated fluid-structure interaction (FSI) code 1. K. Dumont, J.M.A. Stijnen, J. Vierendeels, F.N. van de Vosse and P. Verdonck. Validation of a fluid-structure interaction model of a heart valve using the dynamic mesh method in Fluent. Computer Methods in Biomechanics and Biomedical Engineering: 7: 139-146, 2004. 2. K. Dumont, J. Vierendeels, P. Segers, G. Van Nooten and P. Verdonck. Predicting ATS Open Pivot TM Heart Valve Performance with Computational Fluid Dynamics Journal of Heart Valve Disease: 14(3): 394-399, 2005. 3. J. Vierendeels, K. Dumont and PR. Verdonck. Stabilization of a fluid-structure coupling procedure for rigid body motion. AIAA journal: 43(12) 2549-2557, 2005. • Derive clinical relevant parameters, shear stress distribution using our FSI code. 10

Aim of the Study • Comparison of ATS Open Pivot Valve and St Jude Regent Valve with computational fluid dynamics CFD • Applying a dedicated fluid-structure interaction (FSI) code 1. K. Dumont, J.M.A. Stijnen, J. Vierendeels, F.N. van de Vosse and P. Verdonck. Validation of a fluid-structure interaction model of a heart valve using the dynamic mesh method in Fluent. Computer Methods in Biomechanics and Biomedical Engineering: 7: 139-146, 2004. 2. K. Dumont, J. Vierendeels, P. Segers, G. Van Nooten and P. Verdonck. Predicting ATS Open Pivot TM Heart Valve Performance with Computational Fluid Dynamics Journal of Heart Valve Disease: 14(3): 394-399, 2005. 3. J. Vierendeels, K. Dumont and PR. Verdonck. Stabilization of a fluid-structure coupling procedure for rigid body motion. AIAA journal: 43(12) 2549-2557, 2005. • Derive clinical relevant parameters, shear stress distribution and platelet activation using our FSI code. 11

Introduction (a) 22mm AP ATS Open Pivot Valve Geometry (b) 21mm SJM Regent Valve Geometry FSI Hemodynamic Results 12

Particle Track Results: Forward Flow click on figure to play movie Particle Track Results: Forward Flow Stress Acumulation during Laminar Forward Flow 1 00 2 90 1 .8 Different scale in 80 1 .6 percentage axis percentage of particles percentage of particles 70 1 .4 ATS 60 1 .2 SJM 50 1 ATS 40 0.8 SJM 30 0.6 20 0.4 1 0 0.2 0 0 0 - 5 5 -1 5 1 5 - 25 25 - 35 accumulation intervals in dyne.s/cm 2 13

Particle Track Results: Regurgitation Flow click on figure to play movie Particle Track Results: Regurgitation Flow Stress Accumulation during Laminar Regurgitation Flow 1 00 3 activated 90 2.5 Different scale in 80 percentage axis percentage of particles percentage of particles 70 Hellums 2 ATS Threshold 60 SJM 50 1 .5 ATS 40 SJM 1 30 20 0.5 1 0 0 0 0 - 5 5 - 1 5 1 5 - 25 25 - 35 35 - 45 45 - 55 55 - 65 accumulation intervals in dyne.s/cm 2 14

The role of Finite Element Analysis for stent design Peter Mortier, Matthieu De Beule, Benedict Verhegghe, Pascal Verdonck The achilles heel of stenting is restenosis … 1 day follow-up 180 day follow-up 15

Balloon expandable stents … Cordis Jostent M Jostent Graft ACS RX Duet Jostent Bifurcation Cross Flex LC MiniCrown Penta Pixel Rstent Tenax Tensum 2 Terumo Tetra Velocity Vision Self expandable wire stents … 16

Drug eluting stents … Courtesy ‘t Veer Restenosis in 5 to 30% of the treated lesions … … because of:  Mechanical vascular injury  Non-uniform strut distribution 17

Mechanical vascular injury Courtesy Migliavacca B: “Dogboning” Stent expansion in a non-uniform ends-first manner (dogboning) Courtesy Squire, MIT, 2000 causing high local stresses A: “Balloon – artery” contact Mechanical vascular injury C: “Foreshortening” Δ L Courtesy Squire, MIT, 2000 Courtesy Fortimedix 18

Restenosis after DES partially related to non uniform strut distribution due to Courtesy Hwang et , Circulation A: Inadaquate vessel support B: Sub-optimal drug delivery Why computer simulations? Experimental approach is difficult because of small scale Simulations can accelerate device design Simulations provide additional information (FDA, CE approval) 19

Virtual design loop Traditional approach Our approach Parametric adaptable Parametric adaptable CAD model mesh Mesh generation FEA FEA ADVANTAGES : - two steps instead of three - can be fully automated Virtual Analysis of the mechanical properties • Ideal stent = low foreshortening Δ L Δ L 20

Virtual analysis of the mechanical properties • Ideal stent = low elastic recoil D = 3.3 mm D = 3.2 mm Virtual analysis of the mechanical properties • Ideal stent = high flexibility 21

Virtual analysis of the mechanical properties • Ideal stent = no dogboning Courtesy Migliavacca Virtual analysis of the mechanical properties ° Mortier et al., J. Biom. Eng., 2008 • Foreshortening Can be used to: • Elastic recoil - Compare different existing • Flexibility stent designs • Dogboning - Develop new stent designs • … 22

Personalized medicine Personalized medicine 23

Personalized medicine Personalized medicine 0.4 0.3 0.2 0.1 0.0 Stress [MPa] 24

Opening of sidebranch Opening of sidebranch 25

Opening of sidebranch Taxus Liberté Endeavor Cypher Select (Boston Scientific) (Medtronic) (Cordis, J&J) Gelijkaardig productconcept! From design to clinic Tools are available and validated to analyse a priori stent properties and mechanical behavior Material properties SS316L, CoCr, Mg, Nitinol Dedicated stent design + Stent requirements Flexibility Radial strength,… 26

Simulation based pre-operative planning REPORT CT images (pre- 3D reconstruction Finite element simulation of operative) TAVI Patent application pending Simulation based preoperative planning: case TAVI Peter Mortier, Matthieu De Beule, Benedict Verhegghe, Pascal Verdonck 27

Limitations of current TAVI planning tools Current planning is based on an evaluation of the aortic anatomy using CT imaging Evaluating the anatomy gives only limited insights into the risk on complications related to device / host interaction • Aortic Regurgitation (AR) • Annular rupture • Coronary obstruction • Conduction problems Messika-Zeitoun et al., 2010, JACC Limitations of current TAVI planning tools – case study: John Doe 28

Limitations of current TAVI planning tools – case study: John Doe Dcirc = 22.8 mm Limitations of current TAVI planning tools – case study: John Doe Moderate AR 29

Limitations of current TAVI planning tools – case study: John Doe Moderate AR Could this have been predicted??? Solution: simulation-based TAVI planning FEops developed a web-based pre-operative planning service, called TM TAVIguide , which combines pre-operative CT imaging with advanced computer simulations allowing to predict stent frame deformation and paravalvular regurgitation TM TAVIguide is not yet commercially available 30