Patient-specific Ventricle Models And Virtual Surgery For Optimization Of Tetralogy Of Fallot Patient Pulmonary Valve Replacement Surgery

Tetralogy of Fallot(To F)is a common congenital heart disease.To F patients have severe structural cardiac defects at birth.Open-chest heart surgeries are often needed so that those patients(infants at that time)could survive.Since the size of artificial valves could not change as the patients grow,pulmonary valve insertion/replacement surgery(PVR)is usually performed after the patients reach their adulthood.However,some patients had right ventricle(RV)ejection fractions(EF)improvements post PVR,while some other patients had no EF improvements.Currently,the reason accounting for this phenomenon remains unknown.Seeking potential factors influencing RV function after PVR and exploring methods predicting post-PVR results have become a major clinical challenge.Cardiac magnetic resonance(CMR)images and clinical data from 22 subjects including 6healthy volunteers and 16 To F patients(pre-and post-PVR)were collected to build bi-ventricle mechanical models,perform ventricle mechanical analysis,investigate mechanisms of cardiac functions,and identify predictors for post-PVR results for To F patients.Furthermore,a novel PVR surgical strategy using active contracting bands was introduced to optimize post-PVR outcome.Patient-specific virtual surgeries using biomechanical ventricle models were performed to test the feasibility of the contracting band PVR procedures.Using virtual surgery to avoid testing risky novel surgical procedures on real patients is an important approach in surgical research.Patient-specific ventricle biomechanical models using different diastole and systole zeroload geometries were constructed to reflect active contraction of the myocardium.These models were called two-geometry models since two zero-load geometries were used.In-vivo myocardium material properties,RV inner wall stress and strain values were quantified.Results from two-geometry models and one-geometry models were compared to show their differences.One-geometry models which used the same diastole and systole zero-load geometries were used in most publications.At begin-ejection,mean myocardium effective Young’s Modulus(YM)of RV from 22 two-geometry models was 204.54±227.73 k Pa,mean RV stress was 63.00±25.20 k Pa and mean RV strain was 0.544±0.119.The corresponding RV YM and stress/strain values of one-geometry models were 429.23±464.99 k Pa,60.20±24.17 k Pa and 0.440±0.117,respectively.Wilcoxon signed rank tests indicated that the differences between one-geometry and two-geometry models were statistically different(p<0.001).Knowing the differences of RV stress,strain and myocardium material properties between To F patients and healthy people can be of great help to better understand the mechanism of the disease and find more efficient treatments.These 22 subjects were classified into two groups:6 healthy volunteers and 16 To F patients.At begin-ejection,mean RV myocardium effective YM of 6 healthy volunteers was 45.55±17.08 k Pa,mean RV stress was 44.59±4.63 k Pa and mean RV strain was 0.64±0.048.The corresponding mean RV myocardium effective RV YM,stress and strain values from 16 To F patients were 264.16±241.17 k Pa,69.9±26.28 k Pa and0.508±0.118,respectively.Wilcoxon rank sum test was used to compare the differences between these two groups.RV of 16 To F patients had harder myocardium tissue material properties(p=0.01),higher stress(p=0.03)and strain values(p=0.03)than 6 healthy volunteers.Predicting RV cardiac function after PVR is a great challenge in the clinical study of To F.CMR images before and after PVR from 16 To F patients were used to construct patient-specific biomechanical ventricle models.Ventricle geometry data and biomechanics data from ventricle mechanical models were used as predictors and logistic regression model was used as the method to predict post-PVR results.RV inner wall stress at end-filling was identified as the best predictor.The prediction accuracy was 84.9% for this predictor(RV stress at end-filling).The treatment of RV dysfunction for To F patients is a great challenge for all clinicians and researchers worldwide.An innovative band insertion PVR strategy was further studied using data from 7 To F patients and 5 different band insertion plans.A total of 140 virtual surgery mechanical models were constructed to simulate post-surgery(band insertion surgery)outcome and find the best surgical plan.The 5 band insertion plans used 1-3 contracting bands and different band locations.Band active contraction ratios were set as 0%(passive contraction),10%,15% and 20%.Simulation results indicated that RV EF increased after contracting bands were inserted.The surgery plan with 3 bands and 20% band active contraction ratio was the optimal plan.The mean RV EF of the 7 patients improved to 42.9%±5.68% from38.71%±5.73%.Fluid-structure interaction(FSI)models were constructed to investigate the influence of band insertion on blood flow in RV.There has been no publication in the current literature for FSI simulations of PVR using multi-band insertions.Blood flow in RV changed after the bands were inserted.However,results from FSI models indicated that flow shear stress on the surface of contracting bands was well under 2.2 Pa,flow shear stress in right ventricle was under 4.0Pa,which would not cause damages to blood cells.Pre-and post-PVR CMR and clinical data are extremely difficult to obtain.Patient-specific three-dimension ventricle model construction is also very time consuming.Compared to modeling studies using one-patient data,investigations reported here used multi-patient data and results obtained(mechanical analysis and model comparisons,post-PVR cardiac function predictions,contracting band feasibility study)would have better clinical relevance.The initial results indicated that ventricle mechanical models have wide application potential in clinical diagnosis and surgery optimization.Those preliminary results still need to be validated by more clinical data.The innovations of this dissertation include:(a)Two zero-load geometry active contraction models were used to calculate and compare RV stiffness,stress and strain for multiple To F patients and healthy volunteers;(b)RV wall stress was identified as the best predictor for postPVR EF outcome;(c)Multi-patient multi-band novel PVR surgery strategy using active contraction bands was proposed and virtual surgeries were performed using patient-specific ventricle mechanical models.This demonstrated some initial feasibility evidence of the novel band insertion surgical strategy;(d)Fluid-structure models were used to evaluate the change of blood flow inside right ventricle.