Patient-specific models of dyssynchronous heart failure for assessment of regional work in patients undergoing cardiac resynchronization therapy
Data files
Nov 21, 2024 version files 4.32 GB
-
CircAdapt_Hemodata_Columns_Units.xlsx
11.05 KB
-
README.md
10.62 KB
-
Simplified_Work_Estimate_Units.xlsx
14.96 KB
-
Simplified-Work-Estimates.zip
207.02 MB
-
Simulation_Data.zip
4.12 GB
Abstract
Cardiac resynchronization therapy (CRT) can improve the synchrony of contraction between the right and left ventricles in DHF and reduce morbidity and mortality and increase quality of life. However, regional work before CRT and changes with CRT are not well understood. We used three-dimensional multi-scale patient-specific computational models parameterized by anatomic, functional, hemodynamic, and electrophysiological measurements to assess changes in eight patients with heart failure and left bundle branch block (LBBB) who received CRT. To increase clinical translatability, we also explored the utility of streamlined methods to estimate regional myocardial work. We found that CRT increased global myocardial work efficiency with significant improvements in non-responders. Reverse ventricular remodeling after CRT was greatest in patients with the highest heterogeneity of regional work at baseline. Decreases in early-activated regions of myocardium performing negative myocardial work with CRT best explained patient variations in reverse remodeling. These findings were also observed when regional myocardial work was estimated using ventricular pressure as a surrogate for myocardial stress and changes in endocardial surface area as a surrogate for strain. These findings suggest that CRT promotes reverse ventricular remodeling in human dyssynchronous heart failure by increasing regional myocardial work in early-activated regions of the ventricles and that measurement of these changes can be performed using streamlined approaches.
README
Paper: Patient-Specific Models of Dyssynchronous Heart Failure for Assessment of Regional Work in Patients Undergoing Cardiac Resynchronization Therapy
Authors:
Amanda Craine, Adarsh Krishnamurthy, Christopher T Villongco, Kevin Vincent, David E Krummen, Sanjiv M Narayan, Roy CP Kerckhoffs, Jeffery Omens, Francisco Contijoch, Andrew D McCulloch
Description:
This README file describes the data package accompanying the publication above. The data is split into two parent files corresponding to two different features of this study: Patient-Specific Modeling, and Simplified Work Estimates.
File Structure
Patient-Specific Modeling Dataset:
Files:
Simulation Data folder: This folder contains the simulation results for each subject of this study, separated by subject. Below is an example of the file structure and its contents for one subject. The same file structure applies to all subjects (BiV1-8).
BiV1: Biventricular simulation data for subject 1
Parameter1: Simulation results at baseline (labeled LBBB)
- BiV1_CT_unloaded_DT_CA_LBBB_FullBeat.cont6: Continuity file of biventricular model at baseline, generated with patient-specific parameters
- FullBeat10.zip: All deformed meshes for the 10th beat of simulation
- iphases.mat: timing information for opening and closing of mitral and aortic valves
- SimulationBiV1Par1_HemodataFullBeatCALBBB.xls: Simulated hemodynamics data for all beats at baseline. The measurements for each column are listed in CircAdapt_Hemodata_Columns_Units.xlsx
- stress10.xls: simulated myocardial stress results for the 10th beat
- stretch10.xls: simulated myocardial stretch results for the 10th beat
Parameter4: Simulation results after CRT (labeled CRT)
- BiV1_CT_unloaded_DT_CA_CRT_FullBeat.cont6: Continuity file of biventricular model after CRT, generated with patient-specific parameters
- FullBeat10.zip: All deformed meshes for the 10th beat of simulation
- iphases.mat: timing information for opening and closing of mitral and aortic valves
- SimulationBiV1Par1_HemodataFullBeatCACRT.xls: Simulated hemodynamics data for all beats after CRT. The measurements for each column are listed in CircAdapt_Hemodata_Columns_Units.xlsx
- stress10.xls: simulated myocardial stress results for the 10th beat
- stretch10.xls: simulated myocardial stretch results for the 10th beat
Simplified Work Estimates Dataset
This directory contains data and functions needed to generate simplified myocardial work measurements from finite element meshes of patient-specific left ventricular (LV) endocardium.
Files:
Data folder:
This folder contains the results of all regional strain, stress, and myocardial work estimates as .mat files. This folder also contains a subfolder called Generic Pressure-Strain Estimate, which stores a digitized generic LV pressure waveform as a .csv file. The variables and units for this work at in Simplified_Work_Estimate_Units.xlsx.
- all_seg_strain_all_pats.mat: segmental strain for all patients. Segments are defined by the 17-segment AHA model. Variables contained in this mat file are seg_strain_all_pats (segmental PSM-based strain measurements) and segRSCT_allpats (segmental simplified strain measurements).
- all_seg_work_all_pats.mat: segmental work measurements for all patients. Variables contained in this mat file are segWork_allpats (segmental PSM-based work), segMWCT_allpats (segmental PLHCSA results).
- Generic Pressure-Strain Estimate
- Generic_LVP.csv: generic pressure waveform digitized from Wigger's diagram (Human Bio Media)
- GenericLVP_work_all_patches.mat: generic pressure-based simplified regional work estimates (PgenSA and Pgen,scaledSA) for each patch of the LV geometry for all patients. Variables included are MWCTgenericLVP (PgenSA results) and MWCTgenericLVP_patES (Pgen,scaledSA results).
- genericLVPs.mat: stores the digitized generic LV pressure waveform (variable genericLVP) and generic pressure waveforms scaled to patient-specific peak pressure for each patient (variable genericLVP_ESP).
- Laplace_measurements.mat: stored the local effective radius (variable r_eff_allpats) and wall thickness measurements (variable thickness_data) for each patient to compute wall stresses.
- Laplace_work_all_patches.mat: wall stress-based simplified regional work estimates (WSEDSA and WSTVSA) for each patch of the LV geometry for all patients. Variables included are MWCTLP (WSEDSA results) and MWCTLPTV (WSTVSA results).
- MWCT_tris.mat: catheterization pressure-based simplified regional work estimate (PLHCSA) for each patch of the LV geometry for all patients (variable MWCT_tris).
- PatchAreas.mat: area of each patch on the LV surface for each patient (variable PatchAreas).
- RS_CT.mat: simplified regional strain estimate (RSCT) for each patch of the LV geometry for all patients (variable RSCT_tris).
- seg_strain_allpats.mat: segmental strain measurements for each patient derived from patient-specific model (variable seg_strain_allpats). Also stored in all_seg_strain_all_pats.mat.
- segMWCT_effrad_allpats.mat: segmental wall-stress-based simplified work measurements for all patients. Variables contained in this mat file are segMWCTLP (segmental WSEDSA), segMWCTLPTV (segmental WSTVSA).
- segMWCT_genericLVP_allpats.mat: segmental generic pressure-based simplified work measurements for all patients. Variables contained in this mat file are segMWCT_genericLVP (segmental PgenSA results), and segMWCT_genericLVP_patES (segmental Pgen,scaledSA results).
- dataPSMLBBBCRT.mat: data from patient-specific modeling.
LV Geometric Models folder:
Contains LV geometric models for all 8 patients. The models are stored in the following structure for a sample case: BiV1-8/BiV1/LVPointsXYZ where XYZ is the relative time of the recording of each model. Each case is then represented as a collection of geometric models demonstrating LV motion over one cardiac cycle. Each time frame is stored in three file formats: .obj, .txt, and .xls.
Scripts folder:
Contains the functions and scripts used to compute and analyze regional strain, stress, and myocardial work. User will need to download two packages from MATLAB to process the data: readObj and ellipsoid_fit. Below, the scripts are organized by use.
Function files:
- calculate_RSCT.m: computes simplified regional strain from LV geometric model
- calculatePatchWorkEstimates.m: computes work estimated for every patch of the geometric model
- calculateSegmentalWorkEstimates.m: computes simplified work for each segment of the LV using the 17 segment AHA model.
- data_analysis.m: performs statistical analysis for comparison of two datasets.
- getNegWorkFractions.m: computes the fraction of the LV areas performing negative work
- getSegmentLabels.m: generates an array indicated which patch on the LV model corresponds to each AHA segment.
- PolyAreaSigned.m: computes area of a polygon.
CRT Analysis Scripts: subfolder containing functions to evaluate utility of simplified work estimated to predict CRT response (figures 5 and 9, table 4 in main manuscript)
- calculateNegWorkFraction.m: computes the fraction of the LV volume performing negative work
- plotCorrelationPatientData.m: generates correlation plots between simplified work estimates and PSM-based work.
- plotHistogram.m: generates histogram of work metrics for CRT responders and non-responders.
- plotResponderBoxplot.m: generates boxplots that compare work metrics from PSM-based work estimates between CRT responders and non-responders
- plotResponderBoxplotAllEsimates.m: generates boxplots that compares the ability for work metrics from PSM-based work and simplified work estimates to stratify CRT responders and non-responders.
- plotResponderBoxplotCOV.m: generates boxplots that compares the ability for COVW computed with PSM-based work and simplified work estimates to stratify CRT responders and non-responders.
Scripts:
- calculateSimpleWork.m: calculates 4 of the 5 simplified work measurements: generic LV pressure-regional strain area (PgenSA), generic LV pressure scaled to patient specific LV pressure-regional strain area (Pgen,scaledSA), end-diastolic wall stress-strain area (WSEDSA) and time-varying wall stress-strain area (WSTVSA). Results are stored as Laplace_work_all_patches.mat and GenericLVP_work_all_patches.mat in the data folder.
- CalculateTriMW.m: Computes work as the left heart catheterization derived LV pressure-strain area (PLHCSA). Results stored as MWCT_tris.mat. Also computes segmental work based on 17 AHA segment model for PLHCSA and PSM-based work. Results for each work metric are stored in the data folder as segMWCT_allpats.mat and all_seg_work_all_pats.mat respectively.
- CalculateTriStrain.m: computes simplified regional strain for all patches on the LV surface and for each AHA segment. Regional strain results for all patches are stored as RS_CT.mat. Segmental strain results are stores as seg_strain_allpat.mat. results are stored in the data folder.
- CRT_responder_analysis2.m: compares ability for metrics of PSM-based work and simplified estimates of work to predict CRT responders.
- evaluateLVPestimates.m: computes generic pressure waveform scaled to patient-specific peak LV pressures (Pgen,scaled). Generic pressure waveform stored as genericLVP.mat and Pgen,scaled results for each patient stored in genericLVP_ESP.mat in data folder.
- getPatchAreas.m: computes the areas of each patch on the surface of the geometric LV models. Results stored as PatchAreas.mat in data folder.
- paper_results_figures.m: computes figures in Supplement 2 (Simplified Work Estimates: Methodology for simplified estimates of myocardial work).
- PrincipalCurvatureAnalysis2.m: generates principal curvatures for each case and stores results along with wall thickness measurements in Laplace_measurements.mat in the data folder. Computes segmental wall stress-strain area work estimates (WSEDSA and WSTVSA) and stores them in seg_MWCT_effrad_allpats.mat in the data folder.
- WorkMetrics.m: computes work metrics from PSM-based work measurements.
- WorkPatAllMetric.m: plots correlations between a particular work metric of interest and a clinical parameter of interest.
Segmental Wall Thickness Measurements folder: Contains .txt files of regional LV wall thickness measurements over time for each case. For example, data for one case is stored as BiV1/ WallThicknessWholeLVTimeVar.txt.
Sharing/Access Information
All necessary data is stored within the repository.
Methods
We built three-dimensional multi-scale patient-specific computational models parameterized by anatomic, functional, hemodynamic, and electrophysiological measurements in eight patients with heart failure and left bundle branch block (LBBB) who received CRT.
To increase clinical translatability, we also explored whether streamlined computational methods provide accurate estimates of regional myocardial work.