Dataset and Sample Code for: Quantifying Myocardial Oxygen Consumption and Efficiency with Motion-resolved Cardiac MRI Authors: Li-Ting Huang¹²†, Chia-Chi Yang¹†, Guan Wang³†, Henghui Zhang³, Ranran Zhang³, Hao Ho⁴, Archana Malagi¹, Yuheng Huang⁵⁶, Xinqi Li¹, Ghazal Yoosefian¹, Xinheng Zhang¹, Ziyang Long¹, Xiaoming Bi⁷, Janet Wei⁸, Alan C. Kwan⁹, Michael D. Nelson¹⁰, C. Noel Bairey Merz⁸, Daniel Berman¹¹, Anthony Christodoulou¹² ¹³, Debiao Li¹, Rohan Dharmakumar⁵⁶, and Hsin-Jung Yang¹* Affiliations: ¹ Biomedical Imaging Research Institute, Cedars-Sinai Medical Center, Los Angeles, CA, USA ² Department of Medical Imaging, National Cheng Kung University Hospital, Tainan, Taiwan ³ Department of Radiology, the First Hospital of China Medical University, Shenyang, Liaoning, China ⁴ Department of Statistics, University of California at Los Angeles, Los Angeles, CA, USA ⁵ Krannert Cardiovascular Research Center, Indiana University, Bloomington, IN, USA ⁶ Department of Bioengineering, University of California at Los Angeles, Los Angeles, CA, USA ⁷ Siemens Medical Solutions USA, Inc., Los Angeles, CA, USA ⁸ Division of Cardiology, Women's Heart Center, Cedars-Sinai Medical Center, Los Angeles, CA, USA ⁹ Department of Cardiology, Smidt Heart Institute, Cedars-Sinai Medical Center, Los Angeles, CA, USA ¹⁰ Department of Kinesiology, University of Texas at Arlington, Arlington, TX, USA ¹¹ Departments of Medicine, Biomedical Sciences and Imaging, Cedars-Sinai Medical Center, Los Angeles, CA, USA ¹² Department of Radiological Sciences, David Geffen School of Medicine at UCLA, Los Angeles, CA, USA ¹³ Department of Bioengineering, University of California, Los Angeles, CA, USA † These authors contributed equally to this work. Corresponding author: Hsin-Jung Yang (hsin-jung.yang@cshs.org) 📄 Overview This repository contains the replication dataset and sample code for the paper "Quantifying Myocardial Oxygen Consumption and Efficiency with Motion-resolved Cardiac MRI." Project Description We introduce a rapid, self-calibrated, and motion-resolved cardiac MRI framework to quantify whole-heart Myocardial Oxygen Consumption in under three minutes using standard clinical systems. Validated in a porcine model and demonstrated in heart failure patients, this needle-free approach enables the noninvasive characterization of myocardial oxygen metabolism—including myocardial oxygen extraction fraction (mOEF) and efficiency—to facilitate early disease detection and personalized therapeutic strategies. Repository Scope The included material supports the replication of the study's key technical components: Numerical Simulations: Sensitivity and accuracy analysis of the proposed MR oximetry sequence. Parametric Mapping: Reconstruction of blood oxygen saturation maps from free-breathing cardiac MRI acquisitions using a Low-Rank Tensor (LRT) framework. 📂 Repository Structure The repository is organized into three main components: simulation scripts, reconstruction tools, and sample data. . ├── README.md ├── LICENSE ├── simulation/ ├── Signal_sensitivity.p # MATLAB p-code: BOLD signal sensitivity analysis └── Noise_simulation.p # MATLAB p-code: Monte Carlo accuracy analysis ├── reconstruction/ ├── multitasking_recon.p # MATLAB p-code: Main reconstruction executable └── sample_scan.dat # Sample Siemens TWIX data (Raw MRI data) 💻 1. Numerical Simulations The simulation folder contains obfuscated MATLAB scripts (.p files) to evaluate the performance of the proposed MR oximetry sequence. Signal Sensitivity Analysis Script: simulation/Signal_sensitivity.p Description: This script evaluates the BOLD sensitivity of the sequence by plotting BOLD signal changes across a range of physiological parameters (blood oxygenation , Hct levels) and sequence inter-echo spacing. Key Output: 3D sensitivity plots and cross-section contours used to identify optimal inter-echo spacing values (e.g., 7.5ms to 30ms) that ensure sufficient signal variation across the target range. Noise Simulation and Accuracy Script: simulation/Noise_simulation.p Description: This script assesses the robustness of the method against noise. It simulates blood pool circles with varying Signal-to-Noise Ratios (SNR = 7–100) compared to in vivo levels (SNR ≈ 15). Methodology: A Monte Carlo simulation with 1000 repetitions is performed to fit BOLD curves and derive fitting errors in blood oxygenation. Key Output: Statistical plots showing the mean error and standard deviation of the derived blood oxygenation under different noise conditions. 🧠 2. Parametric Map Generation The reconstruction folder contains the tools required to generate oxygenation maps from raw MRI data. Reconstruction Framework Script: reconstruction/multitasking_recon.p Methodology: The code implements a Low-Rank Tensor (LRT) framework to handle the high-dimensionality of the free-running, motion-resolved 3D acquisition. Modeling: Reconstructed tensor images are fitted to a T1-T2 blood oximetry model (incorporating the Luz-Meiboom T2 model and a Two-compartment T1 model) to quantify Blood Oxygen Saturation in the coronary sinus. Sample Data File: reconstruction /sample_scan.dat Format: Siemens TWIX raw data format. Description: An anonymized sample dataset acquired from a healthy swine using a Siemens 3T scanner at Cedars-Sinai Medical Center. This file allows users to test the multitasking_recon pipeline and verify the generation of parametric maps. Usage The user should input a Siemens MRI raw data file (.dat / twix) to generate the maps. 📬 Contact For questions regarding data processing or usage, please contact the dataset author.