Post-Processing and Workflow

1356055 - A Model for Cardiac Fusion Imaging Software Development for Congenital and Structural Heart Disease

Friday, January 27, 2023

12:20 PM – 1:20 PM

Location: Pacific Jewel

Presenting Author(s)

RL

Ray Lorenzoni, III, MD

Fellow
Children's Hospital Los Angeles
Woodbridge, Connecticut, United States

Primary Author(s)

RL

Ray Lorenzoni, III, MD

Fellow
Children's Hospital Los Angeles
Woodbridge, Connecticut, United States

Co-Author(s)

KM

Kyle McClary, PhD

Software Developer
Microscape, Inc., United States
JF

John Paul Francis

Software Developer
Microscape, Inc., United States
JD

Jon A. Detterich, MD

Associate Professor
Children's Hospital Los Angeles, United States
PW

Pierre Wong, MD

Professor
Children's Hospital Los Angeles, United States

Background: Mainstream three-dimensional (3D) and “4D” (3D movement) visualization tools do not allow fusion of multiple raw 4D imaging datasets. With the availability of powerful 3D software and graphics processors, practicing clinicians may initiate visualization research with a software partner. Window into the Heart (WITH) is a project to create 4D cardiac fusion models from routine clinical imaging.

Methods: A 6-month feasibility phase occurred at Children’s Hospital Los Angeles with a scientific visualization partner, Microscape. Multiple paired clinical 4D pediatric congenital heart disease datasets (3D echocardiography and cross-sectional imaging collected within the same month) were anonymized for software feasibility testing. This feasibility phase sought to establish an interdisciplinary working relationship, customize widely supported Unity software to fuse multiple 4D modalities, and identify relevant clinical applications.

Results: The project established intellectual property and HIPAA agreements, software milestones, and prototype functionality with an intramural grant. Initial in-person meetings allowed the software team to understand physician workflow. Next, the clinical (RL) and software (KM & JPF) leads met virtually twice weekly for one half-hour, first for idea generation or feedback and again for subsequent progress (Figure 1). This structure resulted in concrete headway on milestones, nimble troubleshooting, and timely feedback for iteration. Within 6 months, the proof-of-concept software could render, edit, and fuse multiple 4D cardiac imaging datasets without manual segmentation (Figure 2). This advance in routine patient-specific 4D modeling is amenable to future testing for surgical planning, care transitions, and education.

Conclusion: The WITH feasibility phase demonstrated an effective working relationship between clinical and software teams resulting in a functional fusion imaging software tool. This strategy serves as a template for practicing clinicians to independently develop software for cardiac imaging visualization research.