gms | German Medical Science

Introduction: In burn medicine, the burn injury’s extent is a central parameter for volume replacement therapy, therapy planning, and quality assurance. Errors in the determination of burn surfaces can lead to severe complications such as burn edema, ARDS, wound healing disorders, and abdominal compartment syndrome. The research group has been developing objective 3D methods since 2001 to optimize the treatment process. The presentation gives an insight into a completely new technological approach, which accounts for the recent shift towards web based technology.

Material and Methods: The long-standing research cooperation project BurnCase 3D, which has been developed with support of numerous medical partners, serves as foundation for implementing a new software framework for application-oriented research in burn documentation and treatment. This framework is developed within the ongoing international follow-up research project “SenseBurn” (EUREKA-2 eurostars).

Results:

1.

Platform Independence: The development resulted in a software that runs on Android, iOS, Windows, Linux, macOS, and directly in a web browser – without installation.

2.

Patient-specific 3D models: The creation and adaptation of patient-specific 3D models for different body types (shapes) were implemented using hand crafted expansion vectors deforming a 3D model from a unified model collection. This unified 3D models enable continuous documentation, even if the body proportions changes require a base model switch. Personalized models allow for a significant increase in documentation quality.

3.

Pose adaptation: To facilitate efficient use in clinical routine, an automated pose adaptation of the model was implemented. For this purpose, the location and position of the joints are extracted from a single RGB image using machine learning algorithms. Due to the exact matching of pose and shape of the 3D model, the transfer of the wound surfaces is possible with minimal effort and high accuracy.

4.

Exact transmission of wound areas without artifacts: A new method for wound annotation that is independent of the 3D model mesh resolution was developed. Now the documentation of tiny regions or slight changes over time is possible for the first time.

5.

New Opportunities: The use of new algorithms (e.g., machine learning) opens the possibility of using future technologies.

Conclusion: The latest development of 3D diagnostics results in simplified and optimized options for use in burns treatment. In comparison to previous methods, the estimation error can be reduced further. The resulting high-quality database provides a way to create new and better evidence in the future.