For patients with major burn injuries, an accurate estimation of percent total body surface area (%TBSA) is essential to ascertain vital and functional prognosis and to plan appropriate treatment and minimize medical and surgical complications. However, current clinical methods for burn size estimation lack accuracy and reliability to assess patients with different morphologies. 2D/3D modeling approaches for burn injuries approximate %TBSA assessment and are inaccurate. Only parametric modeling satisfies clinical requirements specifically to major burn patients (limited time, unconscious patient, restricted environment). In order to meet the need for 3D personalization of the patient’s morphologies, the MakeHuman (MH) open-source software is suggested. Thus, this thesis proposes a 3D-based approach with personalized 3D models from a limited set of anthropometric measurements to accurately assess the %TSBA.
First, with predictive performances above 98%, several multiple regression analyses demonstrated that a limited number of anthropometric measurements are sufficient to obtain an accurate 3D modeling of the body surface area (BSA).
Secondly, the relevance of personalizing 3D model to the patient’s morphology for a better assessment of %TBSA was evaluated. Using a 3D rendering interface developed for representation and calculation of the %TBSA, six specific burn patterns were applied automatically on a set of 15 paired 3D models with the same BSA but different morphologies. Small local variations in the morphology of 0.04 m2 ± 0.01 (characterized by burn TBSA) induce overestimation in %TBSA. 8 experts and 6 non-experts estimated the %TBSA of burn patterns simulated on the four commercial mannequins with the clinical method and then in 3D on the interface. The clinical method obtained an average error of 34 % ± 11 whereas the proposed approach is more accurate, repeatable and reproducible with an average error of 5 % ± 1.
Finally, from a stratified sample of 50 000 morphologies randomly generated, several machine learning algorithms were used to identify sets of 3 to 4 anthropometric measurements the most relevant for an accurate 3D modeling of the morphology. The accuracy of the 3D modeling of the morphology was validated on 8 healthy subjects. Results highlighted the height and circumferences of the bust, neck, hips and waist as the best predictors for the body surface area. 3D models built from three to four anthropometric measurements showed good accuracy and were geometrically close to gold standard 3D scans with an average distance error of 1.1 cm ± 0.4 and an average maximal distance error of 7.4cm ± 1.1. Outcomes of this study aim to decrease medical and surgical complications by decreasing errors in %TBSA, and therefore improving patient outcomes by personalizing care.
| Date | 15 May 2019 |
|---|
| Original language | French |
|---|
| Awarding Institution | - École de technologie supérieure
|
|---|
| Supervisor | Jacques A. de Guise (Supervisor) & Isabelle Perreault (Co-supervisor) |
|---|
Desbois, A. (Author),
De Guise, J. (Supervisor) & Perreault, I. (Co-supervisor),
15 May 2019Student thesis: Doctoral thesis › Doctorate in Engineering: Engineering