Method Comparison of In Vitro Wound Area Measurements

Abstract

Abstract Wound area is a primary outcome measure in wound healing studies. This method comparison study evaluates differences of wound area measurements of a newly developed image analysis method based on wound edge contour to an existing method based on contrast tolerance. Digital images of 64 wounds were taken immediately after wounding matured in vitro 3D organotypic tissues with a biopsy punch. Wound area measurements were calculated using each image analysis method and then normalized. The method comparison study evaluates the difference of each paired measurements for all 64 wound areas. Measurement differences are demonstrated and evaluated in normalized data boxplots, scatter plots with a line of equality, data histogram and Normal probability plots, and a Bland-Altman plot of paired measure difference against mean. The measured wound areas using the tolerance method have large variability in comparison to the contour method measures. The tolerance method measures often underestimate and overestimate what is assumed to be an approximately repeatable initial wound size. Skewness in comparison plots are due to the ‘fat tails’ introduced by the variability of measurements of the tolerance method. In contrast, the contour method results in larger wound area measurements on average at a statistically significant level of difference. The relatively less variable range of contour method measurements suggest this method has more potential to agree with the ‘true’ wound area. Future work to improve the method are proposed for application of image analysis methods to distinguish true wound area and measurement error in time for wound healing treatment-control experiments.