GRANU-LATENT study: metrological validation of granulation tissue measurement in chronic wounds

The first metrological study dedicated to measuring the granulation tissue proportion in chronic wounds, led by Dr. Guillaume Maxant. The study draws on 100 wound images assessed by 6 independent expert clinicians and 9 complementary measurement methods, aiming to establish calibrated correction factors applicable to AI algorithms and clinical research.

Study framework

Metrological study led by Dr. Guillaume Maxant, Department of Vascular Surgery, Haguenau Hospital. The first international study dedicated to the rigorous quantification of granulation tissue in chronic wounds.

Timeline

• Sample: 100 chronic wound images
• Raters: 6 independent expert clinicians
• Methods compared: 9 measurement methods (visual assessment, manual segmentation, automated algorithms, colorimetric analysis)
• Status: study ongoing

Study design

Comparative metrological study following a reference-free approach. This original methodology is required by the current absence of a reference method for granulation tissue measurement, in contrast to surface area measurement where digitized planimetry serves as the gold standard.

Objectives

The study pursues two foundational objectives for chronic wound healing assessment:

• Establishing calibrated correction factors applicable to AI-based tissue classification algorithms
• Providing a rigorous reference framework for image annotation in clinical research

Why this study

Assessment of wound tissue composition (granulation, fibrin, necrosis proportions) is a major clinical indicator of wound healing progress. Yet unlike surface area measurement, no international reference method currently exists.

This gap limits the comparability of clinical studies, the quality of machine learning training datasets, and the standardization of clinical practice. GRANU-LATENT addresses this fundamental need by proposing a rigorous metrological approach based on the convergence of multiple independent methods and combined expert assessment.

The findings will contribute to the international standardization of wound assessment, beyond the Pixacare ecosystem alone. The study is part of the logical continuation of the Pixaire-1 and PIXAIRE-Redux research program on AI-based wound measurement validation.

Keywords

Granulation tissue, chronic wound, artificial intelligence, metrology, wound healing, standardization, method validation, digital medical device