Classification of chronic venous diseases based on skin temperature patterns

Stephan Dahlmanns*, Stefanie Reich-Schupke, Franziska Schollemann, Markus Stücker, Steffen Leonhardt, Daniel Teichmann

*Corresponding author for this work
1 Citation (Scopus)

Abstract

Objective. Infrared thermography has the potential to complement the classification of chronic venous diseases (CVD), but lacks sophisticated insights on the association between recorded skin temperatures and the severity of CVD. This research aims to identify temperature patterns in the lower legs of patients that are distinct in specific forms of CVD, including florid ulcers. Approach. Infrared images were acquired in a clinical trial with 36 patients and segmented using a region selection algorithm. The regions were analyzed with respect to seven predefined features. The most prominent thermal features were translated into rules to classify CVD. Main results. Patients with mild forms of CVD show local increases in skin temperature by more than 1.5 °C. These regions were 2.0 °C warmer when CVD is more severe. Temperature variations of on average 0.4 °C occurred within venous leg ulcers. Furthermore, these wounds were 1.1 °C–6.3 °C colder than periwound skin. Significance. Temperature patterns characterized by differences in temperature that occur within a few centimeters or millimeters are distinct to specific stages of CVD. These patterns are present in the locations of varicose veins and tissue damages. Significance. The findings increase the body of knowledge on the potential for the early detection of CVD using infrared thermography. Applying the presented algorithms and rules, infrared thermography may become a complementary tool for the objective classification of CVD.

Original languageEnglish
Article number045001
JournalPhysiological Measurement
Volume42
Issue number4
ISSN0967-3334
DOIs
Publication statusPublished - 04.2021

Fingerprint

Dive into the research topics of 'Classification of chronic venous diseases based on skin temperature patterns'. Together they form a unique fingerprint.

Cite this