Artificial Intelligence–Based Indocyanine Green Lymphography Pattern Classification for Management of Lymphatic Disease

Abstract

Background: Lymphedema diagnosis relies on effective imaging of the lymphatic system. Indocyanine green (ICG) lymphography has become an essential diagnostic tool, but globally accepted protocols and objective analysis methods are lacking. In this study, we aimed to investigate artificial intelligence (AI), specifically convolutional neural networks, to categorize ICG lymphography images patterns into linear, reticular, splash, stardust, and diffuse. Methods: A dataset composed of 68 ICG lymphography images was compiled and labeled according to five recognized pattern types: linear, reticular, splash, stardust, and diffuse. A convolutional neural network model, using MobileNetV2 and TensorFlow, was developed and coded in Python for pattern classification. Results: The AI model achieved 97.78% accuracy and 0.0678 loss in categorizing images into five ICG lymphography patterns, demonstrating high potential for enhancing ICG lymphography interpretation. The high level of accuracy with a low loss achieved by our model demonstrates its effectiveness in pattern recognition with a high degree of precision. Conclusions: This study demonstrates that AI models can accurately classify ICG lymphography patterns. AI can assist in standardizing and automating the interpretation of ICG lymphographic imaging.