Fig. 6: Correlation of tissue prediction by DeepNeo with histopathological findings in rabbits. | Communications Medicine

Fig. 6: Correlation of tissue prediction by DeepNeo with histopathological findings in rabbits.

From: Deep learning model DeepNeo predicts neointimal tissue characterization using optical coherence tomography

Fig. 6

a Confusion matrix of DeepNeo with histopathological based labels of n = 60 quadrants stemming from 12 rabbits. b Representative examples from a rabbit model of neoatherosclerosis with hematoxylin–eosin staining, revealing underlying neointimal tissue characteristics with areas of interest marked with *, respectively +. DeepNeo-based analysis of co-registered OCT frames showed overall good agreement between histopathological findings and AI-based tissue prediction, classifying the tissue into not analyzable (blue), homogenous (green), heterogeneous (orange), and neoatherosclerosis (red). Line thickness represents DeepNeo certainty, where high certainty corresponds to a thicker line.

Back to article page