Fig. 5: Examples of automated analysis by DeepNeo. | Communications Medicine

Fig. 5: Examples of automated analysis by DeepNeo.

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

Fig. 5

Accurate segmentation and prediction of neointimal tissue characteristics on quadrant level. a Predominant homogenous neointima with foam cells in Q3. b Heterogenous neointima in Q1 and Q4 with foam cell infiltration in Q2 and Q3. c No neointima present. d Mixture of homogeneous and heterogenous neointima as well as possible neoatherosclerosis. Note the low confidence in b (Q1 and Q3) and d (Q3), reflecting the difficulty in differentiating heterogeneous neointima from neoatherosclerosis in some cases. Lower row: automated segmentation of lumen, neointima and stent struts.

Back to article page