Skip to main content
Fig. 4 | European Radiology Experimental

Fig. 4

From: MRI-based radiomics to predict response in locally advanced rectal cancer: comparison of manual and automatic segmentation on external validation in a multicentre study

Fig. 4

Examples of three patients segmented by internally developed fully convolutional network model (blue line) and the radiologists (red line). Pt. 2003 with DSC of 0.18 (a) and Pt. 2015 with DSC of 0.76 (b) were wrongly classified by both manual and automatic systems, while Pt. 2028 with DSC of 0.81 (c) was wrongly classified only by the automatic model. DSC Dice similarity coefficient, Pt Patient, Sl. Slice

Back to article page