Skip to main content
Insights into Imaging
Fig. 3 | Insights into Imaging

Fig. 3

From: Standardised lesion segmentation for imaging biomarker quantitation: a consensus recommendation from ESR and EORTC

Fig. 3

Example of good versus poor reproducibility between observers based on lesion definition: Fat-suppressed coronal T2W images through the orbits in a patient with orbital lymphoma (A, B) and in a patient with right idiopathic orbital inflammation (C, D). The reproducibility between the 2 observer segmentations in (A, B) is excellent (Dice similarity coefficient 0.97), whereas it is poor in the observer segmentations in (C, D) where the lesion is poorly defined (Dice similarity coefficient 0.5)

Back to article page