Congress:
ECR24
Poster Number:
C-23271
Type:
EPOS Radiologist (scientific)
Authorblock:
A-I. Iuga1, H. Carolus2, M. Rinneburger1, M. Weisthoff1, T. Klinder2, D. Maintz1, T. Persigehl1; 1Cologne/DE, 2Hamburg/DE
Disclosures:
Andra-Iza Iuga:
Speaker: Philips
Heike Carolus:
Employee: Philips GmbH Innovative Technologies, Hamburg, Germany
Miriam Rinneburger:
Nothing to disclose
Mathilda Weisthoff:
Nothing to disclose
Tobias Klinder:
Employee: Philips GmbH Innovative Technologies, Hamburg, Germany
David Maintz:
Speaker: Philips
Thorsten Persigehl:
Nothing to disclose
Keywords:
Abdomen, CT, Segmentation, Neoplasia
Accurate staging of disease, particularly in determining nodal metastatic disease, is essential in oncologic imaging, as it directly influences patient management and prognosis. 1 In solid tumor nodal disease staging, lymph node (LN) short-axis diameters (SAD) are commonly measured unidimensionally during staging examinations, following standardized diagnostic criteria like RECIST.2 Previous work has shown the potential of algorithms in the detection and segmentation of thoracal LNs.3-5 The aim of our study was to develop an automatic 3D detection and segmentation tool for lymph nodes in CT scans of the abdomen supported by a fully convolutional neural network (CNN).