Segmentation of neck lymph nodes in CT datasets with stable 3D mass-spring models segmentation of neck lymph nodes

Abstract

Rationale and objectives: The quantitative assessment of neck lymph nodes in the context of malignant tumors requires an efficient segmentation technique for lymph nodes in tomographic three-dimensional (3D) datasets. We present a stable 3D mass-spring model for lymph node segmentation in computed tomography (CT) datasets.

Materials and methods: For the first time our model concurrently represents the characteristic gray value range, directed contour information, and shape knowledge, which leads to a robust and efficient segmentation process.

Results: Our model design and the segmentation accuracy were both evaluated with 40 lymph nodes from five clinical CT datasets containing malignant tumors of the neck.

Conclusion: The segmentation accuracy proved to be comparable to that of manual segmentations by experienced users and significantly reduced the time and interaction needed for the lymph node segmentation.