Abstract
Purpose
Exact and reproducible knowledge regarding the position, size, and type of the lymph nodes is often needed for tumor computer-aided diagnosis, treatment planning, and follow-up. An automatic segmentation method for CT data was developed that can identify and delineate normal as well as pathologically altered lymph nodes to satisfy this requirement.
Methods
A semi-automatic lymph node segmentation method was developed using a 3D Stable Mass-Spring Model (SMSM), based on parallel simulation of the shape model on CT scan images. The models are started across the whole dataset at all potential lymph node positions but will only adapt to the data where a lymph node is found. The node positions can be determined by an evaluation of the model’s quality of fit.
Results
Systematically chosen lymph nodes in 5 CT datasets, including enlarged, necrotic, fuzzy-bounded, and deformed lymph nodes, were used to evaluate the segmentation algorithm performance. A test set of 29 lymph nodes taken from 4 typical lymph node regions were included. All lymph nodes were detected automatically, while an additional 31% false-positive (n = 9) candidates were detected. The average calculation time was 2 min per dataset. The segmentation accuracy was comparable to the inter-observer variance of human experts.
Conclusions
Clinically relevant lymph nodes were detected within a few minutes and provided sufficient accuracy to demonstrate the feasibility of a new segmentation method. The test data were diverse, and the robust results suggest potential applicability to many kinds of lymph node abnormalities, except for extremely degenerated lymph nodes.
Similar content being viewed by others
References
Sobin LH, Wittekind C (2002) TNM: classification of malignant tumours. Wiley-Liss
Hintze J, Cordes J, Preim B (2005) Bildanalyse für die präoperative Planung von Neck Dissections. In: BVM
Cordes J, Dornheim J, Preim B, Hertel I, Strauss G (2006) Pre-operative segmentation of neck ct datasets for the planning of neck dissections. In: SPIE: Medical Imaging
Schreyer A et al (2005) Vergleich der Genauigkeit und Effizienz koronarer und axialer Rekonstruktionen zur Lymphknotendetektion bei einem 16-Zeilen-Spiral-CT am Beispiel von Halslymphknoten. RöFo: Fortschritte auf dem Gebiete der Röntgenstrahlen und der Nuklearmedizin 177: 1430–1435
Dornheim L, Tönnies KD, Dixon K (2005) Automatic segmentation of the left ventricle in 3D SPECT data by registration with a dynamic anatomic model. In: MICCAI
Dornheim L, Tönnies KD, Dornheim J (2005) Stable dynamic 3D shape models. In: ICIP
Rogowska J et al (1996) Evaluation of selected two-dimensional segmentation techniques for computed tomography quantification of lymph nodes. Invest Radiol 13: 138–145
Yan J et al (2004) Lymph node segmentation from CT images using fast marching method. Comput Med Imaging Graph 28: 33–38
Honea D, Snyder WE (1999) Three-dimensional active surface approach to lymph node segmentation. In: SPIE: Medical Imaging 3361: 1003–1011
Dornheim J et al (2007) Segmentation of neck lymph nodes in ct datasets with stable 3d mass-spring models. Academic Radiology (Elsevier) MICCAI 2006 Special Issue
Maleike D et al (2008) Lymph node segmentation on ct images by a shape model guided deformable surface method. In: SPIE: Medical Imaging
Felzenszwalb PF, Huttenlocher DP (2005) Pictorial structures for object recognition. Int J Comput Vis 61: 55–79
Carbonetto P, de Freitas N, Barnard K (2004) A statistical model for general contextual object recognition. In: ECCV
Ferrari V, Tuytelaars T, Gool LV (2004) Simultaneous object recognition and segmentation by image exploration. In: ECCV
Belongie S, Malik J, Puzicha J (2002) Shape matching and object recognition using shape contexts. IEEE Trans Pattern Anal Mach Intell 24: 509–522
Coughlan J, Yuille A, English C, Snow D (2000) Efficient deformable template detection and localization without user initialization. Comput Vis Image Underst 78: 303–319
Jain AK, Zhong Y, Lakshmanan S (1996) Object matching using deformable templates. IEEE Trans Pattern Anal Mach Intell 18: 267–278
Keren D, Subrahmonia J, Cooper DB (1992) Robust object recognition based on implicit algebraic curves and surfaces. In: ICPR 791–794
Ballard DH (1981) Generalizing the hough transform to detect arbitrary shapes. Pattern Recognit 13: 111–122
Kitasaka T, Tsujimura Y, Nakamura Y, Mori K, Suenaga Y, Ito M, Nawano S (2007) Automated extraction of lymph nodes from 3-d abdominal ct images using 3-d minimum directional difference filter. In: MICCAI
Cootes TF, Taylor CJ, Cooper DH, Graham J (1995) Active shape models—their training and application. CVIU 61: 38–59
Delingette H (1994) Simplex meshes: a general representation for 3D shape reconstruction. In: CVPR 856–857
Bardinet E, Cohen LD, Ayache N (1998) A parametric deformable model to fit unstructured 3D data. CVIU 71: 39–54
Dornheim L, Dornheim J, Tönnies KD (2006) Automatic generation of dynamic 3d models for medical segmentation tasks. In: SPIE: Medical Imaging
Dornheim L (2008) Quality of fit of stable mass-spring models. Pattern Recognit Image Anal 18: 359–364
Bergner S, Al-Zubi S, Tönnies K (2004) Deformable structural models. In: ICIP
Cohen LD (1991) On active contour models and ballons. CVGIP: Image Underst 53: 211–218
Dornheim L, Dornheim J, Seim H, Tönnies KD (2006) Aktive Sensoren: Kontextbasierte Filterung von Merkmalen zur modellbasierten Segmentierung. In: Bildverarbeitung für die Medizin
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Dornheim, L., Dornheim, J. & Rössling, I. Complete fully automatic model-based segmentation of normal and pathological lymph nodes in CT data. Int J CARS 5, 565–581 (2010). https://doi.org/10.1007/s11548-010-0530-8
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11548-010-0530-8