Validation of automatic landmark identification for atlas-based segmentation for radiation treatment planning of the head-and-neck region Claudia Leavens* a , Torbjørn Vik b , Heinrich Schulz b , Stéphane Allaire a , John Kim a , Laura Dawson a , Brian O’Sullivan a , Stephen Breen a , David Jaffray a , and Vladimir Pekar c a Princess Margaret Hospital, University of Toronto, Canada; b Philips Research Europe, Hamburg, Germany; c Philips Research North America, Markham, Canada ABSTRACT Manual contouring of target volumes and organs at risk in radiation therapy is extremely time-consuming, in particular for treating the head-and-neck area, where a single patient treatment plan can take several hours to contour. As radiation treatment delivery moves towards adaptive treatment, the need for more efficient segmentation techniques will increase. We are developing a method for automatic model-based segmentation of the head and neck. This process can be broken down into three main steps: i) automatic landmark identification in the image dataset of interest, ii) automatic landmark- based initialization of deformable surface models to the patient image dataset, and iii) adaptation of the deformable models to the patient-specific anatomical boundaries of interest. In this paper, we focus on the validation of the first step of this method, quantifying the results of our automatic landmark identification method. We use an image atlas formed by applying thin-plate spline (TPS) interpolation to ten atlas datasets, using 27 manually identified landmarks in each atlas/training dataset. The principal variation modes returned by principal component analysis (PCA) of the landmark positions were used by an automatic registration algorithm, which sought the corresponding landmarks in the clinical dataset of interest using a controlled random search algorithm. Applying a run time of 60 seconds to the random search, a root mean square (rms) distance to the ground-truth landmark position of 9.5 ± 0.6 mm was calculated for the identified landmarks. Automatic segmentation of the brain, mandible and brain stem, using the detected landmarks, is demonstrated. Keywords: automatic segmentation, landmark identification, non-rigid registration 1. INTRODUCTION Manual contouring for radiation treatment planning is typically performed on two-dimensional axial computed tomography (CT) image slices using simple drawing tools. Manual contouring of the head and neck is particularly time- consuming and can take several hours, due to the complexities of the anatomy of the head and neck, complex cancer presentations and their corresponding radiotherapy target volumes, and the large number of organs at risk. With the aim of reducing the time required to contour features of interest in the planning CT scans, we are developing a technique for automatic volume delineation of the head and neck, using model-based segmentation. Segmenting structures from medical images and reconstructing a compact geometric representation of these structures is difficult due to the size of the datasets and the complexity and variability of the anatomic shapes of interest. Shortcomings of the datasets such as image artifacts and noise may cause the boundaries of structures to be indistinct and disconnected. The challenge is to extract boundary elements belonging to the same anatomical structure, and integrating these elements into a coherent and consistent model of the structure. *claudia.leavens@rmp.uhn.on.ca; phone +1 416 603 5800 x6364; fax +1 416 603 5155 Medical Imaging 2008: Image Processing, edited by Joseph M. Reinhardt, Josien P. W. Pluim, Proc. of SPIE Vol. 6914, 69143G, (2008) 1605-7422/08/$18 · doi: 10.1117/12.769710 Proc. of SPIE Vol. 6914 69143G-1 2008 SPIE Digital Library -- Subscriber Archive Copy