Combining registration and active shape models for the automatic segmentation of the lymph node regions in head and neck CT images

Abstract

Purpose: Intensity-modulated radiation therapy ͑IMRT͒ is the state of the art technique for head and neck cancer treatment. It requires precise delineation of the target to be treated and structures to be spared, which is currently done manually. The process is a time-consuming task of which the delineation of lymph node regions is often the longest step. Atlas-based delineation has been proposed as an alternative, but, in the authors' experience, this approach is not accurate enough for routine clinical use. Here, the authors improve atlas-based segmentation results obtained for level II-IV lymph node regions using an active shape model ͑ASM͒ approach. Methods: An average image volume was first created from a set of head and neck patient images with minimally enlarged nodes. The average image volume was then registered using affine, global, and local nonrigid transformations to the other volumes to establish a correspondence between surface points in the atlas and surface points in each of the other volumes. Once the correspondence was established, the ASMs were created for each node level. The models were then used to first constrain the results obtained with an atlas-based approach and then to iteratively refine the solution. Results: The method was evaluated through a leave-one-out experiment. The ASM-and atlas-based segmentations were compared to manual delineations via the Dice similarity coefficient ͑DSC͒ for volume overlap and the Euclidean distance between manual and automatic 3D surfaces. The mean DSC value obtained with the ASM-based approach is 10.7% higher than with the atlas-based approach; the mean and median surface errors were decreased by 13.6% and 12.0%, respectively. Conclusions: The ASM approach is effective in reducing segmentation errors in areas of low CT contrast where purely atlas-based methods are challenged. Statistical analysis shows that the improvements brought by this approach are significant.

Atlas-based automatic segmentation of head and neck organs at risk and nodal target volumes: a clinical validation

“…This may potentially lead to violations of the cranio-caudal limits on the one hand [17] and to non compensation of large volumes overestimations on the other hand [18]. “Active Contour” (AC) or “Active Shape Modelling” post-processing after averaged atlas deformation constrains volumes within their anatomic boundaries, potentially compensating for these problems (AVG-AC) [12,13,21]. Our method takes the advantages of the AC methodology applied to the fast and simple IND atlas.…”

“…Different deformation registration strategies were developed, based on either individual patient data, averaged patient generation, multiple patient data [11] or, more recently, introduction of a volume post-processing by recognition of the key anatomical structures of the head and neck area [12,13]. …”

Auto‐segmentation of normal and target structures in head and neck CT images: A feature‐driven model‐based approach

“…To overcome this problem, various approaches based on optimal atlas selection, and multiatlas segmentation and fusion have been proposed. [9][10][11][12][13] Hybrid approaches combine registration and segmentation into a common framework, [14][15][16] where, for example, evolution of deformable models can serve as a registration constraint or used to compensate for the residual differences after the registration step.…”

Automatic classification of squamosal abnormality in micro-CT images for the evaluation of rabbit fetal skull defects using active shape models

“…As proposed in [9], when applying the ASM on new images, following the same procedure shown in Figure 4, they are globally aligned through rigid registrations with the template image, and then locally aligned through both rigid and For each new shape, we perform a least square fitting by the ASM as a linear combination of the mean shape and the modes of normal shape variations. The fitted shape x v is represented as…”

“…We followed the approach introduced in [9] for constructing the active shape model. We selected 15 micro-CT images classified as normal by the DART evaluators as the training set for the ASM.…”