Automatic segmentation of the spine by means of a probabilistic atlas with a special focus on ribs suppression

Abstract

The study demonstrated that the atlas is able to detect and appropriately eliminate the ribs while improving the segmentation accuracy.

Clinical utility of convolutional neural networks for treatment planning in radiotherapy for spinal metastases

“…Previous studies on automatic spine delineation on CT images [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] reported DSC values in the range of 89–96% and HD in the range 5.8–15 mm, although HD was not always reported. Our sequential approach resulted in DSC values of 97% and 95%, and HD values of 3.6 and 4.5 mm for the internal and external validation respectively.…”

“…The strength of this study lies in the focus on application of automatic spine delineation for radiotherapy treatment planning. Previous studies were conducted mainly on small sets of healthy subjects [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] . However, clinical application of automatic delineation would require the technique to be accurate also for patients with pathological vertebrae, such as spinal metastases and fractures.…”

“…Automatic spine delineation can be divided into two separate tasks: segmentation (correctly distinguishing vertebrae from background) and labeling (correctly identifying each vertebral level). Various approaches for automatic spine delineation on CT have been proposed, yielding promising results [5] , [6] , [7] , [8] , [9] , [10] , [11] , [12] , [13] .…”

Automatic segmentation of vertebral contours from CT images using fuzzy corners

“…Certain morphological operations are allowed to enhance the process. The use of atlas for aiding segmentation is an alternate method [14][15][16]. A three step algorithm is developed in [17] adopting Statistical Shape Models (SSM), that includes a seed placement, initial labeling and final optimization.…”

Learning a Statistical Full Spine Model from Partial Observations

“…Other probabilistic models, different from the shape surface models, such as probabilistic atlas [20], graph models [23], hidden markov models [9] and hier-archical models [22][26] [7] have also been proposed. For example Ruiz et al [20] proposed a probabilistic atlas of the spine. By co-registering 21 CT scans, a probability map is created which can be used to segment and detect the vertebrae with a special focus on ribs suppression.…”