A comparative study of automatic image segmentation algorithms for target tracking in MR‐IGRT

Abstract

On‐board magnetic resonance (MR) image guidance during radiation therapy offers the potential for more accurate treatment delivery. To utilize the real‐time image information, a crucial prerequisite is the ability to successfully segment and track regions of interest (ROI). The purpose of this work is to evaluate the performance of different segmentation algorithms using motion images (4 frames per second) acquired using a MR image‐guided radiotherapy (MR‐IGRT) system. Manual contours of the kidney, bladder, duodenum, and a liver tumor by an experienced radiation oncologist were used as the ground truth for performance evaluation. Besides the manual segmentation, images were automatically segmented using thresholding, fuzzy k‐means (FKM), k‐harmonic means (KHM), and reaction‐diffusion level set evolution (RD‐LSE) algorithms, as well as the tissue tracking algorithm provided by the ViewRay treatment planning and delivery system (VR‐TPDS). The performance of the five algorithms was evaluated quantitatively by comparing with the manual segmentation using the Dice coefficient and target registration error (TRE) measured as the distance between the centroid of the manual ROI and the centroid of the automatically segmented ROI. All methods were able to successfully segment the bladder and the kidney, but only FKM, KHM, and VR‐TPDS were able to segment the liver tumor and the duodenum. The performance of the thresholding, FKM, KHM, and RD‐LSE algorithms degraded as the local image contrast decreased, whereas the performance of the VP‐TPDS method was nearly independent of local image contrast due to the reference registration algorithm. For segmenting high‐contrast images (i.e., kidney), the thresholding method provided the best speed (<1 ms) with a satisfying accuracy (Dice=0.95). When the image contrast was low, the VR‐TPDS method had the best automatic contour. Results suggest an image quality determination procedure before segmentation and a combination of different methods for optimal segmentation with the on‐board MR‐IGRT system.PACS number(s): 87.57.nm, 87.57.N‐, 87.61.Tg

An image regression motion prediction technique for MRI‐guided radiotherapy evaluated in single‐plane cine imaging

“…The distance between the tracked target centroid and the ground‐truth position, dice similarity between the tracked target contour and the ground‐truth target contour, and mutual information between the two registered images were evaluated for each five‐iteration interval. These registrations were compared to registrations initialized using all zeros (“none”), the DVF from the most recently acquired image (“previous”) and a DVF predicted using a linear extrapolation of the motion from the two most recent images (“extrapolation”) …”

“…Targets were tracked using a deformable image registration algorithm similar to that used by the ViewRay system. 3,24 Image registrations were performed using the Elastix package. [25][26][27] A multi-resolution b-spline deformable registration was performed using mutual information as a similarity measure.…”

“…These registrations were compared to registrations initialized using all zeros ("none"), the DVF from the most recently acquired image ("previous") and a DVF predicted using a linear extrapolation of the motion from the two most recent images ("extrapolation"). 24…”

An adaptive Fuzzy C-means method utilizing neighboring information for breast tumor segmentation in ultrasound images

“…). For the final contouring of the breast tumor, necessary morphological image processing steps were used for post‐processing …”

“…As an unsupervised method, the fuzzy c‐means (FCM) method is robust to noise and does not require initial contours and prior information . Therefore, FCM has been widely used to segment US images and images with blurred boundaries . Many modifications and improvements of the FCM method have been proposed, for example, relational FCM, spatial models, probability based FCM, and a neutrosophic c‐means (NCM) clustering algorithm .…”