This is the accepted version of the paper.This version of the publication may differ from the final published version. Permanent repository link Abstract.Purpose: Computed tomographic (CT) colonography is a relatively new technique for detecting bowel cancer or potentially precancerous polyps. CT scanning is combined with 3-dimensional image reconstruction to produce a virtual endoluminal representation similar to optical colonoscopy. Because retained fluid and stool can mimic pathology, CT data is acquired with the bowel cleansed and insufflated with gas and patient in both prone and supine positions. Radiologists then match visually endoluminal locations between the two acquisitions in order to determine whether apparent pathology is real or not. This process is hindered by the fact that the colon, essentially a long tube, can undergo considerable deformation between acquisitions. We present a novel approach to automatically establish spatial correspondence between prone and supine endoluminal colonic surfaces after surface parameterization, even in the case of local colon collapse. Methods:The complexity of the registration task was reduced from a 3D to a 2D problem by mapping the surfaces extracted from prone and supine CT colonography onto a cylindrical parameterization. A non-rigid cylindrical registration was then performed to align the full colonic surfaces. The curvature information from the original 3D surfaces was used to determine correspondence. The method can also be applied to cases with regions of local colonic collapse by ignoring the collapsed regions during the registration.Results: Using a development set, suitable parameters were found to constrain the cylindrical registration method. Then, the same registration parameters were applied to a different set of 13 validation cases, consisting of 8 fully distended cases and 5 cases exhibiting multiple colonic collapses. All polyps present were well aligned, with a mean (± std. dev.) registration error of 5.7 (± 3.4) mm. An additional set of 1175 reference points on haustral folds spread over the full endoluminal colon surfaces resulted in an error of 7.7 (± 7.4) mm. Here, 82% of folds was aligned correctly after registration with a further 15% misregistered by just one fold. Conclusions:The proposed method reduces the 3D registration task to a cylindrical registration representing the endoluminal surface of the colon. Our algorithm uses surface curvature information as a similarity measure to drive registration to compensate for the large colorectal deformations that occur between prone and supine data acquisitions. The method has the potential to both enhance polyp detection and decrease the radiologist's interpretation time.Registration of the endoluminal surfaces of the colon derived from CTC
Abstract. CT colonography is routinely performed with the patient prone and supine to differentiate fixed colonic pathology from mobile faecal residue. We propose a novel method to automatically establish correspondence. Haustral folds are detected using a graph cut method applied to a surface curvature-based metric, where image patches are generated using endoluminal CT colonography surface rendering. The intensity difference between image pairs, along with additional neighbourhood information to enforce geometric constraints, are used with a Markov Random Field (MRF) model to estimate the fold labelling assignment. The method achieved fold matching accuracy of 83.1% and 88.5% with and without local colonic collapse. Moreover, it improves an existing surface-based registration algorithm, decreasing mean registration error from 9.7mm to 7.7mm in cases exhibiting collapse.
This is the accepted version of the paper.This version of the publication may differ from the final published version. Abstract. Matching corresponding location between prone and supine acquisitions for CT colonography (CTC) is essential to verify the existence of a polyp, which can be a difficult task due to the considerable deformations that will often occur to the colon during repositioning of the patient. This can induce error and increase interpretation time. We propose a novel method to automatically establish correspondence between the two acquisitions. A first step segments a set of haustral folds in each view and determines correspondence via a labelling process using a Markov Random Field (MRF) model. We show how the landmark correspondences can be used to non-rigidly transform a 2D source image derived from a conformal mapping process on the 3D endoluminal surface mesh to achieve full surface correspondence between prone and supine views. This can be used to initialise an intensity-based non-rigid B-spline registration method which further increases the accuracy. We demonstrate a statistically significant improvement over the intensity based non-rigid B-spline registration by using the composite method. Permanent repository link
This is the accepted version of the paper.This version of the publication may differ from the final published version. Abstract. Robust registration between prone and supine data acquisitions for CT colonography is pivotal for medical interpretation but a challenging problem. One measure when evaluating non-rigid registration algorithms over the whole of the deformation field is the inverse consistency error, which suggests improved registration quality when the inverse deformation is consistent with the forward deformation. We show that using computed landmark displacements to initialise an intensity based registration reduces the inverse consistency error when using a state-of-the-art non-rigid b-spline registration method. This method aligns prone and supine 2D images derived from CT colonography acquisitions in a cylindrical domain. Furthermore, we demonstrate that using the same initialisation also improves registration accuracy for a set of manually identified reference points in cases exhibiting local luminal collapse. Permanent
Purpose:To evaluate the accuracy of a method of automatic coregistration of the endoluminal surfaces at computed tomographic (CT) colonography performed on separate occasions to facilitate identification of polyps in patients undergoing polyp surveillance. Materials and Methods:Institutional review board and HIPAA approval were obtained. A registration algorithm that was designed to coregister the coordinates of endoluminal colonic surfaces on images from prone and supine CT colonographic acquisitions was used to match polyps in sequential studies in patients undergoing polyp surveillance. Initial and follow-up CT colonographic examinations in 26 patients (35 polyps) were selected and the algorithm was tested by means of two methods, the longitudinal method (polyp coordinates from the initial prone and supine acquisitions were used to identify the expected polyp location automatically at follow-up CT colonography) and the consistency method (polyp coordinates from the initial supine acquisition were used to identify polyp location on images from the initial prone acquisition, then on those for follow-up prone and follow-up supine acquisitions). Two observers measured the Euclidean distance between true and expected polyp locations, and mean per-patient registration accuracy was calculated. Segments with and without collapse were compared by using the Kruskal-Wallace test, and the relationship between registration error and temporal separation was investigated by using the Pearson correlation. Results:Coregistration was achieved for all 35 polyps by using both longitudinal and consistency methods. Mean 6 standard deviation Euclidean registration error for the longitudinal method was 17.4 mm 6 12.1 and for the consistency method, 26.9 mm 6 20.8. There was no significant difference between these results and the registration error when prone and supine acquisitions in the same study were compared (16.9 mm 6 17.6; P = .451). Conclusion:Automatic endoluminal coregistration by using an algorithm at initial CT colonography allowed prediction of endoluminal polyp location at subsequent CT colonography, thereby facilitating detection of known polyps in patients undergoing CT colonographic surveillance.q RSNA, 2014
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.