David J. Vining

Virtual colonoscopy using oral contrast compared with colonoscopy for the detection of patients with colorectal polyps

BACKGROUND & AIMS Virtual colonoscopy using abdominal spiral computed tomography scanning allows total colonic evaluation with minimal invasiveness. Two-dimensional images and selective 3-dimensional images of the colon are used to detect colorectal lesions. This trial used conventional colonoscopy with segmental unblinding to determine the ability of virtual colonoscopy to identify patients with colorectal lesions who need conventional colonoscopy. METHODS We studied 205 patients with virtual colonoscopy using oral iodinated contrast preceding conventional colonoscopy. Colonic lavage was achieved with an oral sodium phosphosoda preparation and colonic distention with a carbon dioxide electronic insufflator. RESULTS The overall sensitivity and specificity of virtual colonoscopy in identifying patients with colorectal lesions was 61.8% and 70.7%, respectively. Virtual colonoscopy was more accurate in identifying patients with lesions >/=6 mm (sensitivity 84.4% and specificity 83.1%) and those with lesions >/=10 mm (sensitivity 90% and specificity 94.6%). The negative predictive value of virtual colonoscopy was 95% for a 6-mm cutoff size and 98.9% for a 10-mm cutoff. Using a 10-mm cutoff, virtual colonoscopy precludes the need for conventional colonoscopy in 86% of patients with a 1% false-negative rate (68% with a 3.4% false-negative rate when using a 6-mm cutoff). CONCLUSIONS Virtual colonoscopy has a high sensitivity and specificity for detecting patients with significant colorectal lesions. Its high negative predictive value may help reduce the number of negative screening colonoscopies. Further studies are needed to determine what lesion cutoff size is clinically acceptable and the appropriate interval time for repeat virtual colonoscopy when it detects lesions below this cutoff size.

Automatic segmentation of the colon for virtual colonoscopy

Virtual colonoscopy is a minimally invasive technique that enables early detection of colorectal polyps and cancer. Normally, a patients bowel is prepared with colonic lavage and gas insufflation prior to computed tomography scanning. An important step for 3D analysis of the image volume is segmentation of the colon. The high-contrast gas/tissue interface that exists in the colon lumen makes segmentation of the majority of the colon relatively easy; however, two factors inhibit automatic segmentation of the entire colon. First, the colon is not the only gas-filled organ in the data volume: lungs, small bowel, and stomach also meet this criterion. User-defined seed points placed in the colon lumen have previously been required to spatially isolate the colon. Second, portions of the colon lumen may be obstructed by peristalsis, large masses, and/or residual feces. These complicating factors require increased user interaction during the segmentation process to isolate additional colonic segments. To automate the segmentation of the colon, we have developed a method to locate seed points and segment the gas-filled lumen sections without user supervision. We have also developed an automated approach to improve lumen segmentation by digitally removing residual contrast-enhanced fluid. Experimental results with 20 patient volumes show that our method is accurate and reliable.

Computing the centerline of a colon: a robust and efficient method based on 3D skeletons.

We present a robust and efficient algorithm for calculating the centerline of a computer-generated colon model created from helical CT image data. The centerline is an essential aid for navigating through complex anatomy such as the colon. Our algorithm involves three steps. In the first step, we generate a 3D skeleton of the binary colon volume using a fast topological thinning algorithm. In the second step, we employ a graph search algorithm to remove extra loops and branches. These loops and branches are caused by holes in the object that are artifacts produced during image segmentation. In the final step, we compute a smooth representation of the centerline by approximating the skeleton with cubic B-splines. This final step is necessary because the skeleton contains many abrupt changes in direction due to the discrete nature of image data. The user supplies two endpoints for the centerline; otherwise, the algorithm is fully automated. Experimental results demonstrate that the algorithm is not only robust but also efficient.

Validation of virtual colonoscopy in the detection of colorectal polyps and masses: rationale for proper study design.

Background: Colorectal cancer, the second-leading cause of cancer-related mortality, is a preventable malignancy in many cases. Despite the availability of several screening modalities, compliance with screening recommendations remains unacceptably low. Virtual colonoscopy is a novel, minimally-invasive technique with the potential to increase colorectal cancer screening rates, but its effectiveness must first be validated. Published studies comparing virtual colonoscopy to conventional colonoscopy have reported varying results. These discrepancies may be attributed to differences in bowel preparation and scanning techniques, as well as errors in endoscopic lesion measurement, endoscopic colonic segmental localization, and the ability of conventional colonoscopy to actually detect lesions. These methodological issues can affect scientific results and ultimately affect the public’s perception of this emerging technique.Aim: The goal of this report is to expose existing methodological shortcomings and propose solutions incorporated in this study design. This article describes the rationale, study design, and outcome definitions of a single-center, blinded, direct comparative trial aiming at assessing the ability of virtual colonoscopy to detect colorectal polyps and masses relative to the criterion standard, conventional colonoscopy.Design Features: Bowel preparation was standardized using oral sodium phosphate lavage, orally administered iodinated contrast, and controlled colonic insufflation. Segmental unblinding allowed a second-look when results were discrepant and polyp matching was performed using an algorithm based on segmental localization and lesion size determination.Conclusions: This methodology could be applied to other studies assessing the accuracy of virtual colonoscopy in order to have uniformity of results.

3D Skeleton for Virtual Colonoscopy

This paper describes an improved algorithm for generating 3D skeletons from binary objects and its clinical application to virtual colonoscopy. A skeleton provides an ideal central path for an auto-piloted examination of a virtual colon rendered from a spiral computed tomography scan.

Lymph-node segmentation using active contours

Node volume analysis is very important medically. An automatic method of segmenting the node in spiral CT x-ray images is needed to produce accurate, consistent, and efficient volume measurements. The method of active contours (snakes) is proposed here as good solution to the node segmentation problem. Optimum parameterization and search strategies for using a two-dimensional snake to find node cross-sections are described, and an energy normalization scheme which preserves important spatial variations in energy is introduced. Three-dimensional segmentation is achieved without additional operator interaction by propagating the 2D results to adjacent slices. The method gives promising segmentation results on both simulated and real node images.

The Virtual Colonoscopy Study: a large multicenter clinical trial designed to compare two diagnostic screening procedures

This paper reviews the design of a large multicenter clinical trial currently being conducted to test the equivalence of two screening procedures for colorectal polyps. The primary outcome is the sensitivity and specificity of the new and standard procedures for detecting subjects with and without polyps of a size > or =6 mm, respectively. An important secondary outcome is the accuracy of these procedures in detecting individual polyps. A total of 619 participants underwent virtual colonoscopy, the new procedure, followed by conventional colonoscopy, the standard procedure. Strategies for the design and implementation of the study are shared as well as the challenges encountered.

Automatic segmentation of the colon

Virtual colonoscopy is a minimally invasive technique that enables detection of colorectal polyps and cancer. Normally, a patients bowel is prepared with colonic lavage and gas insufflation prior to computed tomography (CT) scanning. An important step for 3D analysis of the image volume is segmentation of the colon. The high-contrast gas/tissue interface that exists in the colon lumen makes segmentation of the majority of the colon relatively easy; however, two factors inhibit automatic segmentation of the entire colon. First, the colon is not the only gas-filled organ in the data volume: lungs, small bowel, and stomach also meet this criteria. User-defined seed points placed in the colon lumen have previously been required to spatially isolate only the colon. Second, portions of the colon lumen may be obstructed by peristalsis, large masses, and/or residual feces. These complicating factors require increased user interaction during the segmentation process to isolate additional colon segments. To automate the segmentation of the colon, we have developed a method to locate seed points and segment the gas-filled lumen with no user supervision. We have also developed an automated approach to improve lumen segmentation by digitally removing residual contrast-enhanced fluid resulting from a new bowel preparation that liquefies and opacifies any residual feces.

Segmentation in virtual colonoscopy using a geometric deformable model

The Geometric Deformable Model is developed for accurate colon lumen segmentation as part of an automatic Virtual Colonoscopy system. The deformable model refines the lumen surface found by an automatic seed location and thresholding procedure. The challenges to applying the deformable model are described, showing the definition of the stopping function as the key to accurate segmentation. The limitations of current stopping criteria are examined and a new definition, tailored to the task of colon segmentation, is given. First, a multiscale edge operator is used to locate high confidence boundaries. These boundaries are then integrated into the stopping function using a distance transform. The hypothesis is that the new stopping function results in a more accurate representation of the lumen surface compared to previous monotonic functions of the gradient magnitude. This hypothesis is tested using observer ratings of colon surface fidelity at 100 hundred randomly selected locations in each of four datasets. The results show that the surfaces determined by the modified deformable model better represent the lumen surface overall.

Virtual endoscopy: quicker and easier disease evaluation

The advent of spiral computed tomography (CT) has created the potential to image continuous anatomical volumes during a single breath-hold. The ability to reconstruct overlapping spiral CT images has improved through-plane resolution and contributed to improved diagnostic accuracy. When spiral CT is used to image organ systems such as the colon or airways, it is common to generate up to 500 CT images. We have developed a virtual endoscopy (VE) software system that couples computer-assisted diagnosis capabilities with volume visualization techniques to aid in the analysis of these large datasets. Despite its potential to assist in disease diagnosis, VE faces several important technical and nontechnical challenges that must be addressed before it becomes a clinical reality.