Automatic correction of level set based subvoxel precise centerlines for virtual colonoscopy using the colon outer wall

Immersive Virtual Colonoscopy

Virtual colonoscopy (VC) is a non-invasive screening tool for colorectal polyps which employs volume visualization of a colon model reconstructed from a CT scan of the patient's abdomen. We present an immersive analytics system for VC which enhances and improves the traditional desktop VC through the use of VR technologies. Our system, using a head-mounted display (HMD), includes all of the standard VC features, such as the volume rendered endoluminal fly-through, measurement tool, bookmark modes, electronic biopsy, and slice views. The use of VR immersion, stereo, and wider field of view and field of regard has a positive effect on polyp search and analysis tasks in our immersive VC system, a volumetric-based immersive analytics application. Navigation includes enhanced automatic speed and direction controls, based on the user's head orientation, in conjunction with physical navigation for exploration of local proximity. In order to accommodate the resolution and frame rate requirements for HMDs, new rendering techniques have been developed, including mesh-assisted volume raycasting and a novel lighting paradigm. Feedback and further suggestions from expert radiologists show the promise of our system for immersive analysis for VC and encourage new avenues for exploring the use of VR in visualization systems for medical diagnosis.

A CAD of fully automated colonic polyp detection for contrasted and non-contrasted CT scans

PURPOSE

Computer-aided detection (CAD) systems are developed to help radiologists detect colonic polyps over CT scans. It is possible to reduce the detection time and increase the detection accuracy rates by using CAD systems. In this paper, we aimed to develop a fully integrated CAD system for automated detection of polyps that yields a high polyp detection rate with a reasonable number of false positives.

METHODS

The proposed CAD system is a multistage implementation whose main components are: automatic colon segmentation, candidate detection, feature extraction and classification. The first element of the algorithm includes a discrete segmentation for both air and fluid regions. Colon-air regions were determined based on adaptive thresholding, and the volume/length measure was used to detect air regions. To extract the colon-fluid regions, a rule-based connectivity test was used to detect the regions belong to the colon. Potential polyp candidates were detected based on the 3D Laplacian of Gaussian filter. The geometrical features were used to reduce false-positive detections. A 2D projection image was generated to extract discriminative features as the inputs of an artificial neural network classifier.

RESULTS

Our CAD system performs at 100% sensitivity for polyps larger than 9 mm, 95.83% sensitivity for polyps 6-10 mm and 85.71% sensitivity for polyps smaller than 6 mm with 5.3 false positives per dataset. Also, clinically relevant polyps ([Formula: see text]6 mm) were identified with 96.67% sensitivity at 1.12 FP/dataset.

CONCLUSIONS

To the best of our knowledge, the novel polyp candidate detection system which determines polyp candidates with LoG filters is one of the main contributions. We also propose a new 2D projection image calculation scheme to determine the distinctive features. We believe that our CAD system is highly effective for assisting radiologist interpreting CT.

Post-processing techniques for making reliable measurements from curve-skeletons

Interconnected 3-D networks occur widely in biology and the geometry of such branched networks can be described by curve-skeletons, allowing parameters such as path lengths, path tortuosities and cross-sectional thicknesses to be quantified. However, curve-skeletons are typically sensitive to small scale surface features which may arise from noise in the imaging data. In this paper, new post-processing techniques for curve-skeletons are presented which ensure that measurements of lengths and thicknesses are less sensitive to these small scale surface features. The techniques achieve sub-voxel accuracy and are based on a minimal sphere-network representation in which the object is modelled as a string of minimally overlapping spheres, and as such samples the object on a scale related to the local thickness. A new measure of cross-sectional dimension termed the modal radius is defined and shown to be more robust in comparison with the standard measure (the internal radius), while retaining the desirable feature of capturing the size of structures in terms of a single measure. The techniques are demonstrated by application to trabecular bone and tumour vascular network case studies where the volumetric data was obtained by high resolution computed tomography.

Colon centerline extraction in fragmented segmentations

In virtual colonoscopy, the clinical need is a smooth centered path from the rectum to the cecum, for interactive navigation along the colonic lumen. The primary challenge is breakages in the colon, due to fecal residue, abnormalities, poor insufflation and inadequate electronic cleansing. Here we propose a method, that is a modification of the classic energy minimized geodesic, that extracts centered paths through fragmented colons. To begin, we perform electronic cleansing, automatically localize 4 points: rectum, cecum, sphlenic and hepatic flexures; followed by region growing and heuristic approaches to generate the initial segmentation. This is followed by a daisy chaining procedure to link possibly large colon blobs that may have been missed as weaker candidate segmentations. We then perform a front propagation to extract a minimal energy path through the ordered set of points. This propagation is guided by multiple forces: (a) A strong force given by the distance to the colon segmentation surface (b) A weak force derived from the CT intensity (c) A weak force from the distance to the surface of weaker candidate colon segmentations (d) A geodesic repulsive force, where the other points exhibit an repelling force in their voronoi partition, the force proportional to the geodesic distance to the point. Our contribution is a path extraction method for the colon that is the energy minimized geodesic (a) favouring centeredness (b) punching through gaps, traversing in so far as possible through lower intensity regions and possibly centered within these gaps (c) ordered through the feature points. Results show improvements of the method over the standard minimal energy path approach.

Virtual colon flattening method based on colonic outer surface

Virtual colon flattening (VF) is a minimally invasive viewing mode used to detect colorectal polyps on the colonic inner surface in virtual colonoscopy. Compared with conventional colonoscopy, inspecting a flattened colonic inner surface is faster and results in fewer uninspected regions. Unfortunately, the deformation distortions of flattened colonic inner surface impede the performance of VF. Conventionally, the deformation distortions can be corrected by using the colonic inner surface. However, colonic curvatures and haustral folds make correcting deformation distortions using only the colonic inner surface difficult. Therefore, we propose a VF method that is based on the colonic outer surface. The proposed method includes two novel algorithms, namely, the colonic outer surface extraction algorithm and the colonic outer surface-based distortion correction algorithm. Sixty scans involving 77 annotated polyps were used for the validation. The flattened colons were independently inspected by three operators and then compared with three existing VF methods. The correct detection rates of the proposed method and the three existing methods were 79.6%, 67.1%, 71.9%, and 72.7%, respectively, and the false positives per scan were 0.16, 0.32, 0.21, and 0.26, respectively. The experimental results demonstrate that our proposed method has better performance than existing methods that are based on the colonic inner surface.

Topology preserving thinning of cell complexes

A topology preserving skeleton is a synthetic representation of an object that retains its topology and many of its significant morphological properties. The process of obtaining the skeleton, referred to as skeletonization or thinning, is a very active research area. It plays a central role in reducing the amount of information to be processed during image analysis and visualization, computer-aided diagnosis, or by pattern recognition algorithms. This paper introduces a novel topology preserving thinning algorithm, which removes simple cells-a generalization of simple points-of a given cell complex. The test for simple cells is based on acyclicity tables automatically produced in advance with homology computations. Using acyclicity tables render the implementation of thinning algorithms straightforward. Moreover, the fact that tables are automatically filled for all possible configurations allows to rigorously prove the generality of the algorithm and to obtain fool-proof implementations. The novel approach enables, for the first time, according to our knowledge, to thin a general unstructured simplicial complex. Acyclicity tables for cubical and simplicial complexes and an open source implementation of the thinning algorithm are provided as an additional material to allow their immediate use in the vast number of applications arising in medical imaging and beyond.

Virtual colon flattening based on colonic outer surface

Virtual colon flattening (VF) is a non-invasive procedure to inspect the colonic inner surface for detecting colorectal polyps. Unfortunately, the performance of VF is impeded by deformation distortions of colonic inner surface. Conventionally, the colonic inner surface itself is used to correct deformation distortions. In this paper, we propose a colonic outer surface based VF method to correct distortions instead of colonic inner surface. The proposed method was validated with 60 cases and 200 annotated polyps. Visual inspections were carried out by three operators independently and were compared with three existing VF methods which are based on colonic inner surface. The correct detection rate of the proposed method and the three existing methods were 88.0%, 76.5%, 80.0% and 81.5% respectively. False positives per case were 0.16, 0.32, 0.21, and 0.26 respectively. The proposed method has higher correct detection rate and less false positives than the other three VF methods, demonstrating the usefulness of colonic outer surface as a correction tool for VF results.

Predicting polyp location on optical colonoscopy from CT colonography by minimal-energy curve modeling of the colonoscope path

The ability to accurately locate a polyp found on computed tomographic colonography (CTC) at subsequent optical colonoscopy (OC) is an important task in colorectal cancer screening. We present a method to more accurately match polyp locations at CTC and OC. A colonoscope was modeled as a flexible tube with negligible stretch and minimal strain. The path of the colonoscope was estimated using a minimal-energy curve method. The energy function was defined and optimized by a subdivision scheme. The prediction of polyp locations at OC from CTC was converted to an optimization problem. The prediction performance was evaluated on 134 polyps by comparing the predicted with the true polyp locations at OC. The method can accurately predict polyp locations at OC to within ±0.5 colonoscope mark (5 cm) for more than 58% of polyps and to within ±1 colonoscope mark (10 cm) for more than 96% of polyps, significantly improving upon previously published methods. This method can be easily incorporated into routine OC practice and allow the colonoscopist to begin the examination by targeting locations of potential polyps found at CTC.

Machine learning and radiology

In this paper, we give a short introduction to machine learning and survey its applications in radiology. We focused on six categories of applications in radiology: medical image segmentation, registration, computer aided detection and diagnosis, brain function or activity analysis and neurological disease diagnosis from fMR images, content-based image retrieval systems for CT or MRI images, and text analysis of radiology reports using natural language processing (NLP) and natural language understanding (NLU). This survey shows that machine learning plays a key role in many radiology applications. Machine learning identifies complex patterns automatically and helps radiologists make intelligent decisions on radiology data such as conventional radiographs, CT, MRI, and PET images and radiology reports. In many applications, the performance of machine learning-based automatic detection and diagnosis systems has shown to be comparable to that of a well-trained and experienced radiologist. Technology development in machine learning and radiology will benefit from each other in the long run. Key contributions and common characteristics of machine learning techniques in radiology are discussed. We also discuss the problem of translating machine learning applications to the radiology clinical setting, including advantages and potential barriers.

Fully automated colon segmentation for the computation of complete colon centerline in virtual colonoscopy

Virtual colonoscopy detects polyps by navigating along a colon centerline. Complete colon segmentation based on computed tomography (CT) data is a prerequisite to the computation of complete colon centerline. There are two main problems impeding complete segmentation: overdistention/underdistention of colon and the use of oral contrast agents. Overdistention produces loops in the segmented colon, while underdistention may cause the segmented colon collapse into a series of disconnected segments. Use of oral contrast agents, which have high attenuation on CT, may add redundant structures (bones and small bowels) to the segmented colon. A fully automated colon segmentation method is proposed in this paper to address the two problems. We tested the proposed method in 170 cases, including 37 "moderate" and 133 "challenging" cases. Computer-generated centerlines were compared with human-generated centerlines (plotted by three radiologists). The proposed method achieved a 90.56% correct coverage rate with respect to the human-generated centerlines. We also compared the proposed method with two existing colon segmentation methods: Uitert's method and Nappi's method. The results of these two methods were 75.16% and 72.59% correct coverage rates, respectively. Our experimental results indicate that the proposed method could yield more complete colon centerlines than the existing methods.

Context preserving maps of tubular structures

When visualizing tubular 3D structures, external representations are often used for guidance and display, and such views in 2D can often contain occlusions. Virtual dissection methods have been proposed where the entire 3D structure can be mapped to the 2D plane, though these will lose context by straightening curved sections. We present a new method of creating maps of 3D tubular structures that yield a succinct view while preserving the overall geometric structure. Given a dominant view plane for the structure, its curve skeleton is first projected to a 2D skeleton. This 2D skeleton is adjusted to account for distortions in length, modified to remove intersections, and optimized to preserve the shape of the original 3D skeleton. Based on this shaped 2D skeleton, a boundary for the map of the object is obtained based on a slicing path through the structure and the radius around the skeleton. The sliced structure is conformally mapped to a rectangle and then deformed via harmonic mapping to match the boundary placement. This flattened map preserves the general geometric context of a 3D object in a 2D display, and rendering of this flattened map can be accomplished using volumetric ray casting. We have evaluated our method on real datasets of human colon models.

Virtual eversion and rotation of colon based on outer surface centerline

PURPOSE

Virtual eversion turns the colon's inner surface to its outside while maintaining the original colon path. The virtually everted colon allows both global and local views of the mucosal surface for observation. However, the conventional colon's inner surface centerline commonly used in virtual colonoscopy and virtual flattening is not suitable for virtual eversion. Therefore, the colon's outer surface centerline is introduced for virtual eversion to produce a more accurate representation.

METHODS

An improved level set segmentation method is presented for generating the colon's outer surface. To achieve eversion with fewer errors, the centerline of the colon's outer surface is employed in the proposed virtual eversion method instead of the inner surface centerline. A hybrid sampling method is designed to accelerate the eversion. Virtual rotation is introduced to visualize the lateral and rear views of the colon better. The gathered structures in the high curvature regions can be separated by virtual rotation.

RESULTS

The proposed methods were validated using two three-dimensional phantoms and 87 CT data sets. A study on the observation performance of the everted data showed that the reading times were (63% of time reduction for phantom A, 65% of time reduction for phantom B, and 77% of time reduction for CT data) less than those using virtual colonoscopy, while maintaining the sensibility. The incidence of improperly everted regions in the virtual eversion based on the outer surface centerline was 71% less than that based on the inner surface centerline.

CONCLUSIONS

The virtual eversion based on the outer surface centerline is more accurate than the one based on the inner surface centerline whether the colon's inner surface is smooth or ragged. The time required for polyp detection using the virtual eversion is considerably less than that using the conventional virtual endoscopy. Virtual eversion and virtual rotation are promising methods for the rapid location of colonic polyps. Together with virtual colonoscopy and virtual flattening, virtual eversion and virtual rotation can be integrated to produce a powerful system for diagnosing colonic lesions.

Image analysis for classification of dysplasia in Barrett's esophagus using endoscopic optical coherence tomography

Barrett's esophagus (BE) and associated adenocarcinoma have emerged as a major health care problem. Endoscopic optical coherence tomography is a microscopic sub-surface imaging technology that has been shown to differentiate tissue layers of the gastrointestinal wall and identify dysplasia in the mucosa, and is proposed as a surveillance tool to aid in management of BE. In this work a computer-aided diagnosis (CAD) system has been demonstrated for classification of dysplasia in Barrett's esophagus using EOCT. The system is composed of four modules: region of interest segmentation, dysplasia-related image feature extraction, feature selection, and site classification and validation. Multiple feature extraction and classification methods were evaluated and the process of developing the CAD system is described in detail. Use of multiple EOCT images to classify a single site was also investigated. A total of 96 EOCT image-biopsy pairs (63 non-dysplastic, 26 low-grade and 7 high-grade dysplastic biopsy sites) from a previously described clinical study were analyzed using the CAD system, yielding an accuracy of 84% for classification of non-dysplastic vs. dysplastic BE tissue. The results motivate continued development of CAD to potentially enable EOCT surveillance of large surface areas of Barrett's mucosa to identify dysplasia.

Semi-automated segmentation of the sigmoid and descending colon for radiotherapy planning using the fast marching method

A fast and accurate segmentation of organs at risk, such as the healthy colon, would be of benefit for planning of radiotherapy, in particular in an adaptive scenario. For the treatment of pelvic tumours, a great challenge is the segmentation of the most adjacent and sensitive parts of the gastrointestinal tract, the sigmoid and descending colon. We propose a semi-automated method to segment these bowel parts using the fast marching (FM) method. Standard 3D computed tomography (CT) image data obtained from routine radiotherapy planning were used. Our pre-processing steps distinguish the intestine, muscles and air from connective tissue. The core part of our method separates the sigmoid and descending colon from the muscles and other segments of the intestine. This is done by utilizing the ability of the FM method to compute a specified minimal energy functional integrated along a path, and thereby extracting the colon centre line between user-defined control points in the sigmoid and descending colon. Further, we reconstruct the tube-shaped geometry of the sigmoid and descending colon by fitting ellipsoids to points on the path and by adding adjacent voxels that are likely voxels belonging to these bowel parts. Our results were compared to manually outlined sigmoid and descending colon, and evaluated using the Dice coefficient (DC). Tests on 11 patients gave an average DC of 0.83 (+/-0.07) with little user interaction. We conclude that the proposed method makes it possible to fast and accurately segment the sigmoid and descending colon from routine CT image data.

Reversible projection technique for colon unfolding

Colon unfolding provides an efficient way to navigate the colon in computed tomographic colonography (CTC). Most existing unfolding techniques only compute forward projections. When radiologists find abnormalities or conduct measurements on the unfolded view (which is often quicker and easier), it is difficult to locate the corresponding region on the 3-D view for further examination (which is more accurate and reliable). To address this, we propose a reversible projection technique for colon unfolding. The method makes use of advanced algorithms including rotation-minimizing frames, recursive ring sets, mesh skinning, and cylindrical projection. Both forward and reverse mapping can be computed for points on the colon surface. Therefore, it allows for detecting and measuring polyps on the unfolded view and mapping them back to the 3-D surface. We generated realistic colon simulation data incorporating most colon characteristics, such as curved centerline, variable distention, haustral folds, teniae coli, and colonic polyps. Our method was tested on both simulated data and data from 110 clinical CTC studies. The results showed submillimeter accuracy in simulated data and -0.23 ± 1.67 mm in the polyp measurement using clinical CTC data. The major contributions of our technique are: 1) the use of a recursive ring set method to solve the centerline and surface correspondence problem; 2) reverse transformation from the unfolded view to the 3-D view; and 3) quantitative validation using a realistic colon simulation and clinical CTC polyp measurement.

Registration of prone and supine CT colonography scans using correlation optimized warping and canonical correlation analysis

PURPOSE

In computed tomographic colonography (CTC), a patient will be scanned twice-Once supine and once prone-to improve the sensitivity for polyp detection. To assist radiologists in CTC reading, in this paper we propose an automated method for colon registration from supine and prone CTC scans.

METHODS

We propose a new colon centerline registration method for prone and supine CTC scans using correlation optimized warping (COW) and canonical correlation analysis (CCA) based on the anatomical structure of the colon. Four anatomical salient points on the colon are first automatically distinguished. Then correlation optimized warping is applied to the segments defined by the anatomical landmarks to improve the global registration based on local correlation of segments. The COW method was modified by embedding canonical correlation analysis to allow multiple features along the colon centerline to be used in our implementation.

RESULTS

We tested the COW algorithm on a CTC data set of 39 patients with 39 polyps (19 training and 20 test cases) to verify the effectiveness of the proposed COW registration method. Experimental results on the test set show that the COW method significantly reduces the average estimation error in a polyp location between supine and prone scans by 67.6%, from 46.27 +/- 52.97 to 14.98 mm +/- 11.41 mm, compared to the normalized distance along the colon centerline algorithm (p < 0.01).

CONCLUSIONS

The proposed COW algorithm is more accurate for the colon centerline registration compared to the normalized distance along the colon centerline method and the dynamic time warping method. Comparison results showed that the feature combination of z-coordinate and curvature achieved lowest registration error compared to the other feature combinations used by COW. The proposed method is tolerant to centerline errors because anatomical landmarks help prevent the propagation of errors across the entire colon centerline.

Temporal and multiinstitutional quality assessment of CT colonography

OBJECTIVE

The purpose of this study was to investigate the variability of CT colonography (CTC) scan quality obtained within and between institutions by using previously validated automated quality assessment (QA) software that assesses colonic distention and surface area obscured by residual fluid.

MATERIALS AND METHODS

The CTC scans of 120 patients were retrospectively selected, 30 from each of four institutions. The bowel preparation included oral contrast material for fecal and fluid tagging. Patients at one institution (institution 4) drank half the amount of oral contrast material compared with the patients at the other three institutions. Fifteen of the CTC scans were from the beginning of the protocol studied at each institution and 15 scans were from the same protocol acquired approximately 1 year later in the study. We used previously validated QA software to automatically measure the mean distention and residual fluid of each of five colonic segments (ascending, transverse, descending, sigmoid, and rectum). Adequate distention was defined as a colonic diameter of at least 2 cm. Residual fluid was determined by the percentage of colonic surface area covered by fluid. We compared how the quality varied across multiple institutions and over time within the same institution.

RESULTS

No significant difference in the amount of colonic distention among the four institutions was found (p = 0.19). However, the distention in the prone position was significantly greater than the distention in the supine position (p < 0.001). Patients at institution 4 had about half the amount of residual colonic fluid compared with patients at the other three institutions (p < 0.01). The sigmoid and descending colons were the least distended segments, and the transverse and descending colons contained the most fluid on the prone and supine scans, respectively. More recently acquired studies had greater distention and less residual fluid, but the differences were not statistically significant (p = 0.30 and p = 0.96, respectively).

CONCLUSION

Across institutions, a significant difference can exist in bowel preparation quality for CTC. This study reaffirms the need for standardized bowel preparation and quality monitoring of CTC examinations to reduce poor CTC performance.

Fast path planning in virtual colonoscopy

We propose a fast path planning algorithm using multi-resolution path tree propagation and farthest visible point. Initial path points are robustly generated by propagating the path tree, and all internal voxels locally most distant from the colon boundary are connected. The multi-resolution scheme is adopted to increase computational efficiency. Control points representing the navigational path are successively selected from the initial path points by using the farthest visible point. The position of the initial path point in a down-sampled volume is accurately adjusted in the original volume. Using the farthest visible point, the number of control points is adaptively changed according to the curvature of the colon shape so that more control points are assigned to highly curved regions. Furthermore, a smoothing step is unnecessary since our method generates a set of control points to be interpolated with the cubic spline interpolation. We applied our method to 10 computed tomography datasets. Experimental results showed that the path was generated much faster than using conventional methods without sacrificing accuracy, and clinical efficiency. The average processing time was approximately 1s when down-sampling by a factor of 2, 3, or 4. We concluded that our method is useful in diagnosing colon cancer using virtual colonoscopy.