Franciscus M. Vos

Automatic Detection and Segmentation of Crohn's Disease Tissues From Abdominal MRI

We propose an information processing pipeline for segmenting parts of the bowel in abdominal magnetic resonance images that are affected with Crohns disease. Given a magnetic resonance imaging test volume, it is first oversegmented into supervoxels and each supervoxel is analyzed to detect presence of Crohns disease using random forest (RF) classifiers. The supervoxels identified as containing diseased tissues define the volume of interest (VOI). All voxels within the VOI are further investigated to segment the diseased region. Probability maps are generated for each voxel using a second set of RF classifiers which give the probabilities of each voxel being diseased, normal or background. The negative log-likelihood of these maps are used as penalty costs in a graph cut segmentation framework. Low level features like intensity statistics, texture anisotropy and curvature asymmetry, and high level context features are used at different stages. Smoothness constraints are imposed based on semantic information (importance of each feature to the classification task) derived from the second set of learned RF classifiers. Experimental results show that our method achieves high segmentation accuracy with Dice metric values of 0.90 ± 0.04 and Hausdorff distance of 7.3 ± 0.8 mm. Semantic information and context features are an integral part of our method and are robust to different levels of added noise.

Classifying CT Image Data Into Material Fractions by a Scale and Rotation Invariant Edge Model

A fully automated method is presented to classify 3-D CT data into material fractions. An analytical scale-invariant description relating the data value to derivatives around Gaussian blurred step edges - arch model - is applied to uniquely combine robustness to noise, global signal fluctuations, anisotropic scale, noncubic voxels, and ease of use via a straightforward segmentation of 3-D CT images through material fractions. Projection of noisy data value and derivatives onto the arch yields a robust alternative to the standard computed Gaussian derivatives. This results in a superior precision of the method. The arch-model parameters are derived from a small, but over-determined, set of measurements (data values and derivatives) along a path following the gradient uphill and downhill starting at an edge voxel. The model is first used to identify the expected values of the two pure materials (named and ) and thereby classify the boundary. Second, the model is used to approximate the underlying noise-free material fractions for each noisy measurement. An iso-surface of constant material fraction accurately delineates the material boundary in the presence of noise and global signal fluctuations. This approach enables straightforward segmentation of 3-D CT images into objects of interest for computer-aided diagnosis and offers an easy tool for the design of otherwise complicated transfer functions in high-quality visualizations. The method is applied to segment a tooth volume for visualization and digital cleansing for virtual colonoscopy.

Lesion Conspicuity and Efficiency of CT Colonography with Electronic Cleansing Based on a Three-Material Transition Model

OBJECTIVE The purpose of this article is to report the effect on lesion conspicuity and the practical efficiency of electronic cleansing for CT colonography (CTC). MATERIALS AND METHODS Patients were included from the Walter Reed Army Medical Center public database. All patients had undergone extensive bowel preparation with fecal tagging. A primary 3D display method was used. For study I, the data consisted of all patients with polyps > or = 6 mm. Two experienced CTC observers (observer 1 and observer 2) scored the lesion conspicuity considering supine and prone positions separately. For study II, data consisted of 19 randomly chosen patients from the database. The same observers evaluated the data before and after electronic cleansing. Evaluation time, assessment effort, and observer confidence were recorded. RESULTS In study I, there were 59 lesions partly or completely covered by tagged material (to be uncovered by electronic cleansing) and 70 lesions surrounded by air (no electronic cleansing required). The conspicuity did not differ significantly between lesions that were uncovered by electronic cleansing and lesions surrounded by air (observer 1, p < 0.5; observer 2, p < 0.6). In study II, the median evaluation time per patient after electronic cleansing was significantly shorter than for original data (observer 1, 20 reduced to 12 minutes; observer 2, 17 reduced to 12 minutes). Assessment effort was significantly smaller for both observers (p < 0.0000001), and observer confidence was significantly larger (observer 1, p < 0.007; observer 2, p < 0.0002) after electronic cleansing. CONCLUSION Lesions uncovered by electronic cleansing have comparable conspicuity with lesions surrounded by air. CTC with electronic cleansing sustains a shorter evaluation time, lower assessment effort, and larger observer confidence than without electronic cleansing.

Computational modeling for assessment of IBD: To be or not to be?

The grading of inflammatory bowel disease (IBD) severity is important to determine the proper treatment strategy and to quantify the response to treatment. Traditionally, ileocolonoscopy is considered the reference standard for assessment of IBD. However, the procedure is invasive and requires extensive bowel preparation. Magnetic resonance imaging (MRI) has become an important tool for determining the presence of disease activity. Unfortunately, only moderate interobserver agreement is reported for most of the radiological severity measures. There is a clear demand for automated evaluation of MR images in Crohns disease (CD). This paper aims to introduce a preliminary suite of fundamental tools for assessment of CD severity. It involves procedures for image analysis, classification and visualization to predict the colonoscopy disease scores.

A supervised learning based approach to detect crohn's disease in abdominal MR volumes

Accurate diagnosis of Crohns disease (CD) has emerged as an important medical challenge. Because current Magnetic resonance imaging (MRI) analysis approaches rely on extensive manual segmentation for an accurate analysis, we propose a method for the automatic identification and localization of regions in abdominal MR volumes that have been affected by CD. Our proposed approach will serve to augment results from colonoscopy, the current reference standard for CD diagnosis. Intensity statistics, texture anisotropy and shape asymmetry of the 3D regions are used as features to distinguish between diseased and normal regions. Particular emphasis is laid on a novel entropy based asymmetry calculation method. Experiments on real patient data show that our features achieve a high level of accuracy and perform better than two competing methods.

Semi-Supervised and Active Learning for Automatic Segmentation of Crohn’s Disease

Our proposed method combines semi supervised learning (SSL) and active learning (AL) for automatic detection and segmentation of Crohns disease (CD) from abdominal magnetic resonance (MR) images. Random forest (RF) classifiers are used due to fast SSL classification and capacity to interpret learned knowledge. Query samples for AL are selected by a novel information density weighted approach using context information, semantic knowledge and labeling uncertainty. Experimental results show that our proposed method combines the advantages of SSL and AL, and with fewer samples achieves higher classification and segmentation accuracy over fully supervised methods.

Crohn's disease tissue segmentation from abdominal MRI using semantic information and graph cuts

We propose a graph cut based method to segment regions in abdominal magnetic resonance (MR) images affected with Crohns disease (CD). Intensity, texture, curvature and context information are used with random forest (RF) classifiers to calculate probability maps for graph cut segmentation. The RF classifiers also provide semantic information used to design a novel smoothness cost. Experimental results on 26 real patient data shows our method accurately segments the diseased areas. Inclusion of semantic information significantly improves segmentation accuracy and its importance is reflected in quantitative measures and visual results.

A supervised learning approach for Crohn's disease detection using higher-order image statistics and a novel shape asymmetry measure.

Increasing incidence of Crohn’s disease (CD) in the Western world has made its accurate diagnosis an important medical challenge. The current reference standard for diagnosis, colonoscopy, is time-consuming and invasive while magnetic resonance imaging (MRI) has emerged as the preferred noninvasive procedure over colonoscopy. Current MRI approaches assess rate of contrast enhancement and bowel wall thickness, and rely on extensive manual segmentation for accurate analysis. We propose a supervised learning method for the identification and localization of regions in abdominal magnetic resonance images that have been affected by CD. Low-level features like intensity and texture are used with shape asymmetry information to distinguish between diseased and normal regions. Particular emphasis is laid on a novel entropy-based shape asymmetry method and higher-order statistics like skewness and kurtosis. Multi-scale feature extraction renders the method robust. Experiments on real patient data show that our features achieve a high level of accuracy and perform better than two competing methods.

Automated Detection and Segmentation of Large Lesions in CT Colonography

Computerized tomographic colonography is a minimally invasive technique for the detection of colorectal polyps and carcinoma. Computer-aided diagnosis (CAD) schemes are designed to help radiologists locating colorectal lesions in an efficient and accurate manner. Large lesions are often initially detected as multiple small objects, due to which such lesions may be missed or misclassified by CAD systems. We propose a novel method for automated detection and segmentation of all large lesions, i.e., large polyps as well as carcinoma. Our detection algorithm is incorporated in a classical CAD system. Candidate detection comprises preselection based on a local measure for protrusion and clustering based on geodesic distance. The generated clusters are further segmented and analyzed. The segmentation algorithm is a thresholding operation in which the threshold is adaptively selected. The segmentation provides a size measurement that is used to compute the likelihood of a cluster to be a large lesion. The large lesion detection algorithm was evaluated on data from 35 patients having 41 large lesions (19 of which malignant) confirmed by optical colonoscopy. At five false positive (FP) per scan, the classical system achieved a sensitivity of 78%, while the system augmented with the large lesion detector achieved 83% sensitivity. For malignant lesions, the performance at five FP/scan was increased from 79% to 95%. The good results on malignant lesions demonstrate that the proposed algorithm may provide relevant additional information for the clinical decision process.

Weakly supervised semantic segmentation of Crohn's disease tissues from abdominal MRI

We address the problem of weakly supervised segmentation (WSS) of medical images which is more challenging and has potentially greater applications in the medical imaging community. Training images are labeled only by the classes they contain, and not by the pixel labels. We make use of the Multi Image Model (MIM) for weakly supervised segmentation which exploits superpixel features and assigns labels to every pixel. MIM connects superpixels from all training images in a data driven fashion. Test images are integrated into the MIM for predicting their labels, thus making full use of the training samples. Experimental results on abdominal magnetic resonance (MR) images of patients with Crohns disease show that WSS performs close to fully supervised methods and given sufficient samples can perform on par with fully supervised methods.