Automated detection of cerebral microbleeds in patients with Traumatic Brain Injury.

In this paper a Computer Aided Detection (CAD) system is presented to automatically detect Cerebral Microbleeds (CMBs) in patients with Traumatic Brain Injury (TBI). It is believed that the presence of CMBs has clinical prognostic value in TBI patients. To study the contribution of CMBs in patient outcome, accurate detection of CMBs is required. Manual detection of CMBs in TBI patients is a time consuming task that is prone to errors, because CMBs are easily overlooked and are difficult to distinguish from blood vessels. This study included 33 TBI patients. Because of the laborious nature of manually annotating CMBs, only one trained expert manually annotated the CMBs in all 33 patients. A subset of ten TBI patients was annotated by six experts. Our CAD system makes use of both Susceptibility Weighted Imaging (SWI) and T1 weighted magnetic resonance images to detect CMBs. After pre-processing these images, a two-step approach was used for automated detection of CMBs. In the first step, each voxel was characterized by twelve features based on the dark and spherical nature of CMBs and a random forest classifier was used to identify CMB candidate locations. In the second step, segmentations were made from each identified candidate location. Subsequently an object-based classifier was used to remove false positive detections of the voxel classifier, by considering seven object-based features that discriminate between spherical objects (CMBs) and elongated objects (blood vessels). A guided user interface was designed for fast evaluation of the CAD system result. During this process, an expert checked each CMB detected by the CAD system. A Fleiss' kappa value of only 0.24 showed that the inter-observer variability for the TBI patients in this study was very large. An expert using the guided user interface reached an average sensitivity of 93%, which was significantly higher (p = 0.03) than the average sensitivity of 77% (sd 12.4%) that the six experts manually detected. Furthermore, with the use of this CAD system the reading time was substantially reduced from one hour to 13 minutes per patient, because the CAD system only detects on average 25.9 false positives per TBI patient, resulting in 0.29 false positives per definite CMB finding.

Optimal short-time acquisition schemes in high angular resolution diffusion-weighted imaging.

This work investigates the possibilities of applying high-angular-resolution-diffusion-imaging- (HARDI-) based methods in a clinical setting by investigating the performance of non-Gaussian diffusion probability density function (PDF) estimation for a range of b-values and diffusion gradient direction tables. It does so at realistic SNR levels achievable in limited time on a high-performance 3T system for the whole human brain in vivo. We use both computational simulations and in vivo brain scans to quantify the angular resolution of two selected reconstruction methods: Q-ball imaging and the diffusion orientation transform. We propose a new analytical solution to the ODF derived from the DOT. Both techniques are analytical decomposition approaches that require identical acquisition and modest postprocessing times and, given the proposed modifications of the DOT, can be analyzed in a similar fashion. We find that an optimal HARDI protocol given a stringent time constraint (<10 min) combines a moderate b-value (around 2000 s/mm(2)) with a relatively low number of acquired directions (>48). Our findings generalize to other methods and additional improvements in MR acquisition techniques.

The use of a reference tissue arterial input function with low-temporal-resolution DCE-MRI data.

Pharmacokinetic modeling is a promising quantitative analysis technique for cancer diagnosis. However, diagnostic dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) of the breast is commonly performed with low temporal resolution. This limits its clinical utility. We investigated for a range of temporal resolutions whether pharmacokinetic parameter estimation is impacted by the use of data-derived arterial input functions (AIFs), obtained via analysis of dynamic data from a reference tissue, as opposed to the use of a standard AIF, often obtained from the literature. We hypothesized that the first method allows the use of data at lower temporal resolutions than the second method. Test data were obtained by downsampling high-temporal-resolution rodent data via a k-space-based strategy. To fit the basic Tofts model, either the data-derived or the standard AIF was used. The resulting estimates of K(trans) and v(e) were compared with the standard estimates obtained by using the original data. The deviations in K(trans) and v(e), introduced when lowering temporal resolution, were more modest using data-derived AIFs compared with using a standard AIF. Specifically, lowering the resolution from 5 to 60 s, the respective changes in K(trans) were 2% (non-significant) and 18% (significant). Extracting the AIF from a reference tissue enables accurate pharmacokinetic parameter estimation for low-temporal-resolution data.

Optimal acquisition schemes in high angular resolution diffusion weighted imaging.

The recent challenge in diffusion imaging is to find acquisition schemes and analysis approaches that can represent non-gaussian diffusion profiles in a clinically feasible measurement time. In this work we investigate the effect of b-value and the number of gradient vector directions on Q-ball imaging and the Diffusion Orientation Transform (DOT) in a structured way using computational simulations, hardware crossing-fiber diffusion phantoms, and in-vivo brain scans. We observe that DOT is more robust to noise and independent of the b-value and number of gradients, whereas Q-ball dramatically improves the results for higher b-values and number of gradients and at recovering larger angles of crossing. We also show that Laplace-Beltrami regularization has wide applicability and generally improves the properties of DOT. Knowledge of optimal acquisition schemes for HARDI can improve the utility of diffusion weighted MR imaging in the clinical setting for the diagnosis of white matter diseases and presurgical planning.

A novel 3D multi-scale lineness filter for vessel detection.

The branching pattern and geometry of coronary microvessels are of high interest to understand and model the blood flow distribution and the processes of contrast invasion, ischemic changes and repair in the heart in detail. Analysis is performed on high resolution, 3D volumes of the arterial microvasculature of entire goat hearts, which are acquired with an imaging cryomicrotome. Multi-scale vessel detection is an important step required for a detailed quantitative analysis of the coronary microvasculature. Based on visual inspection, the derived lineness filter shows promising results on real data and digital phantoms, on the way towards accurate computerized reconstructions of entire coronary trees. The novel lineness filter exploits the local first and second order multi-scale derivatives in order to give an intensity-independent response to line centers and to suppress unwanted responses to steep edges.

Quantification of atherosclerotic plaque components using in vivo MRI and supervised classifiers.

In this work we aimed to study the possibility of using supervised classifiers to quantify the main components of carotid atherosclerotic plaque in vivo on the basis of multisequence MRI data. MRI data consisting of five MR weightings were obtained from 25 symptomatic subjects. Histological micrographs of endarterectomy specimens from the 25 carotids were used as a standard of reference for training and evaluation. The set of subjects was divided in a training set (12 subjects) and an evaluation set (13 subjects). Four different classifiers and two human MRI readers determined the percentages of calcified tissue, fibrous tissue, lipid core, and intraplaque hemorrhage on the subject level for all subjects in the evaluation set. Quantification of the relatively small amounts of calcium could not be done with statistical significance by either the classifiers or the MRI readers. For the other tissues a simple Bayesian classifier (Bayes) performed better than the other classifiers and the MRI readers. All classifiers performed better than the MRI readers in quantifying the sum of hemorrhage and lipid proportions. The MRI readers overestimated the hemorrhage proportions and tended to underestimate the lipid proportions. In conclusion, this pilot study demonstrates the benefits of algorithmic classifiers for quantifying plaque components.

Computer-aided diagnosis in chest radiography: a survey.

The traditional chest radiograph is still ubiquitous in clinical practice, and will likely remain so for quite some time. Yet, its interpretation is notoriously difficult. This explains the continued interest in computer-aided diagnosis for chest radiography. The purpose of this survey is to categorize and briefly review the literature on computer analysis of chest images, which comprises over 150 papers published in the last 30 years. Remaining challenges are indicated and some directions for future research are given.

Automatic segmentation of lung fields in chest radiographs.

The delineation of important structures in chest radiographs is an essential preprocessing step in order to automatically analyze these images, e.g., for tuberculosis screening support or in computer assisted diagnosis. We present algorithms for the automatic segmentation of lung fields in chest radiographs. We compare several segmentation techniques: a matching approach; pixel classifiers based on several combinations of features; a new rule-based scheme that detects lung contours using a general framework for the detection of oriented edges and ridges in images; and a hybrid scheme. Each approach is discussed and the performance of nine systems is compared with interobserver variability and results available from the literature. The best performance is obtained by the hybrid scheme that combines the rule-based segmentation algorithm with a pixel classification approach. The combinations of two complementary techniques leads to robust performance; the accuracy is above 94% for all 115 images in the test set. The average accuracy of the scheme is 0.969 +/- 0.0080, which is close to the interobserver variability of 0.984 +/- 0.0048. The methods are fast, and implemented on a standard PC platform.

Geodesic deformable models for medical image analysis.

In this paper implicit representations of deformable models for medical image enhancement and segmentation are considered. The advantage of implicit models over classical explicit models is that their topology can be naturally adapted to objects in the scene. A geodesic formulation of implicit deformable models is especially attractive since it has the energy minimizing properties of classical models. The aim of this paper is twofold. First, a modification to the customary geodesic deformable model approach is introduced by considering all the level sets in the image as energy minimizing contours. This approach is used to segment multiple objects simultaneously and for enhancing and segmenting cardiac computed tomography (CT) and magnetic resonance images. Second, the approach is used to effectively compare implicit and explicit models for specific tasks. This shows the complementary character of implicit models since in case of poor contrast boundaries or gaps in boundaries e.g. due to partial volume effects, noise, or motion artifacts, they do not perform well, since the approach is completely data-driven.

Advances in three-dimensional diagnostic radiology.

The maturity of current 3D rendering software in combination with recent developments in computer vision techniques enable an exciting range of applications for the visualisation, measurement and interactive manipulation of volumetric data, relevant both for diagnostic imaging and for anatomy. This paper reviews recent work in this area from the Image Sciences Institute at Utrecht University. The processes that yield a useful visual presentation are sequential. After acquisition and before any visualisation, an essential step is to prepare the data properly: this field is known as 'image processing' or 'computer vision' in analogy with the processing in human vision. Examples will be discussed of modern image enhancement and denoising techniques, and the complex process of automatically finding the objects or regions of interest, i.e. segmentation. One of the newer and promising methodologies for image analysis is based on a mathematical analysis of the human (cortical) visual processing: multiscale image analysis. After preprocessing the 3D rendering can be acquired by simulating the 'ray casting' in the computer. New possibilities are presented, such as the integrated visualisation in one image of (accurately registered) datasets of the same patient acquired in different modality scanners. Other examples include colour coding of functional data such as SPECT brain perfusion or functional magnetic resonance (MR) data and even metric data such as skull thickness on the rendered 3D anatomy from MR or computed tomography (CT). Optimal use and perception of 3D visualisation in radiology requires fast display and truly interactive manipulation facilities. Modern and increasingly cheaper workstations ( < $10000) allow this to be a reality. It is now possible to manipulate 3D images of 256 at 15 frames per second interactively, placing virtual reality within reach. The possibilities of modern workstations become increasingly more sophisticated and versatile. Examples presented include the automatic detection of the optimal viewing angle of the neck of aneurysms and the simulation of the design and placement procedure of intra-abdominal aortic stents. Such developments, together with the availability of high-resolution datasets of modern scanners and data such as from the NIH Visible Human project, have a dramatic impact on interactive 3D anatomical atlases.

Intravascular ultrasound image subtraction: a contrast enhancing technique to facilitate automatic three-dimensional visualization of the arterial lumen.

At 30 MHz, the intravascular ultrasound backscatter of blood confounds the discrimination of the lumen from the arterial wall. This study validates a subtraction method which creates a still-frame image with a sharp demarcation of the lumen. The method involves subtraction of consecutive images and 2D ensemble averaging of the absolute pixel values. Subtraction exploits the dynamic properties of flowing red blood cells. Three phantom arteries were used, with erythrocytes in their lumens and wall. For this reason, it was not possible, in one single original image, to discriminate the blood in the lumen from the phantom wall. Based on 26 consecutive original images, in the mean subtraction image contrast between lumen and phantom wall grey values increased eightfold from 10.9 (5.3-19.2) (mean and range) in the original image to 87.7 (73.6-107.0) (P < 0.001). A sufficiently large contrast increase to allow automatic segmentation was obtained by using five original images (0.3-s acquisition time) for any single mean subtraction image. Low blood flow velocities (down to 0.5 cm/s) did not alter this result. Automatic segmentation of the lumen allowed fast 3D reconstruction of the lumen in all three phantom arteries. In phantom arteries, the intravascular ultrasound image subtraction technique improved contrast between lumen and wall which enabled automated lumen segmentation and fast 3D visualization of both the lumen and defects in the wall.

Discrimination of the intravascular lumen and dissections in a single 30-MHz US image: use of "confounding" blood backscatter to advantage.

At 30-MHz intravascular ultrasound, the high-intensity backscatter of blood confounds discrimination of the intravascular lumen from the arterial wall. The authors developed and used a subtraction method to visualize the lumen and dissections through cancellation of backscatter from static areas. In two patients who underwent angioplasty of the superficial femoral artery, the residual hyperechoic pattern created by moving blood allowed good differentiation between the wall and lumen or dissections.

Implementation aspects of image management, archiving, and communication systems in routine clinical use.

Implementation of a digital imaging network in routine clinical use is a difficult task. Not only the high technical requirements, but especially the complexity of the organization of the diagnostic information flow in a hospital makes commitment essential in PACS implementation. The application of project management with a dedicated project team is a good approach to establish this complex endeavor. The different phases of a project for implementation are initiative, definition, design, preparation, realization, and evaluation and are exemplified by what we learned from the Dutch PACS project.

Bone growth and remodeling after distraction epiphysiolysis of the proximal tibia of the rabbit. Effect of electromagnetic stimulation.

The effect of pulsed electromagnetic stimulation on bone formation was tested in a lower-limb-lengthening model in the rabbit. Limb lengthening was performed by distraction epiphysiolysis. A specially designed external distraction device allowed 10 mm of lengthening of the tibia. Coils to generate a pulsed electromagnetic field were clipped onto the distractor. Stimulation started after a distraction period of three weeks and was continuous for 18 weeks. A control group received the same treatment without stimulation. Bone formation in the elongated zone was evaluated by computed tomography, scintigraphy, and histology. Bone healing involved accretion of callus followed by a process of remodeling, resulting in the formation of a solid cortex. The formation of a diaphysislike structure at the original site of the metaphysis progressed from the distal end of the elongated zone upward. Electromagnetic stimulation had no effect on the rate or extent of bone formation and remodeling.

Fundamental properties of medical image perception.

With a mind toward the effective acquisition, processing, presentation, and reading of radiological images, a survey of how the human visual system perceives images is presented here. The level is chosen to be suitable for the radiologist, and the relative emphasis on the various visual cues of luminance, color, form, texture, motion, and depth is chosen based on their importance with radiological images. Examples of the radiological relevance of the various visual properties are given. We cover first what the visual system's behavior is and then survey some of the properties of the physiological mechanisms that provide this behavior.

First experiences with the modelling and simulation package MIRACLES applied to a picture archiving and communication system (PACS) in a clinical environment.

Since the construction of picture archiving and communication systems (PACS) appears to be extremely difficult, computer modelling and simulation are used as decision support tools. The package MIRACLES (Medical Image Representation, Archiving and Communication Learned from Extensive Simulation) has been developed at BAZIS in order to support the construction of simulation models of image information systems. This article discusses the application of MIRACLES to a prototypical PACS as being installed in a clinical environment. Attention is focussed to the required system analysis and difficulties which arose during the construction of the simulation model. The emphasis is on the presentation of the results of the simulation study, which show that simulation can be fruitfully used to predict, to analyse and to assist in solving performance problems. The simulation study confirmed assumptions and suppositions concerning both the system performance itself and strategies to improve the performance. The study also resulted in a number of concrete recommendations which might be useful for the set-up of the prototypical PACS.

Bone healing during lower limb lengthening by distraction epiphysiolysis.

A study of limb lengthening by distraction epiphysiolysis in the rabbit tibia is presented. A special external distraction device was developed that allowed 10 mm lengthening of the leg. Bone formation in the elongated zone was studied by computed tomography and [99mTc] methylene diphosphonate (MDP) scintigraphy. Computed tomography showed bone formation proceeding for several weeks after the end of the distraction period, followed by a decrease in the amount of bone during a remodeling phase leading to the formation of a solid cortical structure. The uptake of [99mTc]MDP increased parallel to, but preceeding the actual accretion of bone, followed by a decrease during the bone remodeling phase. Uptake of the tracer will partly reflect bone metabolism, but other factors, like trauma, determine much of the uptake.