3D Regions Of Interest Research Articles

Correlation Between 2- and 3-Dimensional Assessment of Tumor Volume and Vascular Density by Ultrasonography in a Transgenic Mouse Model of Mammary Carcinoma

Visualization and quantification of angiogenesis are instrumental in development of antiangiogenic therapy. Although both 2-dimensional (2D) and 3-dimensional (3D) ultrasonography have been used to monitor tumor growth and vasculature development, the correlation between them has not been sufficiently investigated. We hereby investigated the 2D and 3D sonographic correlation for tumor volume and vascular density confirmed by histologic assessment in the polyoma virus middle T antigen (PyMT) mouse model of mammary carcinoma. Female PyMT mouse tumors were evaluated by ultrasonography in the 2D region of interest (ROI), 3D tumor volume, and 2D and 3D microvascular density after a bolus infusion of a nontargeted contrast-enhanced microbubble agent. Texas Red-dextran was used for quantitative histologic assessment of the tumor microvascular density. The individual 2D tumor ROI area correlated with the 3D tumor volume throughout the 2-week period. However, the extent of the increase in the 3D volume (380%; P < .01; n = 10) was higher than that of the 2D ROI area (72%; P < .01; n = 8-11). A significant and comparable increase in vascular density accessed by both 2D (87%; P < .05; n = 8) and 3D (64%; P < .05; n = 8) imaging was documented. Vascular density obtained through 3D imaging correlated significantly with 2D measurement. These data were confirmed by Texas Red-dextran quantification of vascular density. Conclusions. This study showed a valid application of sonographically based imaging technology in tumor volume and vascular density assessment as well as their 2D and 3D correlation, of which tumor vascular density measured by 2D ultrasonography appeared to be better correlated with the 3D data. Our data indicate that ultrasonography can be applied for real-time, accurate, noninvasive imaging of the tumor volume and vascular density in preclinical models.

Fully automated 3D basal ganglia activity measurement in dopamine transporter scintigraphy (Spectalyzer)

Semiquantitative evaluation of tracer uptake in basal ganglia is superior to visual assessment of images in dopamine transporter (DAT) scintigraphy especially in follow-up of the patients. Manual drawing of regions of interest (ROIs) in two-dimensional (2D) transaxial slices of the single photon emission computed tomography (SPECT) datasets leads to a large inter- and intra-reader variability, while being time consuming. Our aim was to investigate a technique that extracts 3D ROIs in a fully automated fashion and thus might provide reproducible user-independent results allowing better follow-up control and large-scale clinical studies. The highest activity of 123IFP-CIT is expected in the basal ganglia. The proposed method (Spectalyzer) uses the following steps to localize this maximum and extract the ROIs in 3D: (1) Dithers the SPECT volume to obtain a 3D volume with binary only. (2) Models the obtained point distributions as two multivariate Gaussian distributions and estimated their parameters using the expectation maximization algorithm. (3) Using the original SPECT activity values, thresholding is performed using a fixed percentage of maximum activity as a parameter to obtain the 3D ROIs. (4) A reference volume in the occipital region is automatically found based on the location of the two ROIs. (5) From the 3D ROIs, statistical information like mean and median activity and the volume is extracted, relative to the activity in the reference region. The resulting values are compared with values from manual 2D ROIs. Further validation is performed by means of an anthropomorphic striatal phantom. The method was evaluated on 12 SPECT volumes including anthropomorphic striatal phantoms. In all cases the two basal-ganglia were successfully localized and the 3D ROIs estimated, with perfect reproducibility. The obtained values for the mean activity showed the same trend with the values obtained manually and also with the results of the 2D semiautomatic software, but without the substantial inter- and intra-reader variations. The proposed method is successful in finding the 3D ROIs and performing the subsequent measurements automatically. It is proposed as an automatic reproducible approach for semiquantitative analysis of DAT scintigraphy.

CADrx for GBM Brain Tumors: Predicting Treatment Response from Changes in Diffusion-Weighted MRI

The goal of this study was to develop a computer-aided therapeutic response (CADrx) system for early prediction of drug treatment response for glioblastoma multiforme (GBM) brain tumors with diffusion weighted (DW) MR images. In conventional Macdonald assessment, tumor response is assessed nine weeks or more post-treatment. However, we will investigate the ability of DW-MRI to assess response earlier, at five weeks post treatment. The apparent diffusion coefficient (ADC) map, calculated from DW images, has been shown to reveal changes in the tumor’s microenvironment preceding morphologic tumor changes. ADC values in treated brain tumors could theoretically both increase due to the cell kill (and thus reduced cell density) and decrease due to inhibition of edema. In this study, we investigated the effectiveness of features that quantify changes from pre- and post-treatment tumor ADC histograms to detect treatment response. There are three parts to this study: first, tumor regions were segmented on T1w contrast enhanced images by Otsu’s thresholding method, and mapped from T1w images onto ADC images by a 3D region of interest (ROI) mapping tool using DICOM header information; second, ADC histograms of the tumor region were extracted from both pre- and five weeks post-treatment scans, and fitted by a two-component Gaussian mixture model (GMM). The GMM features as well as standard histogram-based features were extracted. Finally, supervised machine learning techniques were applied for classification of responders or non-responders. The approach was evaluated with a dataset of 85 patients with GBM under chemotherapy, in which 39 responded and 46 did not, based on tumor volume reduction. We compared adaBoost, random forest and support vector machine classification algorithms, using ten-fold cross validation, resulting in the best accuracy of 69.41% and the corresponding area under the curve (Az) of 0.70.

FixTag: An algorithm for identifying and tagging fixations to simplify the analysis of data collected by portable eye trackers

Video-based eye trackers produce an output video showing where a subject is looking, the subject's Point-of-Regard (POR), for each frame of a video of the scene. This information can be extremely valuable, but its analysis can be overwhelming. Analysis of eye-tracked data from portable (wearable) eye trackers is especially daunting, as the scene video may be constantly changing, rendering automatic analysis more difficult. A common way to begin analysis of POR data is to group these data into fixations. In a previous article, we compared the fixations identified (i.e., start and end marked) automatically by an algorithm to those identified manually by users (i.e., manual coders). Here, we extend this automatic identification of fixations to tagging each fixation to a Region-of-Interest (ROI). Our fixation tagging algorithm, FixTag, requires the relative 3D positions of the vertices of ROIs and calibration of the scene camera. Fixation tagging is performed by first calculating the camera projection matrices for keyframes of the scene video (captured by the eye tracker) via an iterative structure and motion recovery algorithm. These matrices are then used to project 3D ROI vertices into the keyframes. A POR for each fixation is matched to a point in the closest keyframe, which is then checked against the 2D projected ROI vertices for tagging. Our fixation tags were compared to those produced by three manual coders tagging the automatically identified fixations for two different scenarios. For each scenario, eight ROIs were defined along with the 3D positions of eight calibration points. Therefore, 17 tags were available for each fixation: 8 for ROIs, 8 for calibration points, and 1 for “other.” For the first scenario, a subject was tracked looking through products on four store shelves, resulting in 182 automatically identified fixations. Our automatic tagging algorithm produced tags that matched those produced by at least one manual coder for 181 out of the 182 fixations (99.5% agreement). For the second scenario, a subject was tracked looking at two posters on adjoining walls of a room. Our algorithm matched at least one manual coder's tag for 169 fixations out of 172 automatically identified (98.3% agreement).

Determination of 3D-region of interest using digital in-line holography with astigmatic Gaussian beams

An experimental method to visualize a 3D-region of interest (ROI) by means of an astigmatic Gaussian beam is proposed. This method allows to reduce the amount of image planes to be reconstructed thus saving computational time. ROI determination is performed without any computational step: particles that are located in the ROI can be distinguished from the others according to the hyperbolic shape of their diffraction pattern. Theoretical location of the ROI is determined by using the ABCD approach proposed in a previous paper [Appl. Opt. 47, 4147]. Experimental results are presented.

Self organization of sensor networks for energy-efficient border coverage

Networking together hundreds or thousands of cheap sensor nodes allows users to accurately monitor a remote environment by intelligently combining the data from the individual nodes. As sensor nodes are typically battery operated, it is important to efficiently use the limited energy of the nodes to extend the lifetime of the wireless sensor network (WSN). One of the fundamental issues in WSNs is the coverage problem. In this paper, the border coverage problem in WSNs is rigorously analyzed. Most existing results related to the coverage problem in wireless sensor networks focused on planar networks; however, three dimensional (3D) modeling of the sensor network would reflect more accurately real-life situations. Unlike previous works in this area, we provide distributed algorithms that allow the selection and activation of an optimal border cover for both 2D and 3D regions of interest. We also provide self-healing algorithms as an optimization to our border coverage algorithms which allow the sensor network to adaptively reconfigure and repair itself in order to improve its own performance. Border coverage is crucial for optimizing sensor placement for intrusion detection and a number of other practical applications.

Semi-automated analysis of dynamic changes in myocardial contrast from real-time three-dimensional echocardiographic images as a basis for volumetric quantification of myocardial perfusion

Despite the potential of real-time three-dimensional (3D) echocardiography (RT3DE) to assess myocardial perfusion, there is no quantification method available for perfusion analysis from RT3DE images. Such method would require 3D regions of interest (ROI) to be defined and adjusted frame-by-frame to compensate for cardiac translation and deformation. Our aims were to develop and test a technique for automated identification of 3D myocardial ROI suitable for translation-free quantification of myocardial videointensity over time, MVI(t), from contrast-enhanced RT3DE images. Twelve transthoracic RT3DE (Philips) data sets obtained in pigs during transition from no contrast to steady-state enhancement (Definity) were analysed using custom software. Analysis included: (i) semi-automated detection of left ventricular endo- and epicardial surfaces using level-set techniques in one frame to define a 3D myocardial ROI, (ii) rigid 3D registration to reduce translation and rotation, (iii) elastic 3D registration to compensate for deformation, and (iv) quantification of MVI(t) in the 3D ROI from the registered and non-registered data sets to assess the effectiveness of registration. For each MVI(t) curve we computed % variability during steady-state enhancement (100 x SD/mean) and goodness of fit (r2) to the indicator dilution equation MVI(t) = A[1-exp(-betat)]. Analysis of myocardial contrast throughout contrast inflow was feasible in all data sets. Three-dimensional registration improved MVI(t) curves in terms of both % variability (2.8 +/- 1.8 to 1.5 +/- 0.9%; P < 0.05) and goodness of fit (r2 from 0.79 +/- 0.2 to 0.90 +/- 0.1; P < 0.05). This is the first study to describe a new technique for semi-automated volumetric quantification of myocardial contrast from RT3DE images that includes registration and thus provides the basis for 3D measurement of myocardial perfusion.

Interactive segmentation based on the live wire for 3D CT chest image analysis

The definition of regions of interest (ROIs) such as suspect cancer nodules or lymph nodes in 3D MDCT chest images is often difficult because of the complexity of the phenomena that give rise to them. Manual slice tracing has been used commonly for such problems, but it is extremely time consuming, subject to operator biases, and does not enable reproducible results. Proposed automated 3D image-segmentation methods are generally application dependent, and even the most robust methods have difficulty in defining complex ROIs. The semi-automatic interactive paradigm known as live wire has been proposed by researchers, whereby the human operator interactively defines an ROI’s boundary, guided by an active automated method. We propose 2D and 3D live-wire methods that improve upon previously proposed techniques. The 2D method gives improved robustness and incorporates a search region to improve computational efficiency. The 3D method requires the operator to only consider a few 2D slices, with an automated procedure performing the bulk of the analysis. For tests run with five human operators on both 2D and 3D ROIs in 3D MDCT chest images, the reproducibility was >97% and the ground-truth correspondence was at least 97%. The 2D live-wire approach was ≥14 times faster than manual slice tracing, while the 3D method was ≥28 times faster than manual slice tracing. Finally, we describe a computer-based tool and its application to 3D MDCT-based planning and follow-on live guidance of bronchoscopy. The live-wire methods are efficient, reliable, easy to use, and applicable to a wide range of circumstances.

Lung nodule diagnosis using 3D template matching

In this paper, to utilize the third dimension of Computed Tomography, regions of interest (ROI) slices were combined to form 3D ROI image and a 3D template was determined to find the structures with similar properties of nodules. Convolution of 3D ROI image with the proposed template strengthens the shapes similar to the template and weakens the other ones. False-positive (FP) per nodule and per slice versus diagnosis sensitivity were obtained. The Computer Aided Diagnosis system achieved 100% sensitivity with 0.83 FP per nodule and 0.46 FP per slice, when the nodule thickness was greater than or equal to 5.625 mm.

Application examples of 3D visual interpretation technology in Dagang Oilfield

3D visualization technology is a new seismic interpretation technology that is completely different from traditional seismic data interpretation technologies. This method takes geologic bodies or any 3D region of interest as a unit and integrates the points, lines, and planes of the bodies to make a 3D visual interpretation. This technology can be integrated with the coherence technique to interpret complex faults. Spatial domain automatic horizon tracing can accurately determine structural features. Target sculpting can determine the spatial dimensions and depth extent of perspective 3D lithology bodies using inversion data. I applied this technology to the interpretation of the complex block structures in the Guanxi area of Dagang Oilfield and the prediction of subtle reservoirs in the western slope of the Qinan sag with good results. This study indicated that the visual interpretation technology can be widely and effectively applied in oil field development and to improve the exploration success ratio.

Blood Oxygen Level–Dependent MRI Allows Metabolic Description of Tissue at Risk in Acute Stroke Patients

The delineation of the "penumbra" is of particular interest in acute stroke imaging. The "mismatch concept" applying perfusion-weighted imaging (PWI) and diffusion-weighted imaging (DWI) appears to be an oversimplification of the underlying electrophysiological tissue status. An additional parameter reflecting the metabolic state of the threatened brain tissue would improve our ability to describe the penumbra. One candidate is deoxyhemoglobin (deoxy-Hb) as an indicator of the oxygen extraction fraction that can be visualized by T2*-based blood oxygen level-dependent (BOLD) imaging. We analyzed data from 32 patients with acute stroke in the territory of the middle cerebral artery. MRI included fluid-attenuated inversion recovery, DWI, PWI, time-of-flight angiography, and quantitative T2 and T2* (qT2, qT2*) imaging. Follow-up was performed on day 1 and days 5 to 8. We calculated 1/T2'=1/qT2*-1/qT2. Changes of T2', representing the deoxy-Hb effect, were analyzed by 3D regions of interest (ROIs): apparent diffusion coefficient lesion day 0 (L0), time-to-peak-lesion day 0 (T0), final infarct size days 5 to 8 (F5-8), lesion growth (LG; F5-8-L0), and surviving tissue (ST; T0-F5-8). We observed a clear decrease of T2' in the infarcted hemisphere compared with the unaffected control ROIs. The mean value showed the most pronounced loss of T2' signal intensity in L0 (-15.7%), followed by LG (-10.5%) and ST (-8.0%). The implementation of BOLD imaging in acute stroke MRI offers a noninvasive estimation of the O2 utilization and is able to add additional information concerning the present metabolic state of the threatened brain tissue. The changes in T2' intensity are visually noticeable in the reconstructed T2' images and provide a better estimation of the real penumbra.

Cruciate ligament localization in T1- and T2- weighted MR knee images

Abstract This article shows a new method of the cruciate ligament localization in T1-and T2-weighted MR knee images. Cruciate ligament localization allows to reduce number of sliccs used in 3D rendering. The analysis is performed on T1-and T2-wcightcd MR knee images. Each slice is subjected to two-dimensional entropy and energy measurements of fuzziness. Then, a fuzzy c-means clustering technique modified by introducing additionally to the objective function a median estimator is implemented. The class reflecting soft tissue is subjected to similarity measures which compare a local image intensity. A detailed local analysis, based on image profile models, yields a 3D region of interest (ROI) which includes the posterior and anterior cruciate ligament. Finally, fuzzy techniques applied to the ROI histogram analysis enhances the contrast and views the 3D region.

Projection Extension for Region Of Interest Imaging in Cone-Beam CT 1

For 3D X-ray imaging during interventions, changes of the imaged object are often restricted to a small part of the field of view, suggesting region of interest (ROI) imaging by irradiating this area only. In this article, we present a novel method for extension of truncated projections in order to avoid truncation artifacts in C-arm based 3D ROI imaging. The method makes use of prior knowledge by combining forward projections of a previously acquired, nontruncated 3D reference image with the truncated ROI projections. Rigid registration between the two datasets is achieved by using a technique based on local cross-correlation. To account for a gray value mismatch between the two data sets due to, e.g., differing beam quality and different contributions of scattered radiation, a linear gray level transformation is applied to the forward-projected reference data. The performance of different gray value transformation schemes is systematically assessed by means of numerical simulations. For various simulated scenarios, the best performing transformation has been identified, providing practical guidelines for selecting a scheme depending on the origin of the gray-level mismatch. Experiments prove the high performance of the developed method. The presented technique enables almost artifact-free 3D ROI imaging during interventions. This actually allows for repeated scans at low dose and enables intraprocedural imaging of large objects even with a small detector. However, applicability of the method is limited to scenarios where direct access to a reference image, e.g., a prior CT scan, is available.

Localization of anatomical point landmarks in 3D medical images by fitting 3D parametric intensity models

We introduce a new approach for the localization of 3D anatomical point landmarks. This approach is based on 3D parametric intensity models which are directly fitted to 3D images. To efficiently model tip-like, saddle-like, and sphere-like anatomical structures we introduce analytic intensity models based on the Gaussian error function in conjunction with 3D rigid transformations as well as deformations. To select a suitable size of the region-of-interest (ROI) where model fitting is performed, we also propose a new scheme for automatic selection of an optimal 3D ROI size based on the dominant gradient direction. In addition, to achieve a higher level of automation we present an algorithm for automatic initialization of the model parameters. Our approach has been successfully applied to accurately localize anatomical landmarks in 3D synthetic data as well as 3D MR and 3D CT image data. We have also compared the experimental results with the results of a previously proposed 3D differential approach. It turns out that the new approach significantly improves the localization accuracy.

Image reconstruction in peripheral and central regions‐of‐interest and data redundancy

Algorithms have been developed for image reconstruction within a region-of-interest (ROI) from fan-beam data less than that required for reconstructing the entire image. However, these algorithms do not admit truncated data. In this work, we investigate exact ROI-image reconstruction from fan-beam data containing truncations by use of the so-called fan-beam backprojection-filtration (BPF) algorithm. We also generalize the fan-beam BPF algorithm to exploit redundant information inherent in the truncated fan-beam data. Because the parallel-beam scan can be interpreted as a special case of the fan-beam scan, based upon the fan-beam BPF algorithm, we derive a parallel-beam BPF algorithm for exactly reconstructing ROI images from truncated parallel-beam data. Furthermore, we investigate image reconstruction within two types of distinctive ROIs, which are referred to as the peripheral and central ROIs, respectively, from fan-beam data containing truncations and discuss their potential clinical applications. The results can readily be generalized to reconstructing 3D ROI images from data acquired in circular and helical cone-beam scan. They can also be extended to address ROI-image-reconstruction problems in parallel-, fan-, and cone-beam scans with general trajectories. The work not only has significant implications for clinical and animal-imaging applications of CT, but also may find applications in other imaging modalities.

Evaluation of hypoxia in an experimental rat tumour model by [18F]Fluoromisonidazole PET and immunohistochemistry

This study aimed to evaluate tumour hypoxia by comparing [18F]Fluoromisonidazole uptake measured using positron emission tomography ([18F]FMISO-PET) with immunohistochemical (IHC) staining techniques. Syngeneic rhabdomyosarcoma (R1) tumour pieces were transplanted subcutaneously in the flanks of WAG/Rij rats. Tumours were analysed at volumes between 0.9 and 7.3 cm3. Hypoxic volumes were defined using a 3D region of interest on 2 h postinjection [18F]FMISO-PET images, applying different thresholds (1.2–3.0). Monoclonal antibodies to pimonidazole (PIMO) and carbonic anhydrase IX (CA IX), exogenous and endogenous markers of hypoxia, respectively, were used for IHC staining. Marker-positive fractions were microscopically measured for each tumour, and hypoxic volumes were calculated. A heterogeneous distribution of hypoxia was observed both with histology and [18F]FMISO autoradiography. A statistically significant correlation (P<0.05) was obtained between the hypoxic volumes defined with [18F]FMISO-PET and the volumes derived from the PIMO-stained tumour sections (r=0.9066; P=0.0001), regardless of the selected threshold between 1.4 and 2.2. A similar observation was made with the CA IX staining (r=0.8636; P=0.0006). The relationship found between [18F]FMISO-PET and PIMO- and additionally CA IX-derived hypoxic volumes in rat rhabdomyosarcomas indicates the value of the noninvasive imaging method to measure hypoxia in whole tumours.

3D ROI imaging for cone-beam computed tomography

C-arm X-ray systems can be used to acquire rotational cone-beam projections during interventions, from which CT-like 3D images can be reconstructed. For intraprocedural 3D imaging, changes of the imaged object are often restricted to a certain part of the possible field of view, suggesting region of interest (ROI) imaging by irradiating this area only. Because projections acquired for ROI imaging do not contain the entire object, reconstructed images suffer from truncation-caused artifacts that cannot be fully compensated by commonly used methods. In this paper, we present a novel method that enables low-dose ROI imaging in cone-beam computed tomography by accurately extending truncated projections, in that way, avoiding truncation artifacts. For this purpose, a previously acquired 3D image covering the whole object is combined with the truncated projections of the ROI acquired during the intervention.

A knowledge-driven algorithm for a rapid and automatic extraction of the human cerebral ventricular system from MR neuroimages

A knowledge-driven algorithm for a rapid, robust, accurate, and automatic extraction of the human cerebral ventricular system from MR neuroimages is proposed. Its novelty is in combination of neuroanatomy, radiological properties, and variability of the ventricular system with image processing techniques. The ventricular system is divided into six 3D regions: bodies and inferior horns of the lateral ventricles, third ventricle, and fourth ventricle. Within each ventricular region, a 2D region of interest (ROI) is defined based on anatomy and variability. Each ventricular region is further subdivided into subregions, and conditions detecting and preventing leakage into the extra-ventricular space are specified for each subregion. The algorithm extracts the ventricular system by (1) processing each ROI (to calculate its local statistics, determine local intensity ranges of cerebrospinal fluid and gray and white matters, set a seed point within the ROI, grow region directionally in 3D, check anti-leakage conditions, and correct growing if leakage occurred) and (2) connecting all unconnected regions grown by relaxing growing conditions. The algorithm was validated qualitatively on 68 and quantitatively on 38 MRI normal and pathological cases (30 clinical, 20 MGH Brain Repository, and 18 MNI BrainWeb data sets). It runs successfully for normal and pathological cases provided that the slice thickness is less than 3.0 mm in axial and less than 2.0 mm in coronal directions, and the data do not have a high inter-slice intensity variability. The algorithm also works satisfactorily in the presence of up to 9% noise and up to 40% RF inhomogeneity for the BrainWeb data. The running time is less than 5 s on a Pentium 4, 2.0 GHz PC. The best overlap metric between the results of a radiology expert and the algorithm is 0.9879 and the worst 0.9527; the mean and standard deviation of the overlap metric are 0.9723 and 0.01087, respectively.

Band-selective spin echoes for in vivo localized 1H NMR spectroscopy.

This study describes a new single spin-echo spatial localization sequence, BASSALE or BAnd-Selective Spin echo Acquisition for Localized Editing, that overcomes a number of the limitations of the STEAM and PRESS volume selection pulse sequences. It achieves conformal volume localization in a single shot by spatially tailored suppression of all magnetization outside a 2D region of interest followed by selection of a single orthogonal slice. This separation of spatial localization from the echo formation process has permitted use of a spectrally selective cosine-modulated sinc refocusing pulse to acquire localized 1H spectra with the water suppression efficiency of STEAM and the sensitivity of PRESS. Echoes formed by such spectrally selective pulses have been termed band-selective spin echoes. The BASSALE sequence attains shorter echo times than PRESS, inhibits scalar spin-spin interactions to permit localized editing and T2 relaxometry of metabolites with J-coupled spins (e.g., lactate), is insensitive to homonuclear multiple-quantum and polarization transfer effects, and can be made sensitive or insensitive to spin displacement effects. Applications are shown both with phantoms and in situ in the rat brain.