Automatic model-based contour detection and blood flow quantification in small vessels with velocity encoded magnetic resonance imaging

RATIONALE AND OBJECTIVES: The quantitative assessment of blood flow in peripheral vessels from phase-contrast magnetic resonance imaging studies requires the accurate delineation of vessel contours in cross-sectional magnetic resonance images. The conventional manual segmentation approach is tedious, time-consuming, and leads to significant inter- and intraobserver variabilities. The aim of this study was to verify whether automatic model-based segmentation decreases these problems by fitting a model to the actual blood velocity profile. METHODS: In this study 2 new fully automatic methods (a static and a dynamic approach) were developed and compared with manual analyzes using phantom and in vivo studies of internal carotid and vertebral arteries in healthy volunteers. The automatic segmentation approaches were based on fitting a 3D parabolic velocity model to the actual velocity profiles. In the static method, the velocity profiles were averaged over the complete cardiac cycle, whereas the dynamic method takes into account the velocity data of each cardiac time bin individually. Materials consisted of the magnetic resonance imaging data from 3 straight phantom tubes and the blood velocity profiles of 8 volunteers. RESULTS: For the phantom studies, the automatic dynamic approach performed significantly better than the manual analysis (intraclass correlations [ICC] of 0.62-0.98 and 0.30-0.86, respectively). For the assessment of the total cerebral blood flow in the in vivo studies, the automatic static method performed significantly better than the manual 1 (ICC of 0.98-0.98 and 0.93-0.95, respectively). However, the automatic dynamic method was not significantly better than the manual 1 (ICC = 0.92-0.96) but had the advantage of providing additional parameters. CONCLUSION: Blood flow in magnetic resonance images of small vessels can be assessed accurately, rapidly, and fully automatically using model-based postprocessing by fitting a first approximation of the velocity profile to the actual flow data.     

Model-based registration for dynamic cardiac perfusion MRI

PURPOSE: To assess the accuracy of a model-based approach for registration of myocardial dynamic contrast-enhanced (DCE)-MRI corrupted by respiratory motion. MATERIALS AND METHODS: Ten patients were scanned for myocardial perfusion on 3T or 1.5T scanners, and short- and long-axis slices were acquired. Interframe registration was done using an iterative model-based method in conjunction with a mean square difference metric. The method was tested by comparing the absolute motion before and after registration, as determined from manually registered images. Regional flow indices of myocardium calculated from the manually registered data were compared with those obtained with the model-based registration technique. RESULTS: The mean absolute motion of the heart for the short-axis data sets over all the time frames decreased from 5.3+/-5.2 mm (3.3+/-3.1 pixels) to 0.8+/-1.3 mm (0.5+/-0.7 pixels) in the vertical direction, and from 3.0+/-3.7 mm (1.7+/-2.1 pixels) to 0.9+/-1.2 mm (0.5+/-0.7 pixels) in the horizontal direction. A mean absolute improvement of 77% over all the data sets was observed in the estimation of the regional perfusion flow indices of the tissue as compared to those obtained from manual registration. Similar results were obtained with two-chamber-view long-axis data sets. CONCLUSION: The model-based registration method for DCE cardiac data is comparable to manual registration and offers a unique registration method that reduces errors in the quantification of myocardial perfusion parameters as compared to those obtained from manual registration.     

Cine magnetic resonance microscopy of the rat heart using cardiorespiratory-synchronous projection reconstruction

PURPOSE: To tailor a cardiac magnetic resonance (MR) microscopy technique for the rat that combines improvements in pulse sequence design and physiologic control to acquire high-resolution images of cardiac structure and function. MATERIALS AND METHODS: Projection reconstruction (PR) was compared to conventional Cartesian techniques in point-spread function simulations and experimental studies to evaluate its artifact sensitivity. Female Sprague-Dawley rats were imaged at 2.0 T using PR with direct encoding of the free induction decay. Specialized physiologic support and monitoring equipment ensured consistency of biological motion and permitted synchronization of imaging with the cardiac and respiratory cycles. RESULTS: The reduced artifact sensitivity of PR offered improved delineation of cardiac and pulmonary structures. Ventilatory synchronization further increased the signal-to-noise ratio by reducing inter-view variability. High-quality short-axis and long-axis cine images of the rat heart were acquired with 10-msec temporal resolution and microscopic spatial resolution down to 175 microm x 175 microm x 1 mm. CONCLUSION: Integrating careful biological control with an optimized pulse sequence significantly limits both the source and impact of image artifacts. This work represents a novel integration of techniques designed to support measurement of cardiac morphology and function in rodent models of cardiovascular disease.     

Magnetic resonance microscopic images with 50-mm field-of-view of the medial aspect of the knee

Purpose: To demonstrate the utility of microscopic images with field-of-view of 50 mm in delineation of the medial aspect of the knee, including fascial plane, superficial and deep layers of the medial collateral ligament (MCL), and the medial meniscus. Material and Methods: Using a phantom, the signal-to-noise ratio (SNR) of a magnetic resonance (MR) microscopy coil with a diameter of 47 mm was calculated and compared with that of a regular coil. Four cadaveric knees were imaged by microscopy and resected to confirm the morphologies. Sixty-nine patients with internal derangement were examined by routine and microscopic imaging. Comparing the paired images for delineation of the above-mentioned structures, a qualitative image analysis was performed. Results: SNRs of the MR microscopy coil were higher than those of the regular coil. MR microscopy readily demonstrated the multilayered appearance of the fascial plane and both layers of the MCL in cadavers and patients. In cases with MCL tears, ruptured stumps were identified by microscopy. MR microscopy delineated tiny cleavages in cases with meniscal tears. The mean values of qualitative evaluation of the MR microscopy were significantly higher than those of the routine imaging. Conclusion: High-resolution imaging delineated fine structures of the medial aspect of the knee.     

Evaluation of manual vs semi-automated delineation of liver lesions on CT images

In this paper we compare a semi-automated delineation method with totally manual delineation for area quantification, with respect to efficiency, quality, and intra- and interobserver variability. Liver lesions on 28 CT images were delineated by three observers, twice using completely manual delineation and twice using a semi-automated method. Quantitative comparisons were performed with respect to delineated area and time required for the delineation tasks. Subjective comparisons were performed with respect to efficiency and perceived quality of the semi-automated method. The areas obtained using semi-automated delineation were significantly smaller (11 %) than those obtained using totally manual delineation. Intraobserver and interobserver variability with the semi-automated method were approximately three times lower than with manual delineation. Efficiency of the semi-automated method was subjectively rated favorable, although further improvements are possible. With respect to quality, the semi-automated method was ranked better than the manual method in 73 % of cases. Received 10 January 1996; Revision received 15 May 1996; Accepted 19 July 1996     

Automated registration of dynamic MR images for the quantification of myocardial perfusion

Cardiac dynamic magnetic resonance imaging (MRI) after contrast media injection suffers from motion induced by free breathing during acquisition. This work presents an automated approach for motion correction of the heart. The registration is based on the multipass/multiresolution iterative minimizing of intrinsic differences between each image and a reference image coupled to a two-dimensional/3 parameters rigid body correction. The efficiency of this correction method was evaluated with anatomical landmarks, various cost functions, and for a compartment model fit of the data with 2 parameters: K1, the blood to myocardium transfer coefficient; and Vd, the distribution volume of the contrast media. The variability of K1 and Vd, derived from the fit of the registered images (using the manual correction as a gold standard), was significantly reduced by comparison with the variability obtained from the uncorrected images (P < 0.04). This motion correction method also clearly improves the analysis of dynamic cardiac MRI after contrast media injection in comparison to manual correction.     

Improved image registration by using fourier interpolation

This paper presents a technique for performing two-dimensional rigid-body image registration for functional magnetic resonance images (fMRI). The method provides accurate motion correction without local distortion. The approach is to perform the translation and rotation in the Fourier domain. For images sampled on a grid, such as in echo-planar imaging (EPI), one potential stumbling block to this approach is the computational burden of reconstruction, since the rotated image will no longer be on the Cartesian grid. A method of approximating rotations via local translations (shearing) is presented, which keeps the data on the Cartesian grid. This can provide quite accurate approximations with only a moderate amount of computation. A mean squared error (MSE) criterion is used for determining the registration parameters. This method is tested on several sets of simulated images and shown to have an accuracy ranging from 0.02 to 0.3 pixels for images with SNRs ranging from 100 to 10, respectively. They techniques have been tested on several sets of images. They are shown to work well on real subjects, for both echo-planar and spiral data acquisition schemes. The techniques are used in an activation study in which the subject moved his head during image collection. After use of this registration technique, the activation is easily detected.     

基于VoxelMorph无监督缺失图像配准的无标记射束方向观肿瘤跟踪算法

目的 基于机器学习提出可应用于低图像质量、多叶准直器(MLC)遮挡和非刚性变形兆伏级(MV)图像的无标记射束方向观(BEV)肿瘤放疗跟踪算法。方法 采用窗口模板匹配法和Voxelmorph端到端无监督网络,处理MV图像中的配准问题。使用动态胸部模体,验证肿瘤跟踪算法的准确性。将模体质量保证(QA)计划在加速器上手动设置治疗偏移后执行,收集治疗过程中的682幅电子射野影像系统(EPID)图像作为固定图像;同时采集计划系统中对应射野角度的数字影像重建(DRR)图作为浮动图像,进行靶区跟踪研究。收集21例肺部肿瘤放疗的533对EPID和DRR图像进行肿瘤跟踪研究,提供治疗过程中肿瘤位置变化定量结果。图像相似度用于算法的第三方验证。结果 算法可应对不同程度(10%~80%)的图像缺失,且对数据缺失图像的非刚性配准表现较好。模体验证中86.8%的跟踪误差<3 mm,<2 mm的比例约80%作用。配准后标准化互信息(NMI)由1.18±0.02提高到1.20±0.02(t=-6.78,P=0.001)。临床病例肿瘤运动以平移为主,平均位移3.78 mm,最大位移可达7.46 mm。配准结果显示存在非刚性形变,配准后NMI由1.21±0.03增至到1.22±0.03(t=-2.91,P=0.001)。结论 肿瘤跟踪算法跟踪精度可靠且鲁棒性好,可用于无创、实时、无额外设备和辐射剂量的肿瘤跟踪。     

MR flow imaging by velocity-compensated/uncompensated difference images

The phase shifts acquired by motion of excited spins along magnetic field gradients can result in decreased signal intensity from blood vessels in conventional magnetic resonance images. The imaging technique can be modified with the use of additional gradient pulses so as to either compensate these phase shifts and increase the signal from the vessels or augment the phase shift and decrease the signal, without altering the signal from stationary tissues. Making a difference image from images made with and without sensitization to motion will cancel out the stationary tissues, leaving an image of the vessels alone. The technique does not require cardiac gating, shows veins as well as arteries, and can be performed in an interleaved manner to avoid registration errors due to patient motion.     

Robust segmentation methods with an application to aortic pulse wave velocity calculation

Aortic stiffness has proven to be an important diagnostic and prognostic factor of many cardiovascular diseases, as well as an estimate of overall cardiovascular health. Pulse wave velocity (PWV) represents a good measure of the aortic stiffness, while the aortic distensibility is used as an aortic elasticity index. Obtaining the PWV and the aortic distensibility from magnetic resonance imaging (MRI) data requires diverse segmentation tasks, namely the extraction of the aortic center line and the segmentation of aortic regions, combined with signal processing methods for the analysis of the pulse wave. In our study non-contrasted MRI images of abdomen were used in healthy volunteers (22 data sets) for the sake of non-invasive analysis and contrasted magnetic resonance (MR) images were used for the aortic examination of Marfan syndrome patients (8 data sets). In this research we present a novel robust segmentation technique for the PWV and aortic distensibility calculation as a complete image processing toolbox. We introduce a novel graph-based method for the centerline extraction of a thoraco-abdominal aorta for the length calculation from 3-D MRI data, robust to artifacts and noise. Moreover, we design a new projection-based segmentation method for transverse aortic region delineation in cardiac magnetic resonance (CMR) images which is robust to high presence of artifacts. Finally, we propose a novel method for analysis of velocity curves in order to obtain pulse wave propagation times. In order to validate the proposed method we compare the obtained results with manually determined aortic centerlines and a region segmentation by an expert, while the results of the PWV measurement were compared to a validated software (LUMC, Leiden, the Netherlands). The obtained results show high correctness and effectiveness of our method for the aortic PWV and distensibility calculation.     

Improvement in the quantification of myocardial perfusion using an automatic spline-based registration algorithm

PURPOSE: To improve the quantification of myocardial perfusion by registering the time series of magnetic resonance (MR) images with injection of gadolinium. MATERIALS AND METHODS: Eight patients underwent MR scans to perform myocardial perfusion exam. Two short axis views of the left ventricle (LV) were acquired in free breathing. Two masks for performing the spatial registration of the images were evaluated. The registration was based on pixel intensity in a multi-resolution scheme. The efficiency of this correction was evaluated by calculating geometric residual displacement of the LV and by fitting the data to a compartment model fit with two parameters: K1, the blood-to-myocardium transfer coefficient, and Vd, the distribution volume of the contrast media. RESULTS: The registration stage allowed a decrease in the observed motion of the LV from more than 1.98 +/- 0.68 mm to less than 0.56 +/- 0.18 mm (mean +/- SD). Variability obtained in the perfusion analysis decreased from 46 +/- 103% to 5+/- 4% for K1 parameter and from 18 +/- 21% to 5 +/- 5% for Vd parameter. CONCLUSION: As with manual correction, this automatic motion correction leads to accurate perfusion parameters in dynamic cardiac MR imaging after contrast agent injection. This automatic stage requires placing only one mask over one frame of the perfusion study instead of manually shifting each image to fit a reference image of the perfusion study.     

Automatic motion correction for quantification of myocardial perfusion with dynamic magnetic resonance imaging.

Respiratory motion makes it difficult to quantify myocardial perfusion with dynamic magnetic resonance imaging (MRI). The purpose of this study was to evaluate an automatic registration method for motion correction for quantification of myocardial perfusion with dynamic MRI. The present method was based on the gradient-based method with robust estimation of displacement parameters. For comparison, we also corrected for motion with manual registration as the benchmark. The myocardial kinetic parameters, K1 (rate constant for transfer of contrast agent from blood to myocardium) and k2 (rate constant for transfer from myocardium to blood), were calculated from dynamic images with a two-compartment model. The images corrected by the present method were similar to those corrected by manual registration. The kinetic parameters obtained after motion correction with the present method were close to those obtained after motion correction with manual registration. These results suggest that the present method is useful for motion correction for quantification of myocardial perfusion with dynamic MRI.     

A comparative study on manual and automatic slice-to-volume registration of CT images

In order to assess the clinical relevance of a slice-to-volume registration algorithm, this technique was compared to manual registration. Reformatted images obtained from a diagnostic CT examination of the lower abdomen were reviewed and manually registered by 41 individuals. The results were refined by the algorithm. Furthermore, a fully automatic registration of the single slices to the whole CT examination, without manual initialization, was also performed. The manual registration error for rotation and translation was found to be 2.7±2.8 ° and 4.0±2.5 mm. The automated registration algorithm significantly reduced the registration error to 1.6±2.6 ° and 1.3±1.6 mm (p = 0.01). In 3 of 41 (7.3%) registration cases, the automated registration algorithm failed completely. On average, the time required for manual registration was 213±197 s; automatic registration took 82±15 s. Registration was also performed without any human interaction. The resulting registration error of the algorithm without manual pre-registration was found to be 2.9±2.9 ° and 1.1±0.2 mm. Here, a registration took 91±6 s, on average. Overall, the automated registration algorithm improved the accuracy of manual registration by 59% in rotation and 325% in translation. The absolute values are well within a clinically relevant range.     

Automatic segmentation propagation of the aorta in real-time phase contrast MRI using nonrigid registration

Purpose

To assess the use of a nonrigid registration technique for semi‐automatic segmentation of the aorta from real‐time velocity mapping MRI.

Materials and Methods

Real‐time phase contrast images were acquired to measure flow and stroke volumes in 10 subjects, during free breathing, at rest, and during exercise. A nonrigid registration algorithm was developed to propagate a manually drawn region of interest in the aorta from one frame to all other frames of the real‐time sequence (148 images). Thus the technique provided a semi‐automatic segmentation over the whole sequence of images. The accuracy was assessed by comparison with manual segmentations in terms of Dice overlap measures and stroke volumes (SV).

Results

Semi‐automatic segmentations were comparable to manual ones (Dice score of 0.89 ± 0.04). Inter‐observer reproducibility was similar for manual and semi‐automatic segmentations (Dice score of 0.90 ± 0.04 in both cases, the difference was not significant). SV measurements also showed good agreement between manual and semi‐automatic segmentations (correlation coefficient r > 0.94), and the differences were not statistically significant.

Conclusion

Although real‐time phase contrast images have compromised image quality, a fast and robust segmentation of the aorta was possible using the registration‐based technique. J. Magn. Reson. Imaging 2011;33:232–238. © 2010 Wiley‐Liss, Inc.     

Validation of motion correction techniques for liver CT perfusion studies

Objectives

Motion in images potentially compromises the evaluation of temporally acquired CT perfusion (CTp) data; image registration should mitigate this, but first requires validation. Our objective was to compare the relative performance of manual, rigid and non-rigid registration techniques to correct anatomical misalignment in acquired liver CTp data sets.

Methods

17 data sets in patients with liver tumours who had undergone a CTp protocol were evaluated. Each data set consisted of a cine acquisition during a breath-hold (Phase 1), followed by six further sets of cine scans (each containing 11 images) acquired during free breathing (Phase 2). Phase 2 images were registered to a reference image from Phase 1 cine using two semi-automated intensity-based registration techniques (rigid and non-rigid) and a manual technique (the only option available in the relevant vendor CTp software). The performance of each technique to align liver anatomy was assessed by four observers, independently and blindly, on two separate occasions, using a semi-quantitative visual validation study (employing a six-point score). The registration techniques were statistically compared using an ordinal probit regression model.

Results

306 registrations (2448 observer scores) were evaluated. The three registration techniques were significantly different from each other (p=0.03). On pairwise comparison, the semi-automated techniques were significantly superior to the manual technique, with non-rigid significantly superior to rigid (p<0.0001), which in turn was significantly superior to manual registration (p=0.04).

Conclusion

Semi-automated registration techniques achieved superior alignment of liver anatomy compared with the manual technique. We hope this will translate into more reliable CTp analyses.Image registration and motion correction constitute a general class of challenges that impact on many areas of imaging and radiology. One specific area in which they may have a potential impact is in CT perfusion (CTp).There is increasing interest in the ability of CT to evaluate perfusion of tumours and to better understand the behaviour of tumours and the effects of treatments and therapies on tumours [1,2]. CTp is an evolving technique with potentially wide-ranging applications in oncology, including diagnosis, treatment evaluation and prognostication [3]. The technique relies on the acquisition of time–intensity plots from tissues of interest and vascular supply(s), following intravenous administration of a tracer (iodinated CT contrast medium). The incorporation of this information into appropriate physiological models allows computation of tissue perfusion parameters [4]. Parameters that can be derived include tissue blood flow, blood volume and permeability.CTp data are typically acquired with cine CT of relatively narrow collimation (2–4 cm) performed through the tissue/lesion of interest. A major challenge for CTp is the acquisition of reliable pixel-level time–intensity plots that extend for a sufficient length of time to adequately characterise the perfusion of tissues under consideration. This effectively requires adequate anatomical alignment, or motion correction, of the relevant tissues/lesions of interest. This is particularly problematic in body applications such as the upper abdomen, compared with stationary tissues such as the brain [5,6] and pelvis [7], as movement is inevitable, particularly if acquisitions beyond a single breath-hold are required. The need for prolonged data acquisition is suggested in the literature [8].One approach to acquiring the prolonged data for implementation in CTp is to obtain the necessary length of data during free breathing, which would then be unconstrained by the requirements and rigours of breath-holding. However, in order to maintain spatial fidelity, adequate registration algorithms need to be available.The motivation for this work was that there is, unfortunately, no specific registration software/functionality within the vendor environment of the particular CTp software package being used in our work (CT Perfusion 4; GE Healthcare, Waukesha, WI) to perform registrations of the kind described above. The only technique currently available in this application is manual selection of images. This is clearly extremely time-consuming, and furthermore is prone to errors. Availability of an automated or semi-automated registration algorithm would be advantageous; however, it is clearly a prerequisite that the methodology should first be validated before incorporation into CTp analyses.In this work, we investigate the performance of rigid and non-rigid intensity-based registration techniques to recover liver misregistration in data acquired from liver CTp data sets, as well as comparing the results with those obtained from the currently available manual registration technique.     

Improvement of multislice oxygen-enhanced MRI of the lung by fully automatic non-rigid image registration

Purpose

In oxygen-enhanced magnetic resonance imaging of the lung (O2-MRI), motion artifacts related to breathing hamper the quality of the parametric O2-maps. In this study, fully automatic non-rigid image registration was assessed as a post-processing method to improve the quality of O2-MRI.

Materials and methods

Twenty healthy volunteers were investigated on a 1.5 T MR system. O2-MRI was obtained in four coronal sections using an IR-HASTE sequence with TE/TI of 12/1200 ms. Each section was repeatedly imaged during oxygen and room-air ventilation. Spatial differences among the images were corrected by fully automatic non-rigid registration. Signal variability, relative enhancement ratio between oxygen and room air images, and spatial heterogeneity of lung enhancement were assessed before and after image registration.

Results

Motion artifacts were corrected in 5–10 s. Non-rigid registration reduced signal variability of the source images and heterogeneity of the O2-maps by 1.1 ± 0.2% and 11.2 ± 2.9%, respectively (p < 0.0001). Registration did not influence O2 relative enhancement ratio (p = 0.06).

Conclusion

Fully automatic non-rigid image registration improves the quality of multislice oxygen-enhanced MRI of the lung.     

Automated vs. manual assessment of left ventricular function in cardiac multidetector row computed tomography: comparison with magnetic resonance imaging

We compared semiautomatic contour detection and manual contour tracing in cardiac multidetector row computed tomography (MDCT) with magnetic resonance imaging (MRI) for calculation of left-ventricular (LV) volumes. The study included 30 patients who underwent contrast-enhanced cardiac MDCT and cardiac cine-MRI. Were calculated 8 mm short-axis slices from MDCT data using three-dimensional multiphase image reconstruction. LV volumes including peak ejection rate and peak filling rate were calculated from manually and semiautomatically determined contours. Results were compared to those from cine-MRI with manually drawn contours as the standard of reference. We found good agreement for the LV volumes, with an ejection fraction of 47.1±9.4% for manually drawn contours, 47.9±9.9% for semiautomatically detected contours on MDCT, and 48.0±10.2% for MRI. Except for peak-filling rate analysis of variance revealed no difference between any of these techniques. Bland-Altman plots and Lin’s concordance correlation coefficient showed best agreement between MRI and manual contour tracing in MDCT. Calculation of LV volumes using either semiautomatic or manual contour tracing in cardiac MDCT is therefore feasible when compared to MRI. Automated contour detection needs to be improved to equal manual contour tracing.     

Dynamic three-dimensional undersampled data reconstruction employing temporal registration.

Dynamic 3D imaging is needed for many applications such as imaging of the heart, joints, and abdomen. For these, the contrast and resolution that magnetic resonance imaging (MRI) offers are desirable. Unfortunately, the long acquisition time of MRI limits its application. Several techniques have been proposed to shorten the scan time by undersampling the k-space. To recover the missing data they make assumptions about the object's motion, restricting it in space, spatial frequency, temporal frequency, or a combination of space and temporal frequency. These assumptions limit the applicability of each technique. In this work we propose a reconstruction technique based on a weaker complementary assumption that restricts the motion in time. The technique exploits the redundancy of information in the object domain by predicting time frames from frames where there is little motion. The proposed method is well suited for several applications, in particular for cardiac imaging, considering that the heart remains relatively still during an important fraction of the cardiac cycle, or joint imaging where the motion can easily be controlled. This paper presents the new technique and the results of applying it to knee and cardiac imaging. The results show that the new technique can effectively reconstruct dynamic images acquired with an undersampling factor of 5. The resulting images suffer from little temporal and spatial blurring, significantly better than a sliding window reconstruction. An important attraction of the technique is that it combines reconstruction and registration, thus providing not only the 3D images but also its motion quantification. The method can be adapted to non-Cartesian k-space trajectories and nonuniform undersampling patterns.     

Constrained reverse diffusion for thick slice interpolation of 3D volumetric MRI images

Due to physical limitations inherent in magnetic resonance imaging scanners, three dimensional volumetric scans are often acquired with anisotropic voxel resolution. We investigate several interpolation approaches to reduce the anisotropy and present a novel approach - constrained reverse diffusion for thick slice interpolation. This technique was compared to common methods: linear and cubic B-Spline interpolation and a technique based on non-rigid registration of neighboring slices. The methods were evaluated on artificial MR phantoms and real MR scans of human brain. The constrained reverse diffusion approach delivered promising results and provides an alternative for thick slice interpolation, especially for higher anisotropy factors.