Surgery – Diagnostic testing – Detecting nuclear – electromagnetic – or ultrasonic radiation
Reexamination Certificate
2000-08-30
2002-04-23
Smith, Ruth S. (Department: 3737)
Surgery
Diagnostic testing
Detecting nuclear, electromagnetic, or ultrasonic radiation
C600S420000, C324S306000, C324S309000
Reexamination Certificate
active
06377835
ABSTRACT:
BACKGROUND OF THE INVENTION
1. Field of the Invention
The present invention is directed to a method for contrast enhanced MR angiography, and in particular to a method for producing images wherein arteries and veins are clearly visually separated.
2. Description of the Prior Art
Contrast enhanced MR angiography (MRA) is a valuable clinical tool for the evaluation of various types of vascular diseases Prince et al. “3D Contrast MR Angiography,” Berlin: Springer, 1997. With the high speed gradient systems available today, image acquisition time of a 3D MRA data set can be reduced to a fraction of a breath hold period. Thus, the arrival, passage and wash-out of a contrast agent bolus can be visualized in a series of MRA data sets, that are acquired in a single breath hold.
Since there is a sequential enhancement of arteries and veins, arterial angiograms with virtually no venous overlay can be acquired if the transit time from the arteries to the neighboring veins is longer than the acquisition time of an MRA data set. Venograms are created from later scans, wherein the signal intensity in the veins is at maximum, but with substantial remaining enhancement of the superimposed arteries. To improve the vessel delineation especially in venograms, a subtraction of arterial from venous data sets can be performed, however, a subtraction increases the noise level in the resultant images. In addition, image subtraction crucially depends on the acquisition of a pure arterial phase, since any venous contamination causes artificial elimination of veins in the subtracted venogram.
In MRI, correlation analysis is typically used with functional imaging of the brain (fMRI), where the signal-time-course at a specific location in the brain is correlated with an external reference function, as described in Bandettini et al., “Processing Strategies for Time-course Data Sets in Functional MRI in the Human Brain, ” Magn. Reson. Med. 1993; 30:161-173. Correlation coefficient maps are computed, which are used as an indicator for the strength and the statistical significance of cortical activation. In correlation MRA, the signal-time-course in a selected vessel of interest serves as a reference function. Correlation maps calculated from this reference function highlight structures in the MRA data sets, that show a similar temporal signal behavior as the selected vessel.
Moreover, correlation MRA is a known technique that heretofore has been applied only to two-dimensional projection MRA, where images are sampled at sub-second frame rates as described in Strecker et al., “Functional MRA Combining 2d MR DSA and Correlation Analysis, ” Proceedings ISMRM, Seventh Annual Meeting, Philadelphia, 1999, p 484.
Separate arteriograms and venograms can be computed from a single 3D CE-MRA data set by making use of the spatial connectivity of arteries and veins. Good results have been achieved with volume rendering in combination with connectivity algorithms as described in Cline et al., “Volume Rendering and Connectivity Algorithms for MR Angiography,” Magn. Reson. Med. 1991; 18: 384-394. Semi-automatic volume rendering, however, can become very time consuming in anatomical regions, where arteries and veins are in close proximity. It also often fails in smaller vessels, where signal levels are comparable with those of the surrounding tissue.
A different approach using matched filtering has been proposed for the separation of arteries from veins in time-resolved ECG-triggered time-of-flight MRA in Wang et al., “Generalized Matched Filtering for Time-Resolved MR-Angiography of Pulsatile Flow, ” Magn. Reson. Med. 1993; 30(5):600-608. With matched filtering a set of global or local weighting factors is determined. The resultant arterial or venous image is then formed from the weighted sum of the time-domain data. Mathematically, correlation and matched filtering are very similar, because they can both be described as the product of two vectors. In matched filtering, however, the vectors are formed directly from the data, but correlation first subtracts the temporal average from the data, which intrinsically provides a mechanism for the suppression of static signal.
A separation technique has been described in Mazaheri et al., “Vessel Segmentation in 3D MR Angiography Using Time Resolved Acquisition Curves,” Proceedings ISMRM, Seventh Annual Meeting, Philadelphia, 1999, p 2181, that determines the squared Euclidian distance of the local signal-time vector from two given arterial and venous reference vectors (sum of squared differences). To be independent of the signal intensities, the signal-time vector is scaled by a factor that minimizes the squared distance. In a two-dimensional scatter plot of the arterial-venous distance space, arteries and veins form loosely connected islands. Vessel classification is then performed based on a proximity measure in the distance space. This technique is complicated and requires a classification criterion in the two-dimensional feature space and so far it has only been used in MRA studies of the carotid arteries and veins. It has to be evaluated, whether it can be applied to pulmonary MRA.
SUMMARY OF THE INVENTION
It is an object of the present invention to provide a method for conducting MR angiography (MRA) wherein images are produced containing clearly visually separated veins and arteries.
This object is achieved in accordance with the invention in a method wherein the technique of correlation MRA is used to separate lung arteries from lung veins in contrast-enhanced multi-phase 3D MRA data sets, that are acquired at a temporal resolution of several seconds. In the pulmonary vasculature three-dimensional techniques are required, because the arterial and venous vascular trees are arranged in a complex anatomic orientations in space, where projection methods would inevitably suffer from signal super-positions. Using reference signal-time-courses from manually selected regions of interest (ROI) in the main lung arteries and veins, arterial and venous correlation angiograms are computed. For image display, maximum intensity projections are calculated from the correlation data sets and compared to conventional 3D MRA subtraction angiograms.
Initially, region of interest (ROI) is defined in a slice of an examination subject containing either a major pulmonary artery or vein. The average ROI signal as a function of time is then stored as a reference time curve for the subsequent cross-correlation analysis. For each spatial position in the 3D data set, the unormalized cross-correlation between the reference time curve and the signal-time curve is calculated, negative values are set to zero and the result is stored in a new 3D data set. Subsequently, maximum intensity projections (MIPs) are computed from the arterial and venous correlation maps, to allow visualization of the pulmonary vasculature.
The key issue for time-resolved multi-phase pulmonary angiography is the trade-off in spatial resolution and anatomic coverage for shortening of the acquisition time of each phase. Analysis of pulmonary enhancement kinetics shows that in healthy individuals there is only a short time window of approximately 3s during which only the pulmonary arteries are enhanced (FIG.
1
). With increasing acquisition time per phase pulmonary arteries and veins enhance within the same image and it becomes impossible to separate them on the basis of their enhancement kinetics using conventional techniques.
The initially described inventive method using cross-correlation analysis results in even further improved artery and vein separation when multiple boli are administered to the patient in the procedure for obtaining the 3D MRA data set. In particular, a dual bolus procedure has proven effective.
REFERENCES:
patent: 5115812 (1992-05-01), Sano et al.
patent: 5417213 (1995-05-01), Prince
patent: 5830143 (1998-11-01), Mistretta et al.
“Processing Strategies for Time-Course Data Sets in Functional MRI of the Human Brain,” Bandettini et al., Magnetic Resonance in Medicine, vol. 30 (1993), pp. 161-173.
“Functional
Bock Michael
Knopp Michael V.
Laub Gerhard
Schoenberg Stefan O.
Schiff & Hardin & Waite
Siemens Aktiengesellschaft
Smith Ruth S.
LandOfFree
Method for separating arteries and veins in 3D MR... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method for separating arteries and veins in 3D MR..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method for separating arteries and veins in 3D MR... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2817775