Blood flow measurement using digital angiography and parametric imaging

Synthetic interpolated DSA for radiation exposure reduction via gamma variate contrast flow modeling: a retrospective cohort study

BACKGROUND

Digital subtraction angiography (DSA) yields high cumulative radiation dosages (RD) delivered to patients. We present a temporal interpolation of low frame rate angiograms as a method to reduce cumulative RDs.

METHODS

Patients undergoing interventional evaluation and treatment of cerebrovascular vasospasm following subarachnoid hemorrhage were retrospectively identified. DSAs containing pre- and post-intervention runs capturing the full arterial, capillary, and venous phases with at least 16 frames each were selected. Frame rate reduction (FRR) of the original DSAs was performed to 50%, 66%, and 75% of the original frame rate. Missing frames were regenerated by sampling a gamma variate model (GVM) fit to the contrast response curves to the reduced data. A formal reader study was performed to assess the diagnostic accuracy of the "synthetic" studies (sDSA) compared to the original DSA.

RESULTS

Thirty-eight studies met inclusion criteria (average RD 1,361.9 mGy). Seven were excluded for differing views, magnifications, or motion. GVMs fit to 50%, 66%, and 75% FRR studies demonstrated average voxel errors of 2.0 ± 2.5% (mean ± standard deviation), 6.5 ± 1.5%, and 27 ± 2%, respectively for anteroposterior projections, 2.0 ± 2.2%, 15.0 ± 3.1%, and 14.8 ± 13.0% for lateral projections, respectively. Reconstructions took 0.51 s/study. Reader studies demonstrated an average rating of 12.8 (95% CI 12.3-13.3) for 75% FRR, 12.7 (12.2-13.2) for 66% FRR and 12.0 (11.5-12.5) for 50% FRR using Subjective Image Grading Scale. Kendall's coefficient of concordance resulted in W = 0.506.

CONCLUSION

FRR by 75% combined with GVM reconstruction does not compromise diagnostic quality for the assessment of cerebral vasculature.

RELEVANCE STATEMENT

Using this novel algorithm, it is possible to reduce the frame rate of DSA by as much as 75%, with a proportional reduction in radiation exposure, without degrading imaging quality.

KEY POINTS

• DSA delivers some of the highest doses of radiation to patients. • Frame rate reduction (FRR) was combined with bolus tracking to interpolate intermediate frames. • This technique provided a 75% FRR with preservation of diagnostic utility as graded by a formal reader study for cerebral angiography performed for the evaluation of cerebral vasospasm. • This approach can be applied to other types of angiography studies.

Vulnerability of Antarctica's ice shelves to meltwater-driven fracture

Atmospheric warming threatens to accelerate the retreat of the Antarctic Ice Sheet by increasing surface melting and facilitating 'hydrofracturing'^1-7, where meltwater flows into and enlarges fractures, potentially triggering ice-shelf collapse^3-5,8-10. The collapse of ice shelves that buttress^11-13 the ice sheet accelerates ice flow and sea-level rise^14-16. However, we do not know if and how much of the buttressing regions of Antarctica's ice shelves are vulnerable to hydrofracture if inundated with water. Here we provide two lines of evidence suggesting that many buttressing regions are vulnerable. First, we trained a deep convolutional neural network (DCNN) to map the surface expressions of fractures in satellite imagery across all Antarctic ice shelves. Second, we developed a stability diagram of fractures based on linear elastic fracture mechanics to predict where basal and dry surface fractures form under current stress conditions. We find close agreement between the theoretical prediction and the DCNN-mapped fractures, despite limitations associated with detecting fractures in satellite imagery. Finally, we used linear elastic fracture mechanics theory to predict where surface fractures would become unstable if filled with water. Many regions regularly inundated with meltwater today are resilient to hydrofracture-stresses are low enough that all water-filled fractures are stable. Conversely, 60 ± 10 per cent of ice shelves (by area) both buttress upstream ice and are vulnerable to hydrofracture if inundated with water. The DCNN map confirms the presence of fractures in these buttressing regions. Increased surface melting¹⁷ could trigger hydrofracturing if it leads to water inundating the widespread vulnerable regions we identify. These regions are where atmospheric warming may have the largest impact on ice-sheet mass balance.

Exploring Reperfusion Following Endovascular Thrombectomy

Background and Purpose- Cerebral perfusion in acute ischemic stroke patients is often assessed before endovascular thrombectomy (EVT), but rarely after. Perfusion data obtained following EVT may provide additional prognostic information. We developed a tool to quantitatively derive perfusion measurements from digital subtraction angiography (DSA) data and examined perfusion in patients following EVT. Methods- Source DSA images from acute anterior circulation stroke patients undergoing EVT were retrospectively assessed. Following deconvolution, maps of mean transit time (MTT) were generated from post-EVT DSA source data. Thrombolysis in Cerebral Infarction grades and MTT in patients with and without hemorrhagic transformation (HT) at 24 hours were compared. Receiver operating characteristic modeling was used to classify the presence/absence of HT at 24 hours by MTT. Results- Perfusion maps were generated in 50 patients using DSA acquisitions that were a median (interquartile range) of 9 (8-10) seconds in duration. The median post-EVT MTT within the affected territory was 2.6 (2.2-3.3) seconds. HT was observed on follow-up computed tomography in 16 (32%) patients. Thrombolysis in Cerebral Infarction grades did not differ in patients with HT from those without (P=0.575). Post-EVT MTT maps demonstrated focal areas of hyperperfusion (n=8) or persisting hypoperfusion (n=3) corresponding to the regions where HT later developed. The relationship between MTT and HT was U-shaped; HT occurred in patients at both the lowest and highest extremes of MTT. An MTT threshold <2 or >4 seconds was 81% sensitive and 94% specific for classifying the presence of HT at follow-up. Conclusions- Perfusion measurements can be obtained using DSA perfusion with minimal changes to current stroke protocols. Perfusion imaging post-recanalization may have additional clinical utility beyond visual assessment of source angiographic images alone.

Blood Flow Determinations Utilizing Digital Densitometry

A method of obtaining relative and absolute blood flow measurements from digital densitometry was evaluated with a simulated vessel phantom and a hydrodynamic model. A digital vascular imaging system capable of acquisition in 5122 and 10242 mode was used. Relative and absolute blood flow were measured using parameters derived from the densitometric curve. Since application of densitometric data to absolute flow measurements requires the vessel diameter, an algorithm for vessel size determination was created. Gray scale changes were demonstrated to be linearly related to contrast concentration. The variance of vessel size determination was significantly different in all combinations of 10242 and 5122 imaging with 15 cm or 35 cm field size. The error in vessel size determination was significantly less using the larger 10242 matrix and the smaller 15 cm image intensifier field size, as shown by the smaller variance. In relative flow determinations, there was good correlation between the flow and four parameters of the densitometric curve with no significant differences between 5122 and 10242 imaging. Absolute flow determinations had slightly lower correlation to actual flow but were not significantly different from relative flow determinations. Relative and absolute blood flow determinations can be performed adequately with either 5122 or 10242 imaging. The increased accuracy in vessel size determination with 10242 imaging makes this high resolution system potentially preferable to determine absolute blood flow.

Perfusion Angiography in Acute Ischemic Stroke

Visualization and quantification of blood flow are essential for the diagnosis and treatment evaluation of cerebrovascular diseases. For rapid imaging of the cerebrovasculature, digital subtraction angiography (DSA) remains the gold standard as it offers high spatial resolution. This paper lays out a methodological framework, named perfusion angiography, for the quantitative analysis and visualization of blood flow parameters from DSA images. The parameters, including cerebral blood flow (CBF) and cerebral blood volume (CBV), mean transit time (MTT), time-to-peak (TTP), and T_max, are computed using a bolus tracking method based on the deconvolution of the time-density curve on a pixel-by-pixel basis. The method is tested on 66 acute ischemic stroke patients treated with thrombectomy and/or tissue plasminogen activator (tPA) and also evaluated on an estimation task with known ground truth. This novel imaging tool provides unique insights into flow mechanisms that cannot be observed directly in DSA sequences and might be used to evaluate the impact of endovascular interventions more precisely.

X-ray videodensitometric methods for blood flow and velocity measurement: a critical review of literature

Blood flow rate and velocity are important parameters for the study of vascular systems, and for the diagnosis, monitoring and evaluation of treatment of cerebro- and cardiovascular disease. For rapid imaging of cerebral and cardiac blood vessels, digital x-ray subtraction angiography has numerous advantages over other modalities. Roentgen-videodensitometric techniques measure blood flow and velocity from changes of contrast material density in x-ray angiograms. Many roentgen-videodensitometric flow measurement methods can also be applied to CT, MR and rotational angiography images. Hence, roentgen-videodensitometric blood flow and velocity measurement from digital x-ray angiograms represents an important research topic. This work contains a critical review and bibliography surveying current and old developments in the field. We present an extensive survey of English-language publications on the subject and a classification of published algorithms. We also present descriptions and critical reviews of these algorithms. The algorithms are reviewed with requirements imposed by neuro- and cardiovascular clinical environments in mind.

Velocity measurement based on bolus tracking with the aid of three-dimensional reconstruction from digital subtraction angiography

The problem of blood flow measurement in x-ray angiography using measurements of the leading edge of the contrast bolus as it traverses the vascular bed is considered. A new technique for velocity measurement is presented based upon the ratio of the temporal derivative to the spatial derivative of the contrast bolus in the direction of flow. With the addition of a small correction factor, the value obtained is shown to reflect the transport velocity, or the velocity at which the contrast is transported down the vessel of interest. Most blood flow measurements based on bolus tracking techniques are actually using the contrast transport velocity to represent the blood flow velocity. Because of the streaming that occurs due to laminary flow conditions, the measured transport velocity is found to be somewhere between the average and the peak (central) fluid velocities for measurements taken during the traversal of the bolus leading edge. The spatial and temporal variation of the transport velocity are found to be consistent with the bolus motion expected in the presence of laminar flow. From x-ray images of contrast passage through simple tubes, we find that the derivative method measures the transport velocity during passage of the bolus leading edge. In most cases of laminar blood flow, the leading edge transport velocity can be 20%-40% higher than the average blood velocity.

Three-dimensional knowledge driven reconstruction of coronary trees

A knowledge-driven approach to the three-dimensional reconstruction of coronary artery trees by means of two X-ray projections is proposed. The spatial reconstruction of the tree skeleton is discussed. A binary tree model of the arterial structure and its projections is employed. Consequently, the reconstruction of the three-dimensional tree skeleton is achieved by (a) matching the skeletons of corresponding pairs of vascular segments in the two views and (b) back-projecting the coupled skeleton projections. From a geometrical point of view, the matching problem is, in general, ill-conditioned. For this reason, additional information sources were used. Thus, the matching phase is accomplished by using both the imaging geometry information, as well as anatomical and topological knowledge, about the coronary arteries coded in a rule base. As far as the back-projection phase is concerned, an algorithm was developed based on: (1) the imaging geometry, (2) the bounding of the back-projection error and (3) a contiguity criterion.

Segmentation of coronary arteriograms by iterative ternary classification

A segmentation algorithm for extracting arterial structures in coronary angiograms is presented. The algorithm mimics the process of interactive interpretation in human vision by iteratively implementing a ternary classification and learning process. Two gray-scale thresholds are computed to define three pixel classes: artery, background, and undecided. Then, two new thresholds for undecided pixels are computed using statistics conditioned upon the current classification. The threshold adaptation is governed by a learning algorithm based on the line and consistency measurements around each pixel. The process converges and results in a binary image. The performance of this algorithm on human coronary arteriograms was compared qualitatively to that of a relaxation algorithm and of a scattering based algorithm. Quantitative comparison was also made possible with computer generated images, which were obtained with the help of a model of the imaging chain and a process of interactive visualization of the modeled data. The iterative ternary classifier showed the best performance over a broad range of image quality. The study also demonstrated the use of visualization and user interaction in model building and algorithm development.

Pulsatile coronary flow determination by digital angiography

A new angiographic method to calculate absolute coronary arterial blood flow and its cyclic variations as a function of time, has been designed and evaluated. The method combines densitometric analysis of spatial and temporal aspects concerning the contrast propagation through the arterial system from digital images, and is based on applying the concept of conservation of contrast material in successive images. It requires a standard arteriographic procedure. In simulations with both constant as well as pulsatile flow through a coronary arterial phantom, an excellent agreement with electromagnetic flow measurements was demonstrated (r = 0.993 and r = 0.982, respectively). Preliminary clinical results show, that the method yields reproducible assessment of coronary flow patterns after repeated injections in a patient. In a coronary artery bypass graft, coronary flow patterns in a baseline and a drug-induced hyperemic state were obtained. A consistent coronary flow reserve value was determined after repeated examinations. The method has been shown to be feasible in clinical applications, uncomplicated and fast, but requires further animal and clinical validation in order to determine the possible applications, limitations and accuracy.

Measurement of absolute flow rate in vessels using a stereoscopic DSA system

We used a stereoscopic digital subtraction angiography (DSA) system to measure absolute blood flow rates in vessels. The magnification factor and the three-dimensional orientation of a selected vessel are obtained from automated analysis of stereoscopic DSA images. The cross-sectional area of the vessel is determined from the vessel diameter, which is measured with an iterative deconvolution technique. The time required for fluid to flow through a selected segment of a vessel is determined from the automated analysis of contrast medium 'time-density' curves. The effectiveness of these combined techniques was demonstrated in measurement of rates of both continuous and pulsatile flow in a vessel phantom, with the actual flow rate calibrated volumetrically or by an electromagnetic flowmeter. We have obtained accuracies in measured flow rates of approximately 5% and 18% for continuous and pulsatile flow respectively.

Automated identification of vessel contours in coronary arteriograms by an adaptive tracking algorithm

A tracking algorithm for identification of vessel contours in digital coronary arteriograms was developed and validated. Given an initial start-of-search point, the tracking process was fully automated by utilizing the spatial continuity of the vessel's centerline, orientation, diameter, and density. The incremental sections along a major vessel were sequentially identified, based on the assumptions of geometric similarity and continuation between adjacent incremental sections. The algorithm consisted of an extrapolation-update process which was guided by a matched filter. The filter parameters were adapted to the measured lumen width. The tracking process was robust and extremely efficient as indicated by test results on synthetic images, digital subtraction angiograms, and cineangiograms. The algorithm provided accurate measurement of lumen width and percent stenosis that was relatively invariant to the vessel's orientation, dynamic range, background variation, and degree of blurring.

The accuracy of digital subtraction angiography for the quantification of atherosclerosis

Digital subtraction angiography (DSA) allows the degree of arterial patency or stenosis to be rapidly quantified. We have assessed the accuracy with which a single-plane DSA system is able to quantify area patency by densitometric and geometric methods. Arterial phantoms were designed to test for systematic error; intra-arterial DSA images of critical lesions of the carotid bifurcation and the lower abdominal and peripheral vessels were used to determine intra- and interobserver reproducibility. The densitometric method, which was more accurate than the geometric method, had a mean systematic error of up to 4% and a mean intra-observer variability of about 15% (coefficient of variation). We have identified the principal sources of inaccuracy and ways in which it may be reduced.

Three-dimensional reconstruction of moving arterial beds from digital subtraction angiography

A system for three-dimensional reconstruction of dynamic (moving) vascular bed structures has been developed and is described. Input images are obtained from two-view (bi-plane or ECG correlated) X-ray angiograms. A target structure consisting of vessel branch points (nodes) and lines between the branch points is entered on the first of a sequence of images in one view. The movement of the nodes is indicated on subsequent images and on the images of the second view. The target is linearly warped according to the motion of the node points. Automatic edge detection (with subsequent operator correction) is used to detect centerlines and edges of vessels. Three-dimensional reconstruction is accomplished using a distance minimizing point matching technique. Finally, angle-corrected densitometric methods are used to refine the vessel cross section. Standard shaded surface display techniques are then used to display the moving arterial bed. Flow measurements are obtained by tracking the leading edge of the bolus down the three-dimensional arterial tree.