Digital subtraction angiography (DSA): a neurosurgical perspective

Enhancing cerebral vasculature analysis with pathlength-corrected 2D angiographic parametric imaging: A feasibility study

BACKGROUND

2D angiographic parametric imaging (API) quantitatively extracts imaging biomarkers related to contrast flow and is conventionally applied to 2D digitally subtracted angiograms (DSA's). In the interventional suite, API is typically performed using 1-2 projection views and is limited by vessel overlap, foreshortening, and depth-integration of contrast motion.

PURPOSE

This work explores the use of a pathlength-correction metric to overcome the limitations of 2D-API: the primary objective was to study the effect of converting 3D contrast flow to projected contrast flow using a simulated angiographic framework created with computational fluid dynamics (CFD) simulations, thereby removing acquisition variability.

METHODS

The pathlength-correction framework was applied to in-silico angiograms, generating a reference (i.e., ground-truth) volumetric contrast distribution in four patient-specific intracranial aneurysm geometries. Biplane projections of contrast flow were created from the reference volumetric contrast distributions, assuming a cone-beam geometry. A Parker-weighted reconstruction was performed to obtain a binary representation of the vessel structure in 3D. Standard ray tracing techniques were then used to track the intersection of a ray from the focal spot with each voxel of the reconstructed vessel wall to a pixel in the detector plane. The lengths of each ray through the 3D vessel lumen were then projected along each ray-path to create a pathlength-correction map, where the pixel intensity in the detector plane corresponds to the vessel width along each source-detector ray. By dividing the projection sequences with this correction map, 2D pathlength-corrected in-silico angiograms were obtained. We then performed voxel-wise (3D) API on the ground-truth contrast distribution and compared it to pixel-wise (2D) API, both with and without pathlength correction for each biplane view. The percentage difference (PD) between the resultant API biomarkers in each dataset were calculated within the aneurysm region of interest (ROI).

RESULTS

Intensity-based API parameters, such as the area under the curve (AUC) and peak height (PH), exhibited notable changes in magnitude and spatial distribution following pathlength correction: these now accurately represent conservation of mass of injected contrast media within each arterial geometry and accurately reflect regions of stagnation and recirculation in each aneurysm ROI. Improved agreement was observed between these biomarkers in the pathlength-corrected biplane maps: the maximum PD within the aneurysm ROI is 3.3% with pathlength correction and 47.7% without pathlength correction. As expected, improved agreement with ROI-averaged ground-truth 3D counterparts was observed for all aneurysm geometries, particularly large aneurysms: the maximum PD for both AUC and PH was 5.8%. Temporal parameters (mean transit time, MTT, time-to-peak, TTP, time-to-arrival, TTA) remained unaffected after pathlength correction.

CONCLUSIONS

This study indicates that the values of intensity-based API parameters obtained with conventional 2D-API, without pathlength correction, are highly dependent on the projection orientation, and uncorrected API should be avoided for hemodynamic analysis. The proposed metric can standardize 2D API-derived biomarkers independent of projection orientation, potentially improving the diagnostic value of all acquired 2D-DSA's. Integration of a pathlength correction map into the imaging process can allow for improved interpretation of biomarkers in 2D space, which may lead to improved diagnostic accuracy during procedures involving the cerebral vasculature.

Deep Learning-based Angiogram Generation Model for Cerebral Angiography without Misregistration Artifacts

Background Digital subtraction angiography (DSA) generates an image by subtracting a mask image from a dynamic angiogram. However, patient movement-caused misregistration artifacts can result in unclear DSA images that interrupt procedures. Purpose To train and to validate a deep learning (DL)-based model to produce DSA-like cerebral angiograms directly from dynamic angiograms and then quantitatively and visually evaluate these angiograms for clinical usefulness. Materials and Methods A retrospective model development and validation study was conducted on dynamic and DSA image pairs consecutively collected from January 2019 through April 2019. Angiograms showing misregistration were first separated per patient by two radiologists and sorted into the misregistration test data set. Nonmisregistration angiograms were divided into development and external test data sets at a ratio of 8:1 per patient. The development data set was divided into training and validation data sets at ratio of 3:1 per patient. The DL model was created by using the training data set, tuned with the validation data set, and then evaluated quantitatively with the external test data set and visually with the misregistration test data set. Quantitative evaluations used the peak signal-to-noise ratio (PSNR) and the structural similarity (SSIM) with mixed liner models. Visual evaluation was conducted by using a numerical rating scale. Results The training, validation, nonmisregistration test, and misregistration test data sets included 10 751, 2784, 1346, and 711 paired images collected from 40 patients (mean age, 62 years ± 11 [standard deviation]; 33 women). In the quantitative evaluation, DL-generated angiograms showed a mean PSNR value of 40.2 dB ± 4.05 and a mean SSIM value of 0.97 ± 0.02, indicating high coincidence with the paired DSA images. In the visual evaluation, the median ratings of the DL-generated angiograms were similar to or better than those of the original DSA images for all 24 sequences. Conclusion The deep learning-based model provided clinically useful cerebral angiograms free from clinically significant artifacts directly from dynamic angiograms. Published under a CC BY 4.0 license. Supplemental material is available for this article.

Feasibility study for use of angiographic parametric imaging and deep neural networks for intracranial aneurysm occlusion prediction

Background

Angiographic parametric imaging (API), based on digital subtraction angiography (DSA), is a quantitative imaging tool that may be used to extract contrast flow parameters related to hemodynamic conditions in abnormal pathologies such as intracranial aneurysms (IAs).

Objective

To investigate the feasibility of using deep neural networks (DNNs) and API to predict IA occlusion using pre- and post-intervention DSAs.

Methods

We analyzed DSA images of IAs pre- and post-treatment to extract API parameters in the IA dome and the corresponding main artery (un-normalized data). We implemented a two-step correction to account for injection variability (normalized data) and projection foreshortening (relative data). A DNN was trained to predict a binary IA occlusion outcome: occluded/unoccluded. Network performance was assessed with area under the receiver operating characteristic curve (AUROC) and classification accuracy. To evaluate the effect of the proposed corrections, prediction accuracy analysis was performed after each normalization step.

Results

The study included 190 IAs. The mean and median duration between treatment and follow-up was 9.8 and 8.0 months, respectively. For the un-normalized, normalized, and relative subgroups, the DNN average prediction accuracies for IA occlusion were 62.5% (95% CI 60.5% to 64.4%), 70.8% (95% CI 68.2% to 73.4%), and 77.9% (95% CI 76.2% to 79.6%). The average AUROCs for the same subgroups were 0.48 (0.44–0.52), 0.67 (0.61–0.73), and 0.77 (0.74–0.80).

Conclusions

The study demonstrated the feasibility of using API and DNNs to predict IA occlusion using only pre- and post-intervention angiographic information.

Automatic radiomic feature extraction using deep learning for angiographic parametric imaging of intracranial aneurysms

BACKGROUND

Angiographic parametric imaging (API) is an imaging method that uses digital subtraction angiography (DSA) to characterize contrast media dynamics throughout the vasculature. This requires manual placement of a region of interest over a lesion (eg, an aneurysm sac) by an operator.

OBJECTIVE

The purpose of our work was to determine if a convolutional neural network (CNN) was able to identify and segment the intracranial aneurysm (IA) sac in a DSA and extract API radiomic features with minimal errors compared with human user results.

METHODS

Three hundred and fifty angiographic images of IAs were retrospectively collected. The IAs and surrounding vasculature were manually contoured and the masks put to a CNN tasked with semantic segmentation. The CNN segmentations were assessed for accuracy using the Dice similarity coefficient (DSC) and Jaccard index (JI). Area under the receiver operating characteristic curve (AUROC) was computed. API features based on the CNN segmentation were compared with the human user results.

RESULTS

The mean JI was 0.823 (95% CI 0.783 to 0.863) for the IA and 0.737 (95% CI 0.682 to 0.792) for the vasculature. The mean DSC was 0.903 (95% CI 0.867 to 0.937) for the IA and 0.849 (95% CI 0.811 to 0.887) for the vasculature. The mean AUROC was 0.791 (95% CI 0.740 to 0.817) for the IA and 0.715 (95% CI 0.678 to 0.733) for the vasculature. All five API features measured inside the predicted masks were within 18% of those measured inside manually contoured masks.

CONCLUSIONS

CNN segmentation of IAs and surrounding vasculature from DSA images is non-inferior to manual contours of aneurysms and can be used in parametric imaging procedures.

Effect of injection technique on temporal parametric imaging derived from digital subtraction angiography in patient specific phantoms

Parametric imaging maps (PIM's) derived from digital subtraction angiography (DSA) for the cerebral arterial flow assessment in clinical settings have been proposed, but experiments have yet to determine the reliability of such studies. For this study, we have observed the effects of different injection techniques on PIM's. A flow circuit set to physiologic conditions was created using an internal carotid artery phantom. PIM's were derived for two catheter positions, two different contrast bolus injection volumes (5ml and 10 ml), and four injection rates (5, 10, 15 and 20 ml/s). Using a gamma variate fitting approach, we derived PIM's for mean-transit-time (MTT), time-to-peak (TTP) and bolus-arrivaltime (BAT). For the same injection rates, a larger bolus resulted in an increased MTT and TTP, while a faster injection rate resulted in a shorter MTT, TTP, and BAT. In addition, the position of the catheter tip within the vasculature directly affected the PIM. The experiment showed that the PIM is strongly correlated with the injection conditions, and, therefore, they have to be interpreted with caution. PIM images must be taken from the same patient to be able to be meaningfully compared. These comparisons can include pre- and post-treatment images taken immediately before and after an interventional procedure or simultaneous arterial flow comparisons through the left and right cerebral hemispheres. Due to the strong correlation between PIM and injection conditions, this study indicates that this assessment method should be used only to compare flow changes before and after treatment within the same patient using the same injection conditions.

Can magnetic resonance imaging alone accurately define the arteriovenous nidus for gamma knife radiosurgery?

Object. Current radiosurgical treatment of arteriovenous malformations (AVMs) relies on planning protocols that integrate data from both magnetic resonance (MR) imaging and stereotactic angiography studies. Angiography, however, is invasive and associated with a small but well-defined risk of neurological and systemic complications. Magnetic resonance imaging, on the other hand, is noninvasive with multiplanar capability, demonstrates good anatomical detail, and has been shown to be superior to angiography in the delineation of selected AVMs.

Methods. In this study, MR imaging—related accuracy of defining the AVM nidus in gamma knife radiosurgery is investigated using only T1- and T2-weighted sequences.

Conclusions. Little interobserver variability was observed and AVM nidi, as demonstrated on T1- and T2-weighted MR images, were well correlated in terms of size. The displacement of the new target, however, from the original nidus, was not predictable and occasionally was significant, thus precluding safe radiosurgical planning.

Digital subtraction angiography (DSA) of the extracranial cerebral vessels: a direct comparison between intravenous and intra-arterial DSA

Fifty patients being investigated for ischaemic cerebrovascular disease underwent both intravenous digital subtraction angiography (IV-DSA) and intra-arterial arch digital subtraction angiography (arch-DSA) to enable a direct comparison to be made between the two techniques. The overall quality of the images obtained by arch-DSA was better than that obtained by IV-DSA, but there was no significant difference between the number and gradation of stenoses detected by the two techniques. No significant stenosis was missed by IV-DSA and the technique was found to be sufficiently accurate for it to be used as a screening procedure for carotid bifurcation stenoses suitable for carotid endarterectomy. The advantages and limitations of both techniques are discussed.