Practice enhancements with FastStroke ColorViz analysis in acute ischemic stroke
How to cite this article: Pai V, Ti JP, Tan LQ, Ho TS, Tham C, Sitoh YY. Practice enhancements with FastStroke ColorViz analysis in acute ischemic stroke. J Clin Imaging Sci 2022;12:19.
In acute ischemic stroke (AIS), large vessel occlusion (LVO) and the status of pial collaterals are important factors in decision-making for further treatment such as endovascular therapy. Multiphasic CT Angiogram (mCTA) is the mainstay of AIS imaging, allowing detection of LVO, evaluation of intracranial arterial dynamics, and quantification of pial collaterals. However, thorough mCTA evaluation entails scrutiny of multiple image datasets, individually and then simultaneously, which can be time-consuming, causing a potential delay in treatment. ColorViz (FastStroke, GE Healthcare, Milwaukee, Wisconsin) is a novel CT application which combines mCTA information into a single color-coded dataset for quick, unequivocal evaluation of pial collaterals. In our practice, ColorViz is both time-saving and increases the diagnostic accuracy of LVO and pial collaterals as well as medium vessel, multivessel and posterior circulation occlusions. In this article, we discuss the practical aspects of ColorViz in patients presenting with AIS.
Cerebral ischemia triggers the recruitment of multiple subsidiary anastomotic vascular pathways known as pial collaterals. This homeostatic response helps preserve cerebral perfusion and prolongs sustenance of the penumbra, resulting in reduced infarct volume.[2,3] Greater collateralization has also been associated with the enhanced benefit of endovascular therapy (EVT) and intravenous thrombolysis as well as a reduced risk of hemorrhagic transformation. This ultimately translates into better neurological outcomes and reduced mortality.[3,4] Over the years, the combination of unenhanced CT of the brain followed by a CT angiogram has emerged as the standard investigation in the setting of acute ischemic stroke (AIS), mainly due to its easy availability and speed of image acquisition.[5,6] A single-phase CT angiogram can accurately identify the site of occlusion; however, it lacks temporal resolution, limiting the evaluation of pial collaterals.
It is in this context that multiphasic CT Angiogram (mCTA) was developed with the aim to provide time-resolved images of the intracranial arteries for quantification of collaterals. mCTA consists of an arterial phase angiogram obtained from the arch of the aorta to the vertex followed by two venous phases (equilibrium/peak venous and delayed/post-venous phases) acquired from the base of the skull to the vertex.[7,8] Postprocessed mCTA thus consists of three gray-scale image datasets of the craniocervical vasculature, with multiplanar reconstructions. The technique needs no additional contrast material injection or special vendor package for postprocessing. Evaluation of collaterals requires the operator to either assess the three phases individually or to manually co-register all the phases and scroll through them simultaneously. It also requires expertise and is often time-consuming which is a major limitation in an already time-sensitive clinical setting.
FastStroke and ColorViz
FastStroke (GE Healthcare, Milwaukee, Wisconsin) is a CT software used in AIS imaging. It provides a comprehensive, progressive workflow for image analysis with autogenerated multiplanar reformats of the CT stroke imaging protocol (unenhanced CT of the brain, mCTA). In addition, it auto-registers all phases of the mCTA side-by-side, to facilitate collateral vessel review, and also generates CT perfusion images for analysis.
The ColorViz application of FastStroke combines vascular information from all phases of mCTA into a single color-coded imaging series. It displays consolidated information of intracranial arterial dynamics which helps quantify the overall and phase specific extent of pial collaterals. Vessels are assigned colors based on the temporal distribution of maximum contrast opacification; the arterial phase is coded red, the venous phase in green, and the delayed phase in blue. The display layout is split into three viewports, showing colorized maximum intensity projections (MIP) in the axial, coronal, and sagittal planes. These color maps are available for review within 5–10 seconds.
Evaluation of pial collaterals
Upon localizing the occlusion, the operator can visually assess pial collaterals using ColorViz by determining the predominant color in the affected vascular territory (color present in >40% of vessels). Predominantly red vessels indicate no delay while green or blue vessel predominance indicate a 1-phase or 2-phase delay, respectively. With a single imaging series, ColorViz combines the visual advantages of color and technical advantages of mCTA. It potentially expedites the assessment of pial collaterals (a few seconds) compared to conventional gray-scale mCTA, which impacts further treatment. Verdolotti et al. used a 3-point grading scale for evaluating collaterals:
“Good” collateral circulation (score 3): Prevalence of red vessels distal to the occlusion with only slightly reduced pial collateral representation [Figure 1].
“Intermediate” collateral circulation (score 2): Equal or slightly reduced collaterals, compared to the unaffected hemisphere, but predominantly green colored vessels [Figure 2].
“Poor” collateral circulation (score 1): Significant reduction in the extent of collaterals (even in the presence of few red/green vessels) or if the prevalent color is blue (suggesting significant delay) [Figure 3].
Using this grading system, Verdolotti et al. found ColorViz to be statistically comparable to mCTA with a good interobserver agreement in the evaluation of collateral circulation. In a retrospective study by Lin et al., a 6-point scoring system, previously proposed by Menon et al., was used for the evaluation of collaterals. They found that color-coded mCTA, using ColorViz, had diagnostic performance equivalent to conventional mCTA, but better than CT Perfusion (CTP), in predicting favorable outcome. In another study to compare the role of mCTA color maps versus conventional mCTA in the prediction of final clinical outcome and infarct volume, Ospel et al. used two trichotomized scales for evaluation of collaterals. Their grading system was an extended version of the scoring used by Verdolotti et al. taking into account collateral extent and collateral filing separately, rather than as a single function. In their study of 285 patients, they found color-coded grading of collateral extent performed similarly to conventional mCTA in predicting 90-day outcome and final infarct volume, but grading for color-coded collateral dynamics did not.
Color-coded displays have been proven to be beneficial in facilitating visual search and recognition tasks in cerebral vascular pathologies.[4,13] The potential value of colored display, and hence the color indicator effect, of ColorViz can be exploited in identifying medium vessel occlusions (MeVO) and distal cortical branch occlusions of the middle cerebral artery (MCA).[9,11] Such occlusions are difficult to identify, especially in an emergent situation, but are relevant as multiple studies have recognized such occlusions to be amenable to thrombolysis therapy and thrombectomy.[14-17] The imaging hallmark of a distal occlusion is the delayed contrast opacification, beyond the site of occlusion. This appearance has been termed the “delayed vessel sign.” It is attributed to retrograde filling via pial collaterals. On ColorViz, this manifests in green and/or blue [Figures 4 and 5].[9,11] Lin et al. found that ColorViz may increase the speed and accuracy of detection of MeVO, which could not be sensitively detected on conventional mCTA images.
The visual impact of color may be extended to the identification of anterior cerebral artery (ACA) occlusion. ACA occlusions are rare and pose a challenge clinically (due to their atypical presentation) as well as radiologically (due to the primary focus on the MCA circulation in AIS). CTP may show perfusion delays in the affected ACA territories which would guide inexperienced readers. However, adequate coverage of the ACA territory is necessary for diagnosis; limited coverage, with the imaging slab often placed at the ganglionic level (as in most older CT scanners), is known to limit the evaluation of the ACA territory.[9,19] Clue to identify an ACA occlusion lies in the lack of symmetric opacification, and hence differences in color, of pericallosal arteries on axial images [Figure 6].[9,19] Though there still remains a lack of high-level evidence from a randomized trial to prove the benefit of EVT in ACA occlusions, continued strategies to improve techniques and thrombectomy devices would soon warrant accurate and precise localization of ACA occlusions for prompt EVT.
Another potential role of ColorViz is the identification of occlusions of the posterior circulation, including the vertebrobasilar system [Figure 7]. Infarcts in the posterior fossa are difficult to identify on unenhanced CT scans and CTP evaluation may be limited due to artifacts from the skull base.[9,20] mCTA, being less susceptible to artifacts as compared to CTP, remains an important tool in identifying posterior circulation occlusions. However, it must be acknowledged that further characterization of collateral status in posterior circulation stroke, except for those of the posterior cerebral artery (PCA), is still not established given its variable vascular anatomy and poorly understood pial collateral supply.[7,8]
A proposed use of the color-indicator effect of ColorViz is in the identification of multivessel occlusions (MVO) in downstream or different vascular territories [Figures 8 and 9]. In a study on 720 patients with large vessel occlusion (LVO) subjected to endovascular therapy, MVO was found in 10.7% of patients, however, MVO was only prospectively identified in 54.5% cases. The presence of MVO may impact treatment decisions, angiographic success, and clinical outcomes. While most studies so far evaluate ColorViz in anterior circulation LVO, it does seem to show promise in MVO detection as well. Such an extended application of ColorViz would need validation with more prospective studies.
ColorViz seems to be promising, but its role in clinical practice still needs to be validated using large-scale prospective clinical trials. The limited published studies so far have been retrospective in nature and have not put ColorViz to the test in an emergent setting of AIS. The true utility of ColorViz in clinical practice would need a prospective, multicentric clinical trials. Moreover, its use in posterior circulation stroke and MVO has been speculated, but not proven. The current FastStroke software does not allow manual change of the color-coding system as per the comfort of the operator. This may be confusing for some who are comfortable with a different color scheme.
Adequate evaluation of occlusion relies on the asymmetry between the affected and affected vascular territory. Since the standard of comparison of pial collaterals is based on the assumption that the contralateral side is unaffected, bilateral occlusions may confound interpretation or even result in a missed occlusion. Global or unilateral delayed arterial opacification due to conditions such as congestive heart failure and proximal critical stenosis, respectively, may cause erroneous interpretation of an occlusion or collateral status based on the prevalence of green and/or blue-colored arteries.[7,9]
The automatic synchronization of all mCTA phases, using FastStroke, and subsequent depiction in a color-coded, single composite view (ColorViz) reduces the time for post-processing and evaluation of mCTA. ColorViz is a powerful addition to stroke workflow enabling quick, intuitive evaluation of pial collaterals. Additionally, its promising potential in the identification of MeVO and MVO as well as posterior circulation occlusions would make it a valuable asset, especially for inexperienced radiologists.
Declaration of patient consent
Patient consent is not required as the patient’s identity is not disclosed or compromised.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.