Absolutely useless! Nothing here gets survivors better recovery! I'd have you all fired!
Use of Thick Maximum‐Intensity Projection Brain Computed Tomography Angiography for Evaluation of Baseline Collateral Status Improves Interrater Agreement
Abstract
BACKGROUND
In acute ischemic stroke caused by large‐vessel occlusion, tissue viability is dependent on the blood supply from leptomeningeal collaterals until reperfusion is achieved. Rapid and accurate evaluation of baseline collateral status is a key marker of eligibility for endovascular therapy but can be challenging to interpret using source images of the computed tomography angiography (SI‐CTA). Our objective was to assess whether the use of thick maximum‐intensity projection computed tomography angiography (MIP‐CTA) improves interrater agreement for evaluation of baseline collaterals status between stroke trainees and an expert stroke neurologist.
METHODS
An expert stroke neurologist and 2 stroke trainees independently reviewed images from 40 brain CTA scans with anterior circulation large‐vessel occlusion and assessed collateral status using the Tan collateral scoring system using SI‐CTA in the first reading and then using MIP‐CTA in the second reading. We calculated interrater agreement and recorded the total time needed in each reading.
RESULTS
Interrater agreement was fair between the 2 stroke fellows and stroke expert when using SI‐CTA (κ=0.45 with 52.5% agreement). After using MIP‐CTA, interrater agreement improved to moderate (κ=0.69 with 70% agreement). The median reading time was 1.89 minutes per scan using SI‐CTA and 1.00 minute per scan using MIP‐CTA (P<0.0001).
CONCLUSIONS
We show that using MIP‐CTA, when compared with SI‐CTA, shortens interpretation time and improves interrater agreement between stroke trainees and a stroke imaging expert for the evaluation of baseline collaterals in patients presenting with anterior circulation large‐vessel occlusion.
Endovascular therapy (EVT) is the standard of care for selected patients with acute ischemic stroke (AIS) secondary to large‐vessel occlusion (LVO).1 Baseline collateral status has been shown to influence the outcomes after EVT.2, 3 In patients with LVO, the rate of change in the size of the core and the penumbra is highly influenced by blood flow provided by collateral vessels.4 Recently, MR CLEAN‐LATE trial showed that in patients with anterior circulation LVO stroke presenting in the late time window, EVT is beneficial if good collateral flow is evident on computed tomography angiography (CTA) source images.5
Evaluating the status of the collateral vessels on the source images of the CTA (SI‐CTA) can be challenging, especially when done by trainees or inexperienced physicians. Thick maximum‐intensity projection (MIP) reconstruction uses computed tomography thin slices to generate a 2‐dimensional “stacked” view of the structures on CTA and can visualize a large portion or the full extent of arterial and venous structures on 1 or few images without the addition of contrast or special imaging software.6, 7
It is not known if the addition of MIP imaging to standard CTA will improve visualization and scoring of the collateral status in acute stroke. The aim of this study was to examine whether the use of MIP‐CTA improves interrater reliability and agreement between stroke trainees and an expert stroke neurologist in the assessment of baseline collateral status compared with SI‐CTA in patients with AIS secondary to anterior circulation LVO.
METHODS
We retrospectively analyzed data from the RITE (Riyadh Thrombectomy) registry. The RITE registry is an ongoing prospective citywide registry for EVT of patients with ischemic stroke in Riyadh, Saudi Arabia. The local ethics board approved this study. The data that support the findings of this study are available on request from the corresponding author.
Representative patients from the RITE registry with clinical diagnosis of AIS determined by a focal neurologic deficit measurable on the National Institutes of Health Stroke Scale were included in the study if they presented within 24 hours from last seen normal. Inclusion criteria for the present study were as follows: (1) aged >18 years and (2) had baseline CTA that documented proximal intracranial occlusion of the M1 segment of middle cerebral artery (MCA)±intracranial internal carotid artery (ICA) involvement.
For the present study, we selected representative baseline angiographic images from 40 randomly selected patients with AIS enrolled in the RITE registry. Patients included had a wide variety of collateral statuses, ranging from no collateral flow to excellent collateral flow. A standard aortic arch to vertex CTA was performed using a multidetector computed tomography scanner.
Nonstandard Abbreviations and Acronyms
| AIS | acute ischemic stroke |
| EVT | endovascular therapy |
| LVO | large‐vessel occlusion |
| MIP‐CTA | maximum‐intensity projection computed tomography angiography |
| SI‐CTA | source images of the computed tomography angiography |
CLINICAL PERSPECTIVE
Baseline collateral status in the context of ischemic stroke caused by large‐vessel occlusion has been shown to impact the outcome of endovascular therapy, which makes rapid and accurate evaluation of collateral status an important tool in acute stroke care.
Comparing the use of source imaging computed tomography angiography with maximum‐intensity projection computed tomography angiography in frontline stroke trainees has not been evaluated previously.
The use of maximum‐intensity projection computed tomography angiography results in more accurate and faster evaluation of collateral status when compared with the use of source image data. This may have a positive impact on workflow and care for patients with acute ischemic stroke with large‐vessel occlusion.
Baseline imaging was analyzed in the central imaging core laboratory at the King Faisal Specialist Hospital and Research Center using picture archiving and communication system (PACS) (Sectra Workstation IDS7; Linköping, Sweden).
The 3 readers independently reviewed images in a random order, blind to all demographic and outcome data: an expert stroke neurologist with 10 years’ experience and 2 neurology‐trained stroke fellows with 1 year of experience assessed all scans.
The study consisted of 2 phases 1 week apart. In phase 1, readers were tasked with assessing the Tan score using axial plane SI‐CTA (0.6‐mm slices). In phase 2, readers were tasked with assessing the same images in a rearranged order using MIP‐CTA. MIP reconstruction is a software‐based reconstruction of CTA source images that takes ≈3 to 5 seconds to complete using our institution's imaging PACS system. Previous studies used 20‐ and 24‐mm‐thick MIP to evaluate pial arterial filling.8, 9 On the basis of these studies, we believe that MIP thickness between 20 and 24 mm is likely the ideal range to evaluate collateral blood flow. Hence, we chose an MIP thickness of 22 mm thick to evaluate pial arterial filling. For our study, the total time to complete each phase of image evaluation was recorded. This was used to calculate the average time to evaluate each image.
The Tan collateral scoring system is a scale of 0 to 3 that grades back‐filling pial arteries (Figure 1).8 Case examples for individual pial collaterals, as measured by the Tan score, with SI‐CTA and MIP‐CTA are shown in Figure 2.
Figure 1. Illustrative examples of collateral flow visualized using computed tomography angiography (CTA) across all Tan scores from 0 to 3.
The illustration shows a left‐sided M1 occlusion with color‐coded collateral vessels, where green indicates collaterals in the unaffected hemisphere, and red indicates collaterals in the affected hemisphere. Collateral flow status in the left middle cerebral artery territory is graded as 0 (0% collateral filling), grade 1 (>1%–≤50% collateral filling), grade 2 (>50%–≤99% collateral filling), or grade 3 (100% collateral filling) and scored in comparison to the contralateral middle cerebral artery (MCA) territory collateral status.
Figure 2. Case examples of the collateral status visual assessment for a patient with left‐sided M1 occlusion (red arrows).
Upper panel shows a source image computed tomography angiography (SI‐CTA) using 0.6‐mm slices of CTA raw data, and the lower panel shows a maximum‐intensity projection of the CTA image (MIP‐CTA) using axial plane using 22‐mm‐thick MIP in the same case.
Statistical Analysis
The level of agreement between each of the 2 stroke fellows and the expert stroke neurologist for the collateral scoring of the CTA was measured by the Cohen κ statistic. The interrater agreement level was also expressed as a percentage of agreement. A Cohen κ statistic=0 to 0.20 is considered slight; κ=0.21 to 0.40 is considered fair; κ=0.41 to 0.60 is considered moderate; κ=0.61 to 0.80 is considered substantial; and κ=0.81 to 1 is considered perfect.10 The 95% confidence limits for Cohen κ were also calculated. The total time taken by the 2 fellows to interpret the scans in phase 1 and phase 2 was compared using the Mann‐Whitney test. A 2‐sided P<0.05 was considered significant. The statistical analysis was performed using Stata Statistical Software: Release 17 (2021).
RESULTS
There were 40 CTAs and MIP‐CTAs available for analysis in phase 1 and phase 2. Of these 40 scans, 32 had M1 occlusion and 8 had carotid terminus occlusion. The interrater agreement for accurate detection of pial collaterals was fair with the use of SI‐CTAs only (κ=0.45 with 52.5% agreement). This improved to moderate (κ=0.69 with 70% agreement) when MIP‐CTAs were used (Table 1).
| Raters | Agreement, % | κ | Upper 95% CL | Lower 95% CL |
|---|---|---|---|---|
| Phase 1 | ||||
| ST 1 and SE | 52.5 | 0.18 | −0.05 | 0.41 |
| ST 2 and SE | 52.5 | 0.18 | −0.05 | 0.41 |
| ST 1 and 2 and SE | 52.5 | 0.45 | 0.31 | 0.58 |
| Phase 2 | ||||
| ST 1 and SE | 77.50 | 0.66 | 0.45 | 0.86 |
| ST 2 and SE | 77.5 | 0.65 | 0.43 | 0.86 |
| ST 1 and 2 and SE | 70.0 | 0.69 | 0.57 | 0.82 |
The median time to complete collateral scoring interpretation time was 1.89 minutes per scan when the collateral scoring was done on SI‐CTA and 1.00 minute per scan when the collateral scoring was done on MIP‐CTA (P<0.0001).
DISCUSSION
To our knowledge, the usefulness of MIP‐CTA in the detection of collateral scores has not been previously studied. Our findings show that using MIP‐CTA shortens interpretation time and improves interrater agreement for evaluation of baseline collaterals in patients with LVO. MIP‐CTA reconstruction is a fast and easy procedure that led to improvement in the interrater agreement between stroke trainees and stroke experts in the assessment of collateral status, improves the accurate detection of collateral status in the emergency setting, and can be helpful in making treatment decisions. Furthermore, shortening interpretation time in the evaluation baseline collaterals in the context of acute stroke management may improve workflow and shorten time needed to evaluate radiological images in the acute setting. This may lead to faster administration of appropriate therapy, which, in turn, because of the time‐sensitive nature of acute stroke treatment, may lead to improved outcome in patients with stroke.
ESCAPE (Endovascular Treatment for Small Core and Anterior Circulation Proximal Occlusion With Emphasis on Minimizing CT to Recanalization Times) and MR CLEAN‐LATE (Endovascular Treatment Versus No Endovascular Treatment After 6‐24 Hours in Patients With Ischemic Stroke and Collateral Flow on CT Angiography) trials showed that EVT was efficacious and safe for patients with AIS secondary to anterior circulation LVO who were selected on the basis of the presence of collateral flow on CTA up to 24 hours from time last known well.5, 11
Our study has important limitations. First, scoring was done by readers who had not undergone structured collateral scoring training. Nevertheless, it is likely to represent daily practice where acute stroke code imaging is usually interpreted by a stroke team without formal training. Second, readers did their evaluations in a research environment during working hours. It is uncertain if the performance may be similar in real acute stroke codes and during night on calls. Third, because each reader evaluated each scan just once, intrarater reliability was not considered in this study. Last, our findings are limited to anterior circulation LVO and should not be generalized to posterior circulation stroke or patients with medium or distal vessel occlusions.
In conclusion, MIP‐CTA is a simple tool to evaluate the baseline collateral status that requires no special software or additional administration of contrast material. The inclusion of the MIP‐CTA in AIS imaging protocols is likely to improve interrater agreement and workflow.
Sources of Funding
None.
Disclosures
None.
Acknowledgments
The authors would like to thank their patients for their participation in this study.
Footnotes
REFERENCES
- 1 Powers WJ, Rabinstein AA, Ackerson T, Adeoye OM, Bambakidis NC, Becker K, Biller J, Brown M, Demaerschalk BM, Hoh B, et al. 2018 Guidelines for the early management of patients with acute ischemic stroke: a guideline for healthcare professionals from the American Heart Association/American Stroke Association. Stroke. 2018; 49:e46‐e110. doi:10.1161/STR.0000000000000158LinkGoogle Scholar
- 2 Berkhemer OA, Jansen IG, Beumer D, Fransen PS, van den Berg LA, Yoo AJ, Lingsma HF, Sprengers ME, Jenniskens SF, Lycklama ANGJ, et al. Collateral status on baseline computed tomographic angiography and intra‐arterial treatment effect in patients with proximal anterior circulation stroke. Stroke. 2016; 47:768‐776. doi:10.1161/STROKEAHA.115.011788LinkGoogle Scholar
- 3 Gensicke H, Al‐Ajlan F, Fladt J, Campbell BCV, Majoie C, Bracard S, Hill MD, Muir KW, Demchuk A, San Roman L, et al. Comparison of three scores of collateral status for their association with clinical outcome: the HERMES collaboration. Stroke. 2022; 53:3548‐3556. doi:10.1161/STROKEAHA.122.039717LinkGoogle Scholar
- 4 Liebeskind DS. Collateral circulation. Stroke. 2003; 34:2279‐2284. doi:10.1161/01.STR.0000086465.41263.06LinkGoogle Scholar
- 5 Olthuis SGH, Pirson FAV, Pinckaers FME, Hinsenveld WH, Nieboer D, Ceulemans A, Knapen R, Robbe MMQ, Berkhemer OA, van Walderveen MAA, et al. Endovascular treatment versus no endovascular treatment after 6–24 h in patients with ischaemic stroke and collateral flow on CT angiography (MR CLEAN‐LATE) in The Netherlands: a multicentre, open‐label, blinded‐endpoint, randomised, controlled, phase 3 trial. Lancet. 2023; 401:1371‐1380. doi:10.1016/S0140‐6736(23)00575‐5CrossrefGoogle Scholar
- 6 Fishman EK, Ney DR, Heath DG, Corl FM, Horton KM, Johnson PT. Volume rendering versus maximum intensity projection in CT angiography: what works best, when, and why. Radiographics. 2006; 26:905‐922. doi:10.1148/rg.263055186CrossrefGoogle Scholar
- 7 Harvey EC, Feng M, Ji X, Zhang R, Li Y, Chen GH, Li K. Impacts of photon counting CT to maximum intensity projection (MIP) images of cerebral CT angiography: theoretical and experimental studies. Phys Med Biol. 2019; 64:185015. doi:10.1088/1361‐6560/ab32feGoogle Scholar
- 8 Tan JC, Dillon WP, Liu S, Adler F, Smith WS, Wintermark M. Systematic comparison of perfusion‐CT and CT‐angiography in acute stroke patients. Ann Neurol. 2007; 61:533‐543. doi:10.1002/ana.21130CrossrefMedlineGoogle Scholar
- 9 Menon BK, d'Esterre CD, Qazi EM, Almekhlafi M, Hahn L, Demchuk AM, Goyal M, Multiphase CT. Angiography: a new tool for the imaging triage of patients with acute ischemic stroke. Radiology. 2015; 275:510‐520. doi:10.1148/radiol.15142256CrossrefMedlineGoogle Scholar
- 10 Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics. 1977; 33:159‐174.CrossrefMedlineGoogle Scholar
- 11 Goyal M, Demchuk AM, Menon BK, Eesa M, Rempel JL, Thornton J, Roy D, Jovin TG, Willinsky RA, Sapkota BL, et al. Randomized assessment of rapid endovascular treatment of ischemic stroke. N Engl J Med. 2015; 372:1019‐1030. doi:10.1056/NEJMoa1414905CrossrefMedlineGoogle Scholar
eLetters
eLetters should relate to an article recently published in the journal and are not a forum for providing unpublished data. Comments are reviewed for appropriate use of tone and language. Comments are not peer-reviewed. Acceptable comments are posted to the journal website only. Comments are not published in an issue and are not indexed in PubMed. Comments should be no longer than 500 words and will only be posted online. References are limited to 10. Authors of the article cited in the comment will be invited to reply, as appropriate.
Comments and feedback on AHA/ASA Scientific Statements and Guidelines should be directed to the AHA/ASA Manuscript Oversight Committee via its Correspondence page.