Keywords
cerebral arteriovenous malformation, angioarchitecture, reporting, consensus
A cerebral arteriovenous malformation (cAVM) is an abnormal tangle of cerebral blood vessels. The consensus document by the Joint Writing Group (JWG) highlighted which cAVM features should be recorded. Subsequent publications have reported cAVM angioarchitecture, but it is unknown if all followed the JWG recommendations.
The aim of this systematic review was to describe use of the JWG guidelines.
A database search, using the PRISMA checklist, was performed. We describe the proportion of publications that used JWG reporting standards, which standards were used, whether the definitions used differed from the JWG, or if any additional angiographic features were reported.
Out of 4306 articles identified, 105 were selected, and a further 114 from other sources.
Thirty-three studies (33/219; 15%) specifically referred to using JWG standards.
Since the JWG publication, few studies have used their standards to report cAVMs. This implies that the angioarchitecture of cAVMs are not routinely fully described.
cerebral arteriovenous malformation, angioarchitecture, reporting, consensus
Some changes have been made to the Methods and Results sections.
See the authors' detailed response to the review by Nitin Mukerji
See the authors' detailed response to the review by Johan Wikström and Maria Correia de Verdier
See the authors' detailed response to the review by Mustafa Ismail
Cerebral arteriovenous malformations (cAVMs) cause death and disability mostly in the young.1 A cAVM anatomically consists of a nidus, that shunts blood directly from the arterial feeders to the draining veins, in the absence of capillaries.2–4 It is best described radiologically using digital subtraction angiography (DSA), with the pathognomonic features being a nidus and early venous drainage.5 The angioarchitecture of a cAVM refers to the vascular elements of a cAVM that are demonstrated on angiogram. It encompasses feeding arteries, nidus, draining veins, vascular changes resulting from high blood flow, and any accompanying abnormal vascular appearances.5
The commonest presenting feature is haemorrhage (occurring in 38-71% of cAVMs), and this contributes the most to cAVM morbidity and mortality.4 The annual rate of cAVM rupture, and consequent intracerebral haematoma (ICH), is 2-4% with the highest risk within the first five years of diagnosis.5 There are multiple ICH risk factors, including young age, prior cAVM rupture, deep and infratentorial location, large cAVM size, and deep venous drainage.4,6 Mortality rates are linked to haemorrhage from ruptured cAVMs in a high proportion of cases (10-40%).5,6
Seizures are the second most common presentation (17-30% of cAVMs) with risk factors including absence of aneurysms, temporal, frontal or cortical locations, varices, middle cerebral artery and cortical feeders, previous cAVM haemorrhage, and male gender.4,7–9
Focal neurological deficits (occurring in 5-15%) may be explained by a vascular steal phenomenon, where high shunting through the cAVM causes a reduced vascular supply in the surrounding parenchyma.4,5 It could also be due to mass effect on vulnerable white matter pathways from compressive venous dilatation.5 Risk factors for focal neurological deficits include older age, female gender, deep location, and ectasia.10
The fact that cAVM treatment is associated with significant morbidity and mortality poses additional challenges. Treatment is currently aimed at shrinking or excising the lesion, usually after it has ruptured or caused symptoms. Management decisions require a case-by-case multidisciplinary discussion balancing risks and benefits: there is no definitive algorithm for management. This is because there is little understanding of the pathophysiology underpinning cAVMs, but also there is a lack of consistency in reporting. This inconsistency poses a significant challenge to the progress of cAVM treatment.
Typically, DSA is used to classify cAVMs. cAVM presentation is believed to rely on its angioarchitecture. The latter may also be used to guide treatment. Since cAVMs have a complex morphology with each malformation being unique, reliably classifying cAVMs is challenging for clinicians managing them.
cAVMs are graded using several methods, including Spetzler-Martin (most widely used), Spetzler-Ponce, Lawton-Young or Flickinger-Pollock. They either classify using anatomical grades, or based on the likelihood of success and treatment risks. Though these grading systems incorporate certain essential information required to aid in management decisions, none of them are very detailed or sufficiently extensive. Particularly, in those cases where the grading score does not provide a definitive answer regarding best management, further detail is important. An in-depth description of cAVM angioarchitecture is also vital for cAVM clinical research, which will contribute to improved patient treatment. Furthermore, although reliability studies have shown good intra-observer agreement on the characterisation of cAVM angioarchitecture, inter-observer agreement was poor.11
In an attempt to address this, a consensus document was published by the Joint Working Group (JWG) of the Technology Assessment Committee: this provides elementary and clear definitions of terms and recommends which clinical and radiological cAVM features should be described and recorded ( Table 1).12 The JWG has been very comprehensive in compiling its list of angioarchitecture definitions. This group, consisting of neuroradiologists, neurosurgeons, stroke and interventional neurologists practising in the United States of America, was created to produce guidelines for cAVM research.12 The work done by the JWG has been very significant in establishing a uniform framework for reporting. Complying with their guidance improves clarity and comparability when reporting cAVMs, and when publishing research results.
Standardising the terminology used would not only facilitate clinical trials, but also day-to-day patient management. A detailed understanding of, for instance, venous drainage would facilitate decision-making by better being able to quantify the risks and benefits of operative vs endovascular vs stereotactic radiosurgery management.
The review protocol was sent for registration to PROSPERO but not accepted due to “a perceived lack of direct impact on patient outcomes”. Reporting was in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement and checklist.13
Peer-reviewed publications were searched from 1 Jan 2001 (when JWG standards were published) and limited to human subjects and the English language. There was no restriction on age or sex. Studies were excluded if they were reviews and/or exclusively discussed cavernous malformations, dural arteriovenous fistula, angioma, capillary telangiectasia, Vein of Galen Malformations or other angiographically occult vascular malformations.
A database search was performed using EMBASE and Medline, on 15/07/19 by one reviewing author (SD). It was repeated on 10/9/20 to update the search by a second author (MW). There was independent assessment by a librarian and another reviewing author (HP). In addition to the electronic searches, we conducted citation tracking, checked the reference lists, and reviewed the list of similar articles.
To conduct searches of the Medline electronic bibliographic database, combinations of the following search terms were used.
Medical Subheadings: (Arteriovenous Malformations OR Arteriovenous Malformations, Intracranial) AND (Brain OR Intracranial)) AND (angioarchitecture OR angiogram OR angiographic OR aneurysm OR venous OR ectasia OR nidus OR angiogenesis OR varix).
Text Words: (Arteriovenous Malformations OR Arteriovenous Malformations, Intracranial) AND (Brain OR Intracranial)) AND (angioarchitecture OR angiogram OR angiographic OR aneurysm OR venous OR ectasia OR nidus).
Studies were selected if they included any of the search strategy’s features: titles and abstracts were reviewed. Figure 1 demonstrates how articles were excluded.
Data extraction was performed by two independent reviewers (SD and HP) using the studies’ full text versions and by reviewing inclusion and exclusion criteria. Any disagreements were discussed between the two reviewers and an agreement reached. Pre-designed and piloted proforma were used.
All the individual data items collected from each paper are listed in Table 1.
Definition/categorisation (if present) | |
---|---|
Proposed fields | |
|
|
|
|
|
|
|
|
Venous drainage | |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Arterial supply | |
|
|
|
|
|
|
|
|
|
|
Risk of bias was determined by two independent reviewers (SD and HP) using a version of a score devised to determine the methodological quality of case series and case reports,14 modified to expand its application to other study types. For each study, the quality assessment questions used were:
• Were the JWG standards used?
• Diagnosis: Are diagnostic criteria (as defined) for cAVM clearly identified and met?
• Is the method of cerebral angiography described, including the arterial injections, views taken, and what structures are included in a standard view?
• Is the method of calibration described?
• Are the cerebral angiograms reported on by two blinded neuroradiologist(s)?
• Were the patients reported collected over a short period of time in sufficient numbers?
• Is intra-rater reliability reported for each publication?
The principal summary measures were the number of studies following the JWG terminology and the angioarchitectural fields described.
This review aimed to assess if the description of cAVM angioarchitecture is standardised according to the JWG criteria and to explore if there are additional features that could be used to describe cAVMs and added to the JWG reporting standards. Additional features could be required to describe cAVMs as shown by some studies. Our objectives were to describe:
1. which of the JWG reporting standards were used,
2. adherence to definitions used in the JWG document,
3. novel angiographic features not mentioned by the JWG, and
4. the profession and experience of those reporting cAVMs.
We also compared the inter-observer agreement of different studies for the common criteria studied.
We identified 4306 publications ( Figure 1). 423 full-text articles were assessed for eligibility and 219 articles were selected for full-text review and included in the study, reporting the angiogram findings potentially for 54, 148 individuals in total.
Most studies were retrospective (63%), with the remainder being prospective (27.9%), case reports (2.7%), studies that were both prospective and retrospective (2.7%), and educational (3.7%).
The studies spanned from 2001 to 2020 (the JWG report was published in 2001). The number of patients in each study ranged from 1 to 3923. The median was 120 (interquartile range: 30 to 278). The countries individually publishing the highest number of studies were China (15.5%) and the United States of America (15.1%). Studies from Western Europe, North America and Asia made up 26.5% of studies, with other countries (publishing fewer studies, i.e. one to four studies each) making up 69.9%. The countries from Western Europe were France, UK, Spain, Netherlands, Norway, Italy, Germany, Finland, Belgium, Poland, and Switzerland.
Beijing Tiantan Hospital published more papers than any other centre, with data from this single centre contributing to 26 (74.3%) of the studies from China and 11.9% overall. Several studies used the same study populations, occasionally with a few more cases added due to an extension of study duration by a few years ( Table 2). Out of Beijing Tiantan’s 26 studies, the commonest author groups were Lv et al and Tong et al (five studies for each author group) with the same sample population used three times for each of these two author groups ( Table 2). The University of California (San Francisco) (18; 8.2%), University of Virginia (Charlottesville) (7; 3.2%), and Columbia University (New York) (6; 2.7%) were the most frequently publishing American institutions.
Author group with overlapping study populations | Number of studies (%) | Mean sample size |
---|---|---|
Beijing Tiantan Hospital | ||
Lv, Wu, Jiang, Yang, Li, Sun, Zhang67,182,184 | 3 (1.4) | 302 |
Ma, Kim, Chen, Wu, Ma, Su, Zhao65,66 | 2 (0.9) | 108 |
Tong, Wu, Lin, Cao, Zhao, Wang, Zhang, Zhao64,185, 77 | 3 (1.4) | 225 |
The First Affiliated Hospital of Guangzhou Medical University | ||
Pan, Feng, Vinuela, He, Wu, Zhan69,111 | 2 (0.9) | 152 |
University of California, San Francisco | ||
Du, Dowd, Johnston, Young, Lawton99,187 | 2 (0.9) | 304 |
John Hopkins University, Baltimore | ||
Yang, Liu, Hung, Braileanu, Wang, Caplan, Colby, Coon, Huang76,90 | 2 (0.9) | 123 |
University of Virginia, Charlottesville | ||
Ding, Starke, Quigg, Yen, Xu, Przybylowski, Dodson, Sheehan78,188–190 | 4 (1.8) | 1400 |
University of Illinois, Urbana-Champaign | ||
Shakur, Valyi-Nagy, Amin-Hanjani, Ya qoub, Aletich, Charbel, Alaraj71,72,152,179 | 3 (1.4) | 80 |
Columbia University, New York | ||
Stapf, Mohr, Pile-Spellman, Sciacca, Hartmann, Schumacher, Mast71,72,152,179 | 4 (1.8) | 542 |
Hospital Lariboisiere, Paris | ||
Choi, Mast, Hartmann, Marshall, Stapf177,178 | 2 (0.9) same study population as Stapf et al | 735 |
Given the clinical importance of predicting haemorrhage, the aim of 65 papers (29.7%) was to test for associations between angioarchitecture and risk of bleeding (including location).1,15–78 Twelve studies tested for an association between angioarchitecture and risk of seizure.79–90
Standard imaging was compared against novel imaging techniques in four studies, and pre-operative imaging was investigated in two studies.91–96 Three studies assessed haemodynamics.37,97,98
Eleven papers studied various grading scores.91,99–108 The Spetzler-Martin Grade was the most commonly analysed, but others included the Spetzler-Ponce, Lawton-Young, and Pollock-Flickinger. Out of these 11 studies, seven assessed and proposed different grading systems.101–105,107,108
Four papers assessed the reliability of different cAVM grading scales, and two studies assessed the reliability in describing cAVM angioarchitecture.11,91,99,100,106,109 Agreement ranged from fair to excellent for both inter- and intra-rater comparisons.
Angioarchitectural characteristics were reported in association with treatments: embolisation,26,52,55,101,103,109–138 surgery,15,95,110,119,120,122,123,126,139–154 and stereotactic radiosurgery.53,83,97,60,113,119,122,123,126,129,140,145,155–161
In those studies that assessed inter-observer agreement, the commonest criteria investigated were size, SMG, venous drainage, and arterial feeders. The lowest scores were 0.62,91 0.46,91 0.56,92 and 0.695 respectively. The highest scores were 0.98,162 0.96,163 0.89,164 and 0.91164 respectively.
Overall, the quality of the reporting studies was poor, with several of the study quality criteria not fulfilled (Table 3). Only 48 out of 219 studies (21.9%) used the definitions recommended by the JWG for some of the features reported ( Table 3),16,19,20,23,24,26,28–30,33,35,38,47,49,55–57,61,64,66,67,71,72,75,77,90,152,156,159,165–181 with only 33 publications (15.1%) reporting and specifically mentioning using the JWG standards. Out of the ‘Western’ papers that provide a detailed angioarchitecture description, 21 publications (18.8%) used the JWG standards.
The list excludes studies which do not describe angioarchitecture in detail e.g. only report location. White boxes indicate criteria fulfilled; dark grey boxes indicate criteria absent. a = JWG standard used; b = cAVM diagnostic criteria; c = DSA method: arterial injections; d = DSA method: views; e = Calibration method; f = Inter-rater reliability assessed; g = Statistics performed, including collecting sufficient numbers in a short period of time. None of the studies reported intra-rater reliability.
Study author | a | b | c | d | e | f | g |
---|---|---|---|---|---|---|---|
Abla 201474 | |||||||
Abecassis194 | |||||||
Al-Shahi11 | |||||||
Al-Tamimi195 | |||||||
Alen196 | |||||||
Alexander16 | |||||||
Anderson197 | |||||||
Benson88 | |||||||
Bharatha198 | |||||||
Blanc137 | |||||||
Braileanu199 | |||||||
Brunozzi 2017200 | |||||||
Brunozzi 2019166 | |||||||
Burkhardt176 | |||||||
Chang201 | |||||||
Chen202 | |||||||
Choi 2006177 | |||||||
Choi 2009178 | |||||||
Chowdhury168 | |||||||
Cuong93 | |||||||
D'Aliberti45 | |||||||
De Blasi203 | |||||||
de Castro-Afonso204 | |||||||
Guo46 | |||||||
Halim 200439 | |||||||
Halim 200247 | |||||||
Hernesniemi211 | |||||||
Hetts212 | |||||||
Hofmeister173 | |||||||
Huang32 | |||||||
Hung 201949 | |||||||
Iancu-Gontard109 | |||||||
Illies25 | |||||||
Imbesi213 | |||||||
Iryo214 | |||||||
Jayaraman 2012115 | |||||||
Jiang215 | |||||||
Jiao102 | |||||||
Jin24 | |||||||
Kakizawa216 | |||||||
Kandai19 | |||||||
Kellner17 | |||||||
Khaw29 | |||||||
Kim 200433 | |||||||
Kim 200750 | |||||||
Kim 201430 | |||||||
Kouznetsov217 | |||||||
Dinc 201942 | |||||||
Dinc 2018205 | |||||||
Ding 201736 | |||||||
Ding 201589 | |||||||
Dos Santos170 | |||||||
Du 2005206 | |||||||
Du 2016207 | |||||||
Ellis20 | |||||||
Fierstra87 | |||||||
Fleetwood208 | |||||||
Fok23 | |||||||
Frisoli106 | |||||||
Fukuda 2016209 | |||||||
Fukuda 2017171 | |||||||
Fullerton56 | |||||||
Galletti85 | |||||||
Garcin172 | |||||||
Gauvrit92 | |||||||
Geibprasert210 | |||||||
Griessenauer100 | |||||||
Kubalek63 | |||||||
Kurita218 | |||||||
Lee219 | |||||||
Liew220 | |||||||
Lin221 | |||||||
Liu 201582 | |||||||
Luo 2012183 | |||||||
Lv 2013182 | |||||||
Lv 2011a184 | |||||||
Lv 2011b67 | |||||||
Lv 2015222 | |||||||
Ma 2017a65 | |||||||
Ma 2017b66 | |||||||
Ma 201573 | |||||||
Majumdar57 | |||||||
Miyasaka40 | |||||||
Morgan 2016140 | |||||||
Motebejane223 | |||||||
Neidert104 | |||||||
Nishino224 | |||||||
Nisson 2020105 | |||||||
Niu22 | |||||||
Ognard91 | |||||||
Orning44 | |||||||
Oulasvirta225 | |||||||
Ozyurt162 | |||||||
Pan 201369 | |||||||
Patel226 | |||||||
Pawlikowska35 | |||||||
Pekmezci227 | |||||||
Reitz21 | |||||||
Riordan54 | |||||||
Robert 2014228 | |||||||
Robert 2017103 | |||||||
Robert 2015136 | |||||||
Sahlein61 | |||||||
Schmidt48 | |||||||
Schwartz229 | |||||||
Shakur 2016a41 | |||||||
Shakur 2016b98 | |||||||
Shakur 2018230 | |||||||
Shakur 2015231 | |||||||
Shankar79 | |||||||
Sheng232 | |||||||
Shotar108 | |||||||
Singh94 | |||||||
Stapf 2003179 | |||||||
Stapf 2002b180 | |||||||
Stapf 200672 | |||||||
Stefani 200168 | |||||||
Stefani 200270 | |||||||
Stein 2016b123 | |||||||
Stein 2015181 | |||||||
Sturiale58 | |||||||
Suzuki96 | |||||||
Tanaka144 | |||||||
Taschner164 | |||||||
Tasic59 | |||||||
Todaka27 | |||||||
Togao 2019233 | |||||||
Togao 2020163 | |||||||
Tong 2016a185 | |||||||
Tong 2016b64 | |||||||
Tong 2016c77 | |||||||
Tritt234 | |||||||
Tsuchiya95 | |||||||
Unlu235 | |||||||
Wrede236 | |||||||
Yamada1 | |||||||
Yang 2016b38 | |||||||
Yang 201743 | |||||||
Yang 2016a76 | |||||||
Yang 2015b90 | |||||||
Yang 2015a175 | |||||||
Ye237 | |||||||
Yi167 | |||||||
Yu31 | |||||||
Zwanzger238 |
Biases in the studies were because there was a small population size (less than 100 cases) in 100 studies, a second professional did not independently review angiograms in any of the studies, and there was often a re-analysis of datasets.
The common angioarchitectural features are listed with the associated number of studies ( Figure 2). Most studies described nidus size (175 studies; 78%), location (153; 68%), border (29; 12.9%), venous drainage (173; 76.9%), feeding arteries (88; 39.1%), and the presence of aneurysms (121; 53.8%). No studies described the angiographic features of pial to pial collaterals or Moya-Moya type changes as recommended by the JWG.
Many studies used a variety of the recommended angiographic features, though not necessarily defining these features in the same way as the JWG ( Table 4).
They may have different definitions for these features compared to that stipulated by the JWG.
Almost all the features described using the JWG guidelines were given different definitions, including, type of feeders, arterial feeders, and haemorrhagic presentation.
cAVM location was listed by the JWG in a table, but not described, and this feature had the largest range of definitions. Some specified what constitutes deep, cortical, and/or infratentorial1,24,32,67,182,183 or dichotomised location into supratentorial and infratentorial.29,147,184,185 There were further categorisations into posterior fossa and periventricular by Ma et al.174
Venous ectasia was the feature with the second-most variations of definition. Whereas the JWG defines it as double the calibre change in any draining venous channel, others have described it as 1.5 times larger than the contralateral vessel,186 and two papers are broader in their definitions, describing venous ectasia as a markedly ectatic vein,107 or an abnormal dilatation.68,70
Aneurysms were defined as a saccular luminal dilatation of parent feeding vessels by the JWG. Most papers have essentially stated the aneurysm should be double the width of the artery, with only one definition stating the diameter is at least the same as that of the parent vessel.28
Numerous studies described angioarchitectural features which were not mentioned in the JWG report and these are described in Table 5. These features included perinidal angiogenesis, AVM nidus, deep location, and venous varix/pouch.
We have shown that only 33 studies of 219 (15.1%) included in our systematic review explicitly followed the JWG standards since their publication 20 years ago.12 Out of the ‘Western’ papers that describe angioarchitecture in detail, 21 publications (18.8%) used the JWG standards. Additionally, most studies reported venous drainage (76.9%), cAVM size (78%), and cAVM location (68%), suggesting these features are frequently considered as cAVM angioarchitecture. These parameters were the most widely used, likely due to their relation to the SMG system.
Since 219 publications were reviewed as providing data on angioarchitecture, it appears that this topic is considered important. Most commonly angioarchitectural features were used to test for associations with outcomes relevant to cAVM such as haemorrhage.
In those studies that assessed inter-observer agreement, the criteria most frequently used for comparison were size, SMG, venous drainage, and arterial feeders. It is possible that the other criteria were less used as they were more difficult to analyse on imaging.
Often, certain cAVM features are not reported as they are not present e.g. absence of venous stenosis. However, we argue that important negative findings should be mentioned in all cAVM reports. Adherence to the JWG guidelines will permit more comprehensive cAVM reporting, which will facilitate improved decision-making.
Twenty years have passed since the publication of the JWG definitions and, not unsurprisingly, several papers have reported on additional aspects of angioarchitecture which the JWG had not considered. These additional features may be helpful in understanding cAVMs and consideration should be given for their inclusion in any future update. These features are perinidal angiogenesis, deep location, venous and arterial dilatation. Angiogenesis is important for the formation and development of a cAVM and its presence may be useful in surgical planning.135 The precise location of a cAVM is crucial with well accepted definitions key for a shared understanding when discussing patient management. Venous dilatation is helpful to describe as it indicates if there may be high or low-pressure flow in the cAVM, with a larger vein reducing the pressure in a cAVM.45,69,161,174,182,183,206 This would be relevant to decide on the management approach. Equally, arterial dilatation29,86,160,186 may imply high-pressure flow in a cAVM, particularly if combined with a single vein of regular dimensions and may have clinical implications.
In this review, we also observed that pial-pial collaterals and Moya-moya changes were not recorded. This may reflect the difficulty in identifying these features, but also may suggest that they occur infrequently.
Including the JWG criteria in studies on cAVM angioarchitecture enhances the academic rigour and credibility of research publications. Few studies specifically mentioned adhering to the JWG guidelines, and in those that did not, no reasons were given for omitting them. A possible practical reason was the lack of technical equipment. Facilities will vary in different countries, including the type of biplane machine for angiograms. The widespread use of the JWG standards will have good implications for multi-centre studies and longitudinal research. It was not possible to include perspectives from various stakeholders like clinicians, researchers, and healthcare policymakers, as the few studies that made reference to the JWG, do not make any comments or provide any feedback regarding the JWG.
Given that overall, the technical quality of publications was low, that most studies were retrospective and from small single centre series, the validity of results from these series could be questioned. Data reported from larger series also lacked the full consideration of angioarchitecture and often the same dataset was used for association studies again compromising the associations reported. In addition, as a large proportion of studies were published by single institutions, the results may not be generalisable to other cAVM populations. A similar problem was that a high proportion of publications were based on Chinese populations, particularly conducted by a specific hospital (Beijing Tiantan hospital), making results less generalisable. There was therefore potential publication bias. It may be too harsh to expect reporting by two independent neuroradiologists. There would certainly be variations in the access to technology across different health systems around the world, which would pose additional practical challenges to the universal application of the JWG standards.
The JWG publication did clarify that the definitions were parameters to be used in research studies.19 They have also discussed that there were no minimal criteria that should be used, emphasising that the angioarchitectural criteria were based on reasoned speculation. However, given that many of the criteria are likely to be interdependent, and studies are increasingly used to show associations with clinical presentation, this review would support the need to establish another working group to incorporate additional angiographic features and to include more specific and precise definitions for some of the features that were left open to interpretation. We would argue that these recommendations should then be widely publicised and uniformly incorporated into national and local reporting guidelines to help guide research and to ensure that clinicians can appropriately interpret this research with the understanding of the common language.
This review was not registered as described above. The review protocol can be accessed on BioStudies. The link is https://www.ebi.ac.uk/biostudies/studies/S-BSST1168. The accession number is S-BSST1168.
BioStudies. Review of angioarchitecture literature SD. DOI: https://www.ebi.ac.uk/biostudies/studies/S-BSST1168BioStudies.
The raw data is in two files titled:
- ‘Review of angioarchitecture literature SD’ and
- ‘Review of angioarchitecture literature quality SD’.
The content is raw data from the papers used in the systematic review. The accession number is S-BSST1168.
Biostudies. PRISMA checklist. DOI: https://www.ebi.ac.uk/biostudies/studies/S-BSST1168. The flow diagram is in Figure 1.
The authors would like to thank the Natalie Kate Moss Trust for their generous support. PK and AP-J are supported by the Stroke Association (TSA LECT 2017/02; SA L-RC 19\100000).
Views | Downloads | |
---|---|---|
F1000Research | - | - |
PubMed Central
Data from PMC are received and updated monthly.
|
- | - |
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Neurosurgery
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: MRI of the brain.
Are the rationale for, and objectives of, the Systematic Review clearly stated?
Yes
Are sufficient details of the methods and analysis provided to allow replication by others?
Yes
Is the statistical analysis and its interpretation appropriate?
Not applicable
Are the conclusions drawn adequately supported by the results presented in the review?
Yes
If this is a Living Systematic Review, is the ‘living’ method appropriate and is the search schedule clearly defined and justified? (‘Living Systematic Review’ or a variation of this term should be included in the title.)
Not applicable
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: MRI of the brain.
Are the rationale for, and objectives of, the Systematic Review clearly stated?
Partly
Are sufficient details of the methods and analysis provided to allow replication by others?
Yes
Is the statistical analysis and its interpretation appropriate?
Not applicable
Are the conclusions drawn adequately supported by the results presented in the review?
Yes
If this is a Living Systematic Review, is the ‘living’ method appropriate and is the search schedule clearly defined and justified? (‘Living Systematic Review’ or a variation of this term should be included in the title.)
Yes
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Vascular Neurosurgery, General Neurosurgery,
Are the rationale for, and objectives of, the Systematic Review clearly stated?
Yes
Are sufficient details of the methods and analysis provided to allow replication by others?
Yes
Is the statistical analysis and its interpretation appropriate?
Partly
Are the conclusions drawn adequately supported by the results presented in the review?
Partly
Competing Interests: No competing interests were disclosed.
Reviewer Expertise: Cerebrovascular Neurosurgery
Alongside their report, reviewers assign a status to the article:
Invited Reviewers | |||
---|---|---|---|
1 | 2 | 3 | |
Version 3 (revision) 25 Nov 24 |
read | read | |
Version 2 (revision) 07 Oct 24 |
read | ||
Version 1 29 Sep 23 |
read | read |
Provide sufficient details of any financial or non-financial competing interests to enable users to assess whether your comments might lead a reasonable person to question your impartiality. Consider the following examples, but note that this is not an exhaustive list:
Sign up for content alerts and receive a weekly or monthly email with all newly published articles
Already registered? Sign in
The email address should be the one you originally registered with F1000.
You registered with F1000 via Google, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Google account password, please click here.
You registered with F1000 via Facebook, so we cannot reset your password.
To sign in, please click here.
If you still need help with your Facebook account password, please click here.
If your email address is registered with us, we will email you instructions to reset your password.
If you think you should have received this email but it has not arrived, please check your spam filters and/or contact for further assistance.
Comments on this article Comments (0)