Volume measurement of cerebral arteriovenous malformations from angiography

Application of artificial intelligence in brain arteriovenous malformations: Angioarchitectures, clinical symptoms and prognosis prediction

BACKGROUND

Artificial intelligence (AI) has rapidly advanced in the medical field, leveraging its intelligence and automation for the management of various diseases. Brain arteriovenous malformations (AVM) are particularly noteworthy, experiencing rapid development in recent years and yielding remarkable results. This paper aims to summarize the applications of AI in the management of AVMs management.

METHODS

Literatures published in PubMed during 1999-2022, discussing AI application in AVMs management were reviewed.

RESULTS

AI algorithms have been applied in various aspects of AVM management, particularly in machine learning and deep learning models. Automatic lesion segmentation or delineation is a promising application that can be further developed and verified. Prognosis prediction using machine learning algorithms with radiomic-based analysis is another meaningful application.

CONCLUSIONS

AI has been widely used in AVMs management. This article summarizes the current research progress, limitations and future research directions.

Segmentation techniques of brain arteriovenous malformations for 3D visualization: a systematic review

BACKGROUND

Visualization, analysis and characterization of the angioarchitecture of a brain arteriovenous malformation (bAVM) present crucial steps for understanding and management of these complex lesions. Three-dimensional (3D) segmentation and 3D visualization of bAVMs play hereby a significant role. We performed a systematic review regarding currently available 3D segmentation and visualization techniques for bAVMs.

METHODS

PubMed, Embase and Google Scholar were searched to identify studies reporting 3D segmentation techniques applied to bAVM characterization. Category of input scan, segmentation (automatic, semiautomatic, manual), time needed for segmentation and 3D visualization techniques were noted.

RESULTS

Thirty-three studies were included. Thirteen (39%) used MRI as baseline imaging modality, 9 used DSA (27%), and 7 used CT (21%). Segmentation through automatic algorithms was used in 20 (61%), semiautomatic segmentation in 6 (18%), and manual segmentation in 7 (21%) studies. Median automatic segmentation time was 10 min (IQR 33), semiautomatic 25 min (IQR 73). Manual segmentation time was reported in only one study, with the mean of 5-10 min. Thirty-two (97%) studies used screens to visualize the 3D segmentations outcomes and 1 (3%) study utilized a heads-up display (HUD). Integration with mixed reality was used in 4 studies (12%).

CONCLUSIONS

A golden standard for 3D visualization of bAVMs does not exist. This review describes a tendency over time to base segmentation on algorithms trained with machine learning. Unsupervised fuzzy-based algorithms thereby stand out as potential preferred strategy. Continued efforts will be necessary to improve algorithms, integrate complete hemodynamic assessment and find innovative tools for tridimensional visualization.

Highlights on Cerebral Arteriovenous Malformation Treatment Using Combined Embolization and Stereotactic Radiosurgery: Why Outcomes are Controversial?

Cerebral arteriovenous malformations (AVMs) are abnormal tangling between brain arteries and veins causing an arteriovenous shunt called nidus with an intervening network of vessels from the region of formation and spans through the brain. AVM effect is debilitating to the affected individual due to associated persistent intracerebral hemorrhage, resulting in significant occurrences of seizures and neurological damage. Recent innovative treatments involve a combination of embolization (Embo) procedures followed by stereotactic radiosurgery (SRS), designed to optimize less-invasive practice for the obliteration of the AVMs. Three groups of investigators reported different outcomes based on obliteration rates and adverse events, making the effectiveness of options for therapy, controversial. We have taken the case-oriented-approach to highlight on varying outcomes from various studies and provide insights as to why findings from different operation settings could be so conflicting. We chose 18 articles for systematic analysis based on initial electronic database selection of 40 key papers already identified for inclusion, followed by independent blinding assessment by two co-authors. Our evaluation was based first on our specific inclusion criteria, examining method quality, obliteration rates, serious adverse events (SAEs) and mortality rates. Second, we made a comparison between SRS or embo alone treatments versus combined embo/SRS procedures, relative to AVM sizes, following Spetzler-Martin (SM) method. Third, we considered publications which had concrete statistics with well-defined P-values and clarified outcomes for accurate evaluation. We found that patients with small to medium-sized AVM were susceptible to either embo alone or SRS alone treatments, yielding obliteration rates from 71%-100%. Except for one report, giant sizes AVMs were not amenable to these single treatments, subjecting patients to embo/SRS procedures, which yielded mixed results: One group reported 52%-65% obliteration rates, compared to 23%-28% embo alone treatment. A second group contradicted this apparent beneficial outcome, obtaining obliteration rates of 53% with combined treatment compared to 71% with SRS alone, four-year postoperative. A third group reported there was no difference between single and combined treatments and obtained complete obliteration of 70%-82%, ranging from three-five-years postoperative follow-up. In all the cases analyzed, obliteration rates improved with time. SAEs, such as persistent hemorrhage and permanent neurologic deficits (P-NDs), as well as mortality, were minimal during intraoperative and postoperative follow-ups. The problem of conflicting outcomes in combined treatments of AVM by EMBO/SRS exists. Previous investigators, however, have overlooked to address this issue satisfactorily. Our analysis found that the reported inconsistencies in AVM treatment outcomes are attributable to key factors making therapy unpredictable, which includes: the size of the AVM, nidus localization and accessibility of either Embo or radiation dose applied, certain Embo materials lowering obliteration rates by masking radioactive effect on the nidus during SRS and follow-up timing for obtaining obliteration rates determine the extent of obliteration. We have indicated critical factors which require consideration when planning strategies for treatment of AVM patients and have made suggestions of how to overcome such hurdles.

Clinico-radiological outcomes following gamma knife radiosurgery for pediatric arteriovenous malformations

PURPOSE

The aim of this study was to evaluate clinico-radiological outcomes following gamma knife radiosurgery (GKS) for pediatric arteriovenous malformations (AVMs).

METHODS

The present series included 39 children (3-17 years of age) who underwent GKS for cerebral AVMs between January 2002 and February 2008. Twenty-five patients presented with hemorrhages. The median AVM volume was 1.5 cm³, and the median marginal dose was 20 Gy. All patients continued to have follow-up for more than 24 months with serial magnetic resonance images (MRIs)/angiograms. Current school performance has been evaluated using a telephone survey answered by the patients' parents.

RESULTS

Follow-up angiograms, available in 34 patients, confirmed complete obliteration in 16 patients. Serial MRIs indicated obliteration of the nidus in one of five patients without angiography. The diffuse nidus structure and low marginal dose were significantly associated with incomplete obliteration. Twelve patients underwent a second GKS, and subsequent angiographies, available in six patients, demonstrated complete obliteration in two of them. Complications included new-onset seizures (n = 1), apraxia (n = 1), and temporal horn entrapment requiring a shunt operation (n = 1). School performance declined in 14 out of 32 patients. Both the AVM volume and modified AVM score were found to be reliable predictors of school performance.

CONCLUSIONS

The results of the present study substantiate the diffuse and other atypical features of pediatric AVMs as major determinants of treatment failure following GKS. Considering the apparent declination of school performance, future prospective studies would be required to investigate the possible late-effects of GKS on neuropsychological function.

A methodology for validating a 3D imaging modality for brain AVM delineation: application to 3DRA

A general methodology is described to validate a 3D imaging modality with respect to 2D digital subtracted angiography (DSA) for brain AVMs (BAVM) delineation. It relies on the assessment of the statistical compatibility of the radiosurgical target delineated in 3D with its delineations in 2D. This methodology is demonstrated through a preliminary evaluation of 3D rotational angiography (3DRA). Generally speaking, BAVM delineation cannot be performed on 3DRA alone. However, in our study, 3DRA showed similar performances to DSA for rather easy cases, and even better for three patients. Conversely, three problematic cases are identified and discussed.

Clinical significance of 3D reconstruction of arteriovenous malformation using digital subtraction angiography and its modification with CT information in stereotactic radiosurgery

PURPOSE

A three-dimensional (3D) reconstruction method of arteriovenous malformation (AVM) nidus from digital subtraction angiography (DSA) in combination with CT and/or MRI was developed, and its usefulness was evaluated in this study.

MATERIALS AND METHODS

The contour of the AVM nidus was delineated on two orthogonal projected DSA images. First, the volume and center of the AVM nidus were calculated in a classic DSA plan using three maximal lengths of the nidus in three perpendicular directions, assuming that the nidus had a prolate ellipsoid shape. Second, in the 3D-DSA plan, the contours of the AVM nidus on the two orthogonal projected DSA images were segmented to be compatible with the slice thickness of the CT image. Assuming that each segment of the nidus has an ellipsoid pillar shape, the volume and center of each segment were calculated. The volume and 3D shape of the nidus were calculated by 3D reconstruction in the 3D-DSA plan. Third, in the CT-DSA plan, the contour based on the segmented DSA was superimposed on the corresponding transaxial CT image slice by slice. The cylindrical shape of the nidus in the transaxial image was modified using the enhanced CT images in the CT-DSA plan. These three planning methods were compared using dose-volume statistics from real patients' data. Eighteen patients with intracranial AVMs in different brain locations who had been treated by radiosurgery were the subjects of this study. To examine the visibility (validity) of the nidus on the CT image, the "nidus" was delineated on an enhanced CT image without DSA superposition in the CT plan and compared with the CT-DSA plan.

RESULTS

The variance in the distance between coordinates determined by the CT plan and those determined by the classic DSA plan was significantly larger than the variance in the CT-DSA plan (p < 0.0001 for lateral, AP, and craniocaudal directions). The difference in the variance was not reduced by the addition of MRI (p < 0.0001 for each direction). The mean volume +/- SD of the nidus calculated was 5.9 +/- 8.0 cm(3) in the classic DSA plan, 4.0 +/- 5.6 cm(3) in the 3D-DSA plan, and 3.6 +/- 5.2 cm(3) in the CT-DSA plan. The 3D-DSA plan significantly reduced the mean nidus volume 31.8% +/- 12.7% from the classic DSA plan (p = 0.0054). The CT-DSA plan further significantly reduced the volume 9.8% +/- 8.8% from the 3D-DSA plan (p = 0.0021). The mean overlapping volume of the nidus between the CT plan and CT-DSA plan was 2.6 +/- 4.3 cm(3) (range 0.17-18.9), corresponding to 63.7% +/- 19.2% (range 11.4-85.3%) of the volume in the CT-DSA plan.

CONCLUSIONS

The superposition of the segmented DSA information on CT was shown to be an important tool to determine the precise shape of the nidus and is suggested to be useful to reduce partial occlusion of the AVM or radiation complications in radiosurgery.

Observer agreement in the angiographic assessment of arteriovenous malformations of the brain

PURPOSE

We aimed to determine intraobserver and interobserver agreement in the characterization of brain arteriovenous malformation (AVM) angioarchitecture on intra-arterial digital subtraction angiograms.

METHODS

Five experienced interventional neuroradiologists independently reviewed 40 anonymized angiograms obtained at the time of first-ever AVM diagnosis. The allocation of the films to observers was balanced for AVM size and complexity. Every observer was compared with himself and all the others by distributing the films in 2 batches 3 months apart. The observers used standard forms to collect both quantitative and categorized qualitative angiographic data. To measure agreement we used the kappa statistic (kappa) for nominal data, weighted kappa for ordinal and discrete interval data, and Bland & Altman analysis for continuous data.

RESULTS

Intraobserver agreement was generally moderate to substantial, with 95% confidence intervals ranging from fair to almost perfect. However, for every characteristic, interobserver agreement was less than intraobserver agreement. Interobserver agreement was generally slight to moderate, with 95% confidence intervals ranging from less than chance to almost perfect.

CONCLUSION

This study demonstrates the need for robust and generalizeable definitions of AVM angioarchitecture and methods of nidus size measurement-with proof of good intraobserver and interobserver agreement-for future efforts to understand the prognosis and best treatment of AVMs.

Compartments in arteriovenous malformation nidi demonstrated with rotational three-dimensional digital subtraction angiography by using selective microcatheterization. Report of three cases

Although in several histological studies of arteriovenous malformation (AVM) nidi the presence of compartments has been documented, no clinical study has been published. The present study was conducted to determine the presence of nidus compartments in clinical cases by using a new radiographic method. Two patients with unruptured and one with a ruptured AVM (all Spetzler-Martin Grade III) were studied. A microcatheter was superselectively introduced into each of two or three feeding arteries of the AVMs under three-dimensional (3D) angiographic guidance to obtain 3D images of the nidus by using rotational digital subtraction angiography. On 3D images the different feeding arteries were found to be independent from one another, which allowed the authors to confirm the presence of compartments. On the other hand, separate feeding arteries often had a common draining vessel. Compartments in AVM nidi were demonstrated by a novel rotational 3D angiographic procedure by using superselective microcatheterization, which should be useful for designing treatment strategies for large and complex AVMs.