Residual brain AVMs after surgical resection: A literature review of causes and treatment considerations

1. Introduction

Cerebral arteriovenous malformations (AVMs) are congenital lesions that form due to failure in the development of capillaries and veins during the third week of pregnancy.¹ AVMs do not usually cause symptoms until patients are in their second to fourth decade of life. The most common initial symptoms are bleeding and seizures.² There are various modalities to treat AVMs, including microscopic surgical removal, endovascular embolization, and stereotactic radiosurgery.^3,4 The most effective treatment option is still a matter of debate. However, complete surgical removal is the only method that can immediately eliminate the risk of bleeding, particularly in AVMs that presented with bleeding.

Current studies indicate that around 4% of patients may still have AVM remaining after the microscopic surgical removal.^5,6 Despite this, the risk of hemorrhage from remaining AVMs, whether ruptured or unruptured, appears to be comparable to or even lower than the yearly risk of bleeding from unruptured AVMs.⁵ The likelihood of having residual AVM and sustaining a permanent new neurologic deficit as a result of surgery depends on several factors, including the size of the AVM, whether or not it has deep venous drainage and its location in an eloquent area.^6,7

The prevailing recommendation within neurosurgeons is to perform a subsequent surgical resection in the case that remnant AVMs are discerned on postoperative angiography. Nevertheless, when the residual AVM is situated in an eloquent brain area, a conundrum arises - excising the remnant AVM may augment the risk of neurological deficits, whereas abstaining from intervention could precipitate hemorrhage. The management of residual AVMs involves multiple modalities, including microsurgical resection, endovascular embolization, or stereotactic radiotherapy has not been extensively explored in the literature and continues to be a matter of considerable debate. The primary aim of this study is to conduct a comprehensive literature review to identify and analyze factors influencing the occurrence of residual arteriovenous malformations (AVMs) following microsurgical resection. Moreover, the study endeavors to offer valuable insights into diverse management strategies aimed at addressing these residual AVMs.

2. Methodology

2.1. Literature search

A systematic search of relevant literature was conducted to identify studies on residual AVMs following microsurgical resection. The search included electronic databases, such as PubMed, Scopus, and Web of Science, and the search strategy utilized the following terms: (((Brain Arteriovenous Malformations [MeSH Terms]) OR (AVM[Title/Abstract])) AND ((Residual [Title/Abstract]) OR (remnant [Title/Abstract]) OR (recurrence [Title/Abstract])) AND (surgery [Title/Abstract])). The initial search was conducted on January 17,2023, and no restrictions were applied to the publication date to capture all relevant studies up to the time of the search.

2.2. Inclusion and exclusion criteria

Studies were included based on their relevance to the research objective. The inclusion criteria encompassed research articles, case studies, and reviews reporting on residual AVMs in patients who underwent microsurgical resection. Exclusion criteria involved studies with inadequate data, non-English publications, animal studies, conference abstracts, and articles without full-text availability.

2.3. Study selection process

Two independent reviewers screened the identified studies by titles and abstracts. The full texts of potentially relevant articles were then reviewed based on the inclusion and exclusion criteria. Any discrepancies or uncertainties during the selection process were resolved through discussion and consultation with a third reviewer.

2.4. Data extraction

Data extraction involved the systematic collection of relevant information from the included studies. The following data were extracted: demographics of patients (age, gender), clinical features (preoperative symptoms, presence of ruptured AVM), AVM characteristics (size, location, Spetzler Martin grade, feeding arteries, venous drainage, eloquent location), and management strategies (type of treatment, timing of diagnosis, outcomes after treatment).

2.5. Data analysis

Data were analyzed by SPSS version 25 using descriptive statistics, including means and standard deviations (SD) for continuous variables and frequencies and percentages for categorical variables.

3. Results

3.1. Demographics and clinical features of included cases

The study included 37 patients diagnosed with residual AVM, with a mean age of 24.7 years (SD ± 12.49) and a female prevalence was 56.5% (Supplementary File 1⁽8, 9, 10, 11, 12, 13^,7,14, 15, 16, 17, 18, 19⁾ and Table 1). The most common symptom reported prior to microsurgical resection of the AVM was altered consciousness (33.4%), followed by hemiplegia (20.0%), headache and vomiting (20.0%). Additionally, ruptured AVMs were the majority (78.6%) of cases included in this study.

Table 1.

Original AVM characters.

Number of Cases	37
Demographics
Age (mean ± ST)	24.7 ± 12.49
Gender (No.=23)	No. (%)
Female	13 (56.5%)
Male	10 (43.5%)
Presenting Symptoms of original AVM (No.=15)	No. (%)
Disturbed level of consciousness	5 (33.4%)
Headache and vomiting	3 (20.0%)
Body sided weakness	3 (20.0%)
Seizure	2 (13.3%)
Transient visual deficits	2 (13.3%)
Rupture status (hemorrhage) of original AVM (No.=14)	No. (%)
Yes	11 (78.6%)
No	3 (21.4%)
Location of original AVM (No.=28)	No. (%)
Left frontal lobe	5 (17.8%)
Left cerebellar hemisphere	4 (14.2%)
Right frontal lobe	4 (14.2%)
Right temporal lobe	3 (10.7%)
Left occipital lobe	2 (7.1%)
Left parietal lobe	2 (7.1%)
Right parietal lobe	2 (7.1%)
Right cerebellar hemisphere	2 (7.1%)
Right occipital lobe	1 (3.5%)
Left temporal lobe	1 (3.5%)
Right perisylvian	1 (3.5%)
Thalamus	1 (3.5%)
AVM Size (No.=17)	No. (%)
0–3 cm	7 (41.2%)
3.1–6 cm	7 (41.2%)
>6 cm	3 (17.6%)
Eloquent Location (Yes/No) (No.=19)	No. (%)
No	11 (57.8%)
Yes	8 (42.2%)
Venous Drainage (superficial or deep) (No.=19)	No. (%)
Deep	12 (63.1%)
Superficial	7 (36.9%)
Spetzler Martin Grade (No.=23)	No. (%)
Grade 1	4 (17.3%)
Grade 2	9 (39.2%)
Grade 3	7 (30.4%)
Grade 4	1 (4.3%)
Grade 5	2 (8.8%)
Feeding Artery (No.=20)	No. (%)
MCA	10 (50.0%)
ACA	5 (25.0%)
AchA	1 (5.0%)
PCA	3 (15.0%)
AICA & PICA	1 (5.0%)
SCA	1 (5.0%)
Perforaters: Thalamoperforating branch of the basilar artery & medial posterior choroidal artery & lenticulostriate branches	3 (15.0%)

Abbreviations: AVM; Arteriovenous malformation, MCA; Middle cerebral artery, ACA; Anterior cerebral artery, PCA; Posterior cerebral artery, AchA; Anterior choroidal artery, AICA; Anteroinferior cerebellar artery, PICA; Posteroinferior cerebellar artery, SCA; Superior cerebellar artery.

3.2. Original arteriovenous malformation features of included cases

The mean size of AVM was 3.96 cm (SD ± 1.55). Most AVMs were originally less than 6 cm (82.4%) in size. The most frequent location was the left frontal lobe of the cerebrum (17.8%) and left cerebellar hemisphere (14.2%). 42.2% of the residual AVMs were located in an eloquent area and (63.1%) had deep venous drainage. The most common Spetzler Martin Grade was found to be grade 2 (39.2%), followed by grade 3 (30.4%) and grade 1 (17.3%). The most common feeding artery for AVM cases was MCA (50.0%), followed by ACA (25.0%).

3.3. Characteristics and management of residual AVMs after microsurgical resection

Table 2 shows that regarding the size of the remnant, most of them were less than 1 cm (58.8%), while 41.2% of the remnants were equal to or more than 1 cm. Regarding the diagnosis of the residual AVMs, most of the cases were diagnosed postoperatively (59.5%). Regarding the treatment modality of the remnant, 67.6% of reported case were treated by surgery, 18.9% by embolization and 8.1% by spontaneous thrombosis and 5.4% by stereotactic radiosurgery ablation. The management of cases diagnosed postoperatively was by surgery in 45.5% of cases and embolization in 31.8% of patients. In cases diagnosed postoperatively 13.6% of the cases were treated conservatively and resolved by spontaneous thrombosis while 9.1% were treated by stereotactic radiosurgery. AVM remnant which was diagnosed intraoperatively during microsurgical resection were mainly managed surgically in the same setting (86.6%). Radiological outcome after treatment of the remnant shows complete obliteration in 83.8% of cases. The modified Rankin Scale after microsurgical resection was 2 in the majority of cases (46.1%). All cases with AVM remnant size (<1) were diagnosed postoperatively, and only 85% of cases with AVM remnant ≥1 cm were diagnosed postoperatively.

Table 2.

AVM remnant size, diagnosis, and modalities of treatment.

Diagnosis of remnant (No. = 37)	No. (%)
Intraoperative	15 (40.5%)
Postoperative	22 (59.5%)
Size of remnant (No.=19)	No. (%)
<1 cm	11 (58.8%)
≥1 cm	8 (41.2%)
Treatment modality of the remnant (No.=37)	No. (%)
Surgery	25 (67.6%)
At same sitting of original AVM surgery	13 (52.0%)
At second surgery sitting	12 (48.0%)
Embolization	7 (18.9%)
Spontaneous thrombosis	3 (8.1%)
Stereotactic radiosurgery ablation	2 (5.4%)
Radiological outcome after treatment of the remnant (No.=31)	No. (%)
Complete obliteration	26 (83.8%)
Incomplete obliteration	5 (16.2%)
Modified Rankin Scale (No.=13)	No. (%)
0	4 (30.7 %)
1	0 (0%)
2	6 (46.1%)
3	1 (7.6%)
4	0 (0%)
5	2 (15.3%)
6	0 (0%)

4. Discussion

While surgical treatment of AVMs has been a topic of much discussion over recent years, there have been limited studies describing features characterizing residual AVMs after surgical resection. Residual AVMs, also referred to as remnant AVMs, largely occur due to a lack of awareness of the full extent of the AVM or proximity to eloquent areas that are wished to be preserved. They are thought to arise from perinidal dilated capillary networks that connect to the nidus, feeding arteries and draining veins via arterioles and venules.²⁰ The incidence of residual AVMs following surgical resection is around 4%, and the risk of bleeding in remanent AVMs appears to be similar to or slightly lower than the annual risk of hemorrhage in unruptured AVMs.¹⁸ In our study, we aimed to provide an overview of the characteristics of post-surgical residual AVMs including patient demographics, location, size, Spetzler-Martin grade, presence of deep venous drainage and associated symptoms or complications, in addition to the commonly used management strategies.

The mean age of individuals with residual AVMs after surgical resection was found to be 25 years of age. This is younger than the mean age of 37 years reported in a study conducted by Englot et al involving 440 patients with supratentorial AVM, and 30.3 years reported in another study conducted by Chowdhury et al involving 60 patients with brain AVMs.^21,22 The percentage of females (56.5%) reported to have residual AVM was almost similar to that of males (43.5%), which is consistent with data from the study conducted by Englot et al (48.6% were females and 51.4% were males).²¹

The most common symptom mentioned before undergoing microsurgical resection of the AVM was altered consciousness (33.4%), followed by headache and vomiting (20.0%), limb weakness (20.0%), seizure (13.3%) and finally transient visual deficits (13.3%). The rates of symptoms reported in our study were overall less than those reported in other studies; Chowdhury et al reported that 50% experienced altered consciousness and 50% had headache, which were the two most common symptoms reported. This was followed by vomiting (40%), seizure (40%) and visual disturbance (10%).²² In the study by Englot et al, residual AVMs were not found to be associated with seizure recurrence.²¹ However, another study did find that remnant AVMs were related to higher rates of seizure recurrence.¹⁴ The risk of seizure is found to be increased in males, individuals with a history of AVM hemorrhage, frontotemporal location of the lesion as well as AVMs with deep artery perforators.²¹

We also found that more residual AVMs were associated with hemorrhage (78.6%), similar to findings from Chowdhury et al (70.0%) and Englot et al (54.8%).^21,22 Hemorrhage is a well-documented complication in cases of incomplete AVM obliteration. The risk is increased in patients with previous AVM hemorrhage, deep venous drainage, infratentorial or deep AVM location and presence of an aneurysm. Interestingly, size of AVM is not usually associated with risk of hemorrhage.²³ With a mortality rate of 10–15% and a morbidity rate of 30–50%, hemorrhage presents as a serious complication in patients with AVM.²⁴ Hence, to prevent this complication, it is recommended for residual AVMs to be treated definitively, possibly in the same surgical setting if possible.²⁰

The original lesion for the majority of the residual AVMs was found to be located in the left frontal lobe (17.8%), followed by the left cerebellar hemisphere (14.2%), the right frontal lobe (14.2%) and right temporal lobe (10.7%). This is consistent with other data showing that AVM is slightly more frequently found on the left when compared to the right cerebrum and more likely to be located supratentorially than infratentorially.^21,22 Supratentorially, our study found that the frontal lobe was the most common site for residual AVM (40.9% of supratentorial sites), followed by the temporal (18.2%), parietal (18.2%) and lastly occipital (13.6%) lobe. This is consistent with data showing the rate of AVM being highest in the frontal lobe (39.3%) followed by the temporal (27.4%), parietal (20.2%) and lastly occipital (13.1%) lobe.²¹

Taking into account the size, presence of deep venous drainage and eloquence of the area involved, the Spetzler-Martin grade is useful in assessing surgical risk. A higher grade is predictive of increased rate of morbidity post-surgery and greater prospective risk of hemorrhage.^25,26 Given the lower operative risk with grade 1 and 2 AVMs, surgical resection is the treatment of choice.²⁷ For grade 3 to 5 AVMs, multiple treatment options are available in addition to surgery, including embolization and stereotactic radiosurgery.^{28, 29, 30} In our study, most of the AVM remnants were originated from AVM with Spetzler-Martin grade 2 (39.2%), followed by grade 3 (30.4%), grade 1 (17.3%), grade 5 (8.8%) and lastly grade 4 (4.3%). Resection of AVMs of higher grade tend to be more complex in nature, which may explain why grade 3 AVMs were more common than grade 1 AVMs. However, variability in treatment and location may have affected the sub-categorical frequencies. Similarly, Englot et al also reported similar data showing that grade 2 AVM (37.6%) were most common, followed by grade 3 (32.8%) then grade 1 (18.3%). However, grade 5 AVMs (<0.1%) were found to be exceedingly rare and the least common.²¹

Presence of deep venous drainage is correlated with higher risk of hemorrhage and worse MRS score at discharge and follow up at ≥12 months.^15,21 Our study found that a higher proportion of residual AVMs were associated with deep venous drainage (63.1%) compared to superficial venous drainage (36.9%). This is at odds with data from Chowdhury et al and Englot et al showing that superficial venous drainage was more common than deep venous drainage.²¹

The most common feeding artery for AVM was MCA (50.0%), followed by ACA (25.0%), PCA (15.0%), perforators (thalamoperforating branch of the basilar artery, the medial posterior choroidal artery and lenticulostriate branches) (15.0%), AchA (5.0%), SCA (5.0%) and AICA and PICA (5.0%). Overall, remnants tended to occur in AVMs that were supplied by the anterior circulation more than the posterior circulation, while AVMs that were supplied by perforators tended to have very few residuals. Similarly, Chowdhury et al found that feeding vessels were most likely to come from the anterior circulation when compared to the posterior circulation, with main vessel involvement being more frequent than branch vessel involvement.¹¹ Furthermore, several studies have found that the presence of deep perforating arteries predicts worse outcomes after surgical resection.^31,32

AVMs in eloquent areas of the brain, such as the motor, language and visual areas as well as deep structures, present a challenge as they are associated with increased morbidity.²⁵ While our data showed that most residual AVMs occurred in a non-eloquent area of the brain (57.8%), data from Englot et al found that AVMs occurred more frequently in eloquent areas of the brain (60.3%) when compared to non-eloquent areas (39.7%).²¹ On the other hand, the percentage of AVMs that were positive versus negative for eloquence was found to be equal in the study by Chowdhury et al²² One would expect that location in an eloquent area of the brain would warrant higher level of caution during resection to preserve healthy tissue, and would hence lead to a higher potential of residual AVM from occurring. Our unexpected finding might perhaps be attributed to the tendency of AVMs in eloquent areas to be referred to tertiary centers that are more well-equipped with use of extensive or advanced intraoperative monitoring (e.g., indocyanine green video angiography), resulting in lower rates of residual AVMs occurring. Alternatively, the higher frequency of post-surgical residual AVMs in non-eloquent areas seen in our study might also be partly attributed to preference of treating residual AVMs in eloquent areas with stereotactic radiosurgery as opposed to surgical resection.³³ One thing to note about our study, as well as most other data determining eloquence in risk assessment for AVM, is that they do not differentiate the type of eloquence. The significance of this lies in the nuance that sensorimotor and language eloquence are associated with worse outcomes compared to visual eloquence following AVM resection.³⁴

Regarding the original AVM, the size was variable. The majority of the original AVMs fell between 0 and 3 cm in size (41.2%) or 3.1–6 cm (41.2%), followed by > 6 cm (17.6%). This is the opposite of the findings from Chowdhury et al, where most AVMs were >6 cm followed by those that were 3.1–6 cm and finally <6 cm²². Our unexpected finding that original AVMs that were larger in size tended to have less reported residuals could perhaps be due to the small sample size. The mean size was 3.96 cm, which is slightly higher than the mean size of 2.75 cm reported by Englot et al²¹ In a prospective study conducted by Stapf et al involving 240 patients undergoing surgical resection for AVM, no association was found between the size and location of the original AVM and the rate of residual AVM.³⁵ Of note, the role of surgical resection of large AVMs has been limited. Resection has traditionally been preferred for smaller AVMs, with some studies recommending a focus on resecting AVMs smaller than 3 cm as they are associated with better outcomes post-resection.²⁶

With regards to the size of the remnant after surgery, our study found that most remnants were less than 1 cm (58.8%) in size, with the rest (41.2%) being equal to or more than 1 cm. Increased difficulty in identifying smaller residual AVMs during surgery might potentially contribute to higher residual AVM occurrence, but we found that residual AVMs that were smaller (<1 cm) were more likely to be diagnosed intraoperatively than larger AVMs (>1 cm). Our small sample size might be an influencing factor contributing to these findings. In the current study, we found that, the original AVM of most of the smaller remnants were classified as grade 2, whereas the original AVMs of most of the larger remnants were grade 1.

The Modified Ranken Scale (MRS) score after microsurgical resection was 2 in the majority of cases (46.1%). This is higher than the most frequently reported score of 0 for compact AVM and 1 for diffuse AVM in a study conducted by Du et al involving 304 patients who underwent microsurgical resection. Either way, a score of 0–2 is considered to be a good outcome, with poor outcomes defined as a score greater than 2³².

Most of the cases were diagnosed postoperatively (59.5%). Table 3 shows that the management of postoperative cases was by second surgery in 45.5% of cases and embolization in 31.8% of cases. 15% of the cases diagnosed postoperatively were treated conservatively and resolved by spontaneous thrombosis while 9.1% of cases were treated by stereotactic radiosurgery. All the AVMs that were treated in the same setting were larger than 1 cm. Our study showed that for remnants of originally low-grade AVMs (Spetzler–Martin grades 1, 2 and some grade 3), surgery was the most common treatment modality, with the exception of remnants of grade 1 AVMs. This is generally consistent with guidelines suggesting that the initial recommended treatment for low grade residual AVMs is surgery. According to literature, stereotactic radiosurgery is an alternative for patients with surgically or endovascularly inaccessible lesions, or can be used as an adjunct with other treatment modalities. Embolization is usually used as part of a multimodal treatment strategy but can be curative in a minority of patients.²⁷ Commonly used embolic agents include N-butylcyanoacrylate (NBCA) and ethylene-vinyl alcohol copolymer (Onyx), which have demonstrated equivalent safety and efficacy. Many remnant AVMs following surgery are small with single feeding arteries and low Spetzler-Martin grade, making them suitable candidates for treatment with embolisation.¹⁸ Meanwhile, high-grade (Spetzler–Martin 4 and 5) lesions are best managed conservatively.²⁷

Table 3.

Different treatment modalities of AVM remnant are dependent on the time of the diagnosis.

Intraoperative diagnosis (No. = 15)		Postoperative diagnosis (No. = 22)
At same sitting of original AVM surgery	13 (86.6%)	At second surgery sitting	10 (45.5%)
At second surgery sitting	2 (13.4%)	Spontaneous thrombosis	3 (13.6%)
		Embolization	7 (31.8%)
		Stereotactic radiosurgery ablation	2 (9.1%)

On the other hand, our study showed that 40.5% of the remnant AVMs were diagnosed intraoperatively during microsurgical resection. Among those, most of them were managed surgically in the same setting (86.6%), with just 13.4% of cases managed by second surgery. The remaining cases were diagnosed postoperatively and of those, most were treated by second surgery (45.5%), followed by conservative with spontaneous thrombosis (13.6%), embolization (31.8%), and finally stereotactic radiosurgery (9.1%). Most of the treated AVM remnant cases show complete obliteration (83.8%) after being managed intraoperatively or postoperatively, consistent with the high cure rates reported in other studies; embolization was able to cure all residual AVMs in one retrospective study, while stereotactic radiosurgery was reported to have a cure rate of 85% for residual AVMs.^18,30 With regards to surgical treatment, data from a prospective study showed that 97% of patients with AVM improved or were unchanged from their pre-operative MRS scores. However, this study included data on primary AVMs and did not exclusively involve residual AVMs.³⁶ Incomplete obliteration, which includes cases of recurrent AVMs discovered during follow-up after a second microsurgical resection aimed at removing the remnant.¹⁶

A limitation of our study is our small sample size involving only 35 cases of residual AVMs post-resection, the data of which were mostly extracted from published case studies and small case series. This serves as a challenge to generalize our findings to a wider population. A literature review of this nature is limited by the availability of relevant published data; data on characteristics of residual AVMs is exceedingly scarce in literature. Much of the contextualization of our data involves comparison with data on primary AVM, since there is remarkably more data on characteristics of primary AVMs rather than residual AVMs. There was also missing data regarding the parameters of the residual AVMs such as the demographics, presenting symptoms, presence of hemorrhage, location and size of AVM, etc. Finally, there is lack of general consensus on the treatment strategies for residual AVMs after surgery, making it hard to draw definitive conclusions in this regard.

5. Conclusion

The occurrence of residual AVMs following microsurgical resection is multifactorial, influenced by variables such as age, gender, AVM size, location, Spetzler-Martin grade, and vascular anatomy. Addressing these remnants requires a tailored approach, considering factors like size, location, and the timing of diagnosis. Treatment options encompass a spectrum of interventions, including microsurgical resection, endovascular embolization, and stereotactic radiotherapy.

CRediT authorship contribution statement

Mostafa H. Algabri: Writing – original draft, Resources, Methodology, Data curation. Maliya Delawan: Writing – original draft, Methodology, Formal analysis, Data curation. Mayur Sharma: Writing – review & editing, Methodology. Mohammed S. Al-Hilfi: Writing – original draft, Resources, Methodology. Muntadher H. Almufadhal: Writing – original draft, Data curation. Noor M. Shaker: Writing – original draft, Visualization, Data curation. Zainab I. Abdualmurttafie: Resources, Methodology. Mustafa Ismail: Writing – review & editing, Resources, Methodology. Norberto Andaluz: Writing – review & editing, Supervision. Samer S. Hoz: Writing – review & editing, Writing – original draft, Resources, Methodology, Data curation, Conceptualization.

Declaration of Competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Abbreviations

ACA

Anterior cerebral artery

AchA

Anterior choroidal artery

AICA

Anterior inferior cerebellar artery

AVM

Arteriovenous malformation

MCA

Middle cerebral artery

MRS

Modified Rankin Scale

PCA

Posterior cerebral artery

PICA

Posterior inferior cerebellar artery

SCA

Superior cerebellar artery

Appendix A. Supplementary data

The following is the Supplementary data to this article: