Intraoperative detection of residual arteriovenous malformation using narrow-band imaging: illustrative case

Abstract

BACKGROUND

Complete obliteration of the nidus is the primary goal of microsurgical treatment for arteriovenous malformations (AVMs), yet small residual components may be difficult to detect under conventional white-light visualization.

OBSERVATIONS

A 46-year-old man presented with a ruptured Spetzler-Martin grade III AVM with a large temporoparietal intracerebral hemorrhage. After endovascular embolization, microsurgical resection was performed using a 4K 3D ORBEYE exoscope system. Although the nidus appeared completely removed under white light, narrow-band imaging (NBI) revealed a small vascular lesion within the resection cavity. Indocyanine green videoangiography (ICG-VA) confirmed residual blood flow, and the suspected remnant was additionally resected. Histopathological analysis demonstrated arterialized vessels with irregular internal elastic laminae consistent with residual AVM components. Postoperative angiography confirmed complete obliteration. The patient was transferred for rehabilitation on hospital day 70 with mild aphasia and sensory disturbance (modified Rankin Scale score 2).

LESSONS

NBI served as a useful adjunct for detecting small residual AVM components during microsurgical resection. Combined use with ICG-VA may improve intraoperative assessment of nidus removal.

https://thejns.org/doi/10.3171/CASE26265

Arteriovenous malformations (AVMs) are abnormal vascular lesions characterized by direct arteriovenous shunts without an intervening capillary bed. Microsurgical resection remains the definitive treatment for AVMs, particularly in patients presenting with hemorrhage. Complete obliteration of the nidus is essential because residual components may increase the risk of postoperative hemorrhage. However, identifying small residual nidus components within the resection cavity can be challenging under conventional white-light visualization.

Narrow-band imaging (NBI) is an optical imaging technique that enhances vascular contrast by using narrow-band wavelengths corresponding to hemoglobin absorption peaks. Although NBI has been widely used in endoscopic procedures, its application in cerebrovascular surgery has rarely been reported. The ORBEYE, an exoscope developed by Olympus Medical Science, utilizes NBI technology as an optional feature.

We report a case of ruptured AVM in which NBI suggested the presence of residual AVM components within the resection cavity and contributed to complete nidus removal.

Illustrative Case

A 46-year-old man presented with headache, vomiting, leftward conjugate eye deviation, and right-sided hemiparesis. His past medical history was unremarkable, and the patient was not taking any medications. The patient had a prestroke modified Rankin Scale (mRS) score of 0. On admission day, he developed a sudden headache followed by agitation. His level of consciousness fluctuated, with episodes of decreased consciousness accompanied by snoring-like breathing. Emergency medical services were activated, and right-sided hemiparesis was noted on their arrival.

The patient was 158 cm in height and weighed 60.2 kg. His level of consciousness ranged from Japan Coma Scale (JCS) score I-3 to II-10, corresponding to a Glasgow Coma Scale score of E3V3M5. Neurological examination revealed right-sided hemiparesis with a muscle strength of approximately 1–2/5 on manual muscle testing (MMT). The National Institutes of Health Stroke Scale score was 21. Vital signs were stable (blood pressure 147/104 mm Hg, heart rate 76 beats per minute, respiratory rate 17 breaths per minute, oxygen saturation 99% on room air, body temperature 35.4°C).

Head CT revealed a large left temporoparietal subcortical intracerebral hemorrhage measuring approximately 5 × 5 × 9 cm. A slight midline shift was observed (Fig. 1A). CT angiography demonstrated an AVM with a nidus measuring approximately 3.5 × 3.1 cm (Fig. 1B). The lesion was primarily supplied by a branch of the anterior parietal artery arising from the left middle cerebral artery. The nidus was located anterior to the hematoma and centered in the postcentral gyrus, suggesting that the AVM was the source of the hemorrhage. Two dilated cortical draining veins were identified, both draining into the junction of the superior sagittal sinus and the transverse sinus (Fig. 1C and D). No deep venous drainage was observed. Based on these findings, the patient was diagnosed with a ruptured Spetzler-Martin grade III AVM associated with subcortical intracerebral hemorrhage.

FIG. 1.

Preoperative imaging studies. A: Axial head CT image showing a subcortical intracerebral hemorrhage measuring approximately 5 cm in diameter in the left parietal and temporal lobes, accompanied by slight midline shift. B: CT angiogram demonstrating an AVM with a nidus measuring approximately 3.5 × 3.1 cm. The nidus was located anterior to the hematoma and centered in the postcentral gyrus. C and D:Three-dimensional CT angiograms in the anteroposterior (C) and left anterior oblique (D) views demonstrating an AVM nidus primarily supplied by the anterior parietal artery and drained by two cortical veins coursing toward the transverse sinus and the superior sagittal sinus.

Following admission, the patient was initially managed conservatively with antiepileptic medication and osmotic diuretics. His neurological symptoms gradually improved; motor strength recovered to 5/5 on MMT, and his level of consciousness improved to a JCS score of 3. However, aphasia and headache persisted. Decreased pain and temperature sensation in the right upper and lower extremities was also noted.

On hospital day 28, endovascular treatment was performed. Under general anesthesia, the right femoral artery was punctured, and a 6-Fr FUBUKI XF Neurovascular Long Sheath (Asahi Intecc) was guided into the left internal carotid artery. A Marathon microcatheter (Medtronic) was then navigated to the nidus. Angiography demonstrated that a branch of the anterior parietal artery served as the main feeder and functioned as a terminal feeder to the nidus. Another branch of the anterior parietal artery and the posterior parietal artery acted as passing feeders, supplying the nidus through side branches while continuing distally toward the cortical surface (Fig. 2A and B). Endovascular embolization was performed using ethylene-vinyl alcohol copolymer (Onyx 18, Medtronic) and n-butyl-2-cyanoacrylate (NBCA; Histoacryl, B. Braun). Embolization was performed sequentially through multiple feeders to achieve targeted devascularization. A total of 2.5 mL of Onyx was delivered across four injections with a cumulative injection time of 1631 seconds (approximately 27 minutes). NBCA was administered at concentrations of 25% (0.11 mL) and 20% (0.07 mL). Approximately 60% angiographic obliteration of the nidus was achieved (Fig. 2C–H).

FIG. 2.

Intraoperative fluoroscopic images obtained during endovascular embolization. A and B: Anteroposterior (A) and left anterior oblique (B) arterial phase views of left internal carotid angiograms obtained before endovascular embolization demonstrating the AVM nidus. C and D: Anteroposterior (C) and left anterior oblique (D) fluoroscopic views obtained during embolization demonstrating the Onyx cast within the nidus. E–H: Anteroposterior (E) and left anterior oblique (F) arterial phase images and corresponding anteroposterior (G) and left anterior oblique (H) venous phase images of left internal carotid artery obtained after Onyx embolization, demonstrating approximately 60% obliteration of the AVM nidus.

Two days after embolization, microsurgical resection of the AVM was performed. A curvilinear incision was made over the left parietal region followed by a 5 × 5–cm craniotomy. The ORBEYE exoscope system was introduced, and the dura was opened in a cruciate fashion. From the cortical surface, the Onyx cast and part of the embolized nidus were visible. The lesion was circumferentially dissected from the surrounding parenchyma after coagulation of the remaining arterial feeders. Finally, the draining vein was divided, and the nidus was removed en bloc. The posterior aspect of the nidus was contiguous with the hematoma cavity, consistent with hemorrhage secondary to AVM rupture. The hematoma was evacuated as much as possible.

Under standard white-light visualization, no obvious abnormal vessels appeared to remain in the resection cavity (Fig. 3A). However, NBI suggested the presence of residual abnormal vasculature (Fig. 3B), which was subsequently confirmed by intraoperative indocyanine green videoangiography (ICG-VA) (Fig. 3C). Additional tissue was therefore resected and submitted for histopathological examination (Fig. 3D) (Video 1).

FIG. 3.

Intraoperative exoscopic views. A:After AVM resection, the nidus appeared to be completely removed under conventional white-light visualization. B: Narrow-band image demonstrating a small vascular lesion with a characteristic dark green appearance suggestive of a residual nidus (arrowhead). C: ICG videoangiogram demonstrating residual blood flow within the lesion (arrow). D: The residual lesion was additionally resected and submitted for histopathological examination.

VIDEO 1. Clip showing residual AVM under white light, NBI, and ICG. Click here to view.

Histopathological examination demonstrated a vascular lesion composed of vessels of varying calibers. Artery-like vessels with muscular walls and relatively thin-walled vein-like vessels were intermixed (Fig. 4A and B). Elastica van Gieson (EVG) staining demonstrated internal elastic laminae in the artery-like vessels, with focal irregularities and disruption (Fig. 4C). No evidence of neoplastic proliferation or features characteristic of cavernous malformation were observed. In the clinical context of additional resection following AVM removal, these findings were considered consistent with residual AVM tissue.

FIG. 4.

Histopathological findings of the resected specimen. A and B: Low-power (A) and high-power (B) hematoxylin and eosin–stained sections demonstrating numerous vascular channels within the brain parenchyma. The vessels show arterial features with irregular architecture, consistent with components of an AVM. C: EVG staining demonstrating arterialized vessels with irregular internal elastic laminae, consistent with residual components of an AVM nidus.

The postoperative course was uneventful. Postoperative cerebral angiography demonstrated no residual vascular malformation (Fig. 5). No new neurological deficits were observed. Although sensory aphasia and decreased superficial sensation persisted, the patient achieved an mRS score of 2. On hospital day 70, he was transferred to a rehabilitation hospital for further recovery.

FIG. 5.

One-month postoperative cerebral angiograms demonstrating no residual AVM. A and B: Anteroposterior (A) and lateral (B) arterial phase images. C and D: Corresponding anteroposterior (C) and lateral (D) venous phase images of left internal carotid artery.

Informed Consent

The necessary informed consent was obtained in this study.

Discussion

Observations

In the case of hemorrhagic AVMs, if the waiting approach is feasible, it is preferable to defer treatment until the cerebral edema and the reduced visibility of the lesion caused by vascular compression have subsided. A variable rest period of 1–6 weeks between hemorrhage and definitive treatment is generally advocated in the literature.¹ The primary objective of microsurgical treatment for AVMs is complete obliteration of the nidus. Residual nidus components, even if small, may increase the risk of postoperative hemorrhage due to altered hemodynamics.^2–5 However, identifying small remnants within the resection cavity can be challenging under conventional white-light visualization.

NBI is commonly used in gastroenterology and otolaryngology to enhance the visualization of mucosal and vascular patterns.^6,7 NBI has also been used in neurosurgery for neuroendoscopy and exoscopy.^8,9 NBI can enhance the contrast between blood vessels and neural tissues and help to depict minuscule vascular structures. Because the condition of the lesion could be monitored in real time during surgery, repeatedly checking the lesion site was very useful.⁹ To our knowledge, the use of NBI for intraoperative identification of residual intracranial AVM components has rarely been reported. Re-verification of the NBI-positive area with ICG-VA allowed confirmation of residual blood flow, directly leading to an improved precision of the total nidus resection.

NBI utilizes filtered wavelengths—specifically 415 nm (blue) and 540 nm (green)—which coincide with the peak light absorption of hemoglobin.⁷ AVM vessels are densely aggregated and possess relatively thick muscular walls, which may increase hemoglobin absorption and reduce light reflection under NBI, thereby enhancing the contrast between the nidus and surrounding normal vessels.^6,7 Consequently, AVM vessels appear as a deeper and darker green color than normal vessels on NBI. In the present case, normal cortical vessels appeared as bright cyan-green linear structures, whereas the suspected residual nidus appeared as a relatively darker green irregular vascular structure. This difference in color and vascular pattern raised suspicion of residual abnormal vasculature and prompted further evaluation. Furthermore, while areas of pooled blood typically appear as dark or blackened regions on NBI, in this case the lesion appeared as a darker green region with partially preserved vascular architecture, which differed from simple blood accumulation.

Histopathological examination in this case demonstrated a cluster of vascular structures composed of arteries with muscular walls of varying calibers intermixed with relatively thin-walled vein-like vessels. EVG staining revealed internal elastic laminae within the artery-type vessels, although focal irregularity and fragmentation were observed. These findings were consistent with a residual nidus of an AVM.^10–12 Based on these pathological characteristics, it was considered that excessive hemoglobin absorption occurred compared with the surrounding normal brain tissue, thereby enhancing the contrast of the lesion.

ICG-VA remains the gold standard for intraoperative hemodynamics assessment; yet it possesses inherent practical limitations. ICG-VA provides only a snapshot and requires repeated intravenous injections with mandatory washout periods.¹³ In contrast, NBI is noninvasive, contrast free, and integrated into the digital workflow of the ORBEYE exoscope system, allowing for instantaneous and repeated switching between modes.^8,9 Our experience indicates that NBI and ICG-VA are synergistic: NBI can provide the initial morphological clue for residual pathology, which is then hemodynamically validated by ICG-VA.

Technical proficiency is essential to maximize the efficacy of NBI. Because visualization is easily obscured by surface blood or clots, maintaining a clean surgical field through meticulous hemostasis and active suction is paramount. Furthermore, the integration of NBI with the ORBEYE exoscope allows the entire surgical team to share high-definition views on a large monitor. This collective situational awareness acts as a safety net, facilitating a more reliable and radical resection of complex AVMs.

Lessons

In the context of AVM surgery, complete obliteration of the nidus is essential. However, the presence of small residual components may not be discernible through white-light visualization alone. Intraoperative NBI visually suggested the presence of residual abnormal vasculature, a finding that subsequently was confirmed by ICG-VA. The evidence is based on a single case; larger prospective data sets are required to define the routine utility of NBI.

Disclosures

The authors report no conflict of interest concerning the materials or methods used in this study or the findings specified in this paper.

Author Contributions

Conception and design: Hayashi, Goda. Acquisition of data: Hayashi, Goda, Imanishi, Murai. Analysis and interpretation of data: Hayashi, Goda. Drafting the article: Hayashi, Goda. Critically revising the article: Hayashi, Goda. Reviewed submitted version of manuscript: Hayashi, Goda, Horikawa, Murai. Approved the final version of the manuscript on behalf of all authors: Hayashi. Statistical analysis: Hayashi, Goda. Administrative/technical/material support: Hayashi, Goda, Maeda. Study supervision: Hayashi, Goda, Yoshioka, Kobayashi, Murai.

Supplemental Information

Videos

  Video 1. https://vimeo.com/1187015871.

Correspondence

Hideki Hayashi: Uji Tokushukai Medical Center, Kyoto, Japan. nrsrgnhd@gmail.com.