Uses and limitations of indocyanine green videoangiography for flow analysis in arteriovenous malformation surgery

Uses and limitations of indocyanine green videoangiography for flow analysis in arteriovenous malformation surgery

Journal of Clinical Neuroscience 20 (2013) 224–232 Contents lists available at SciVerse ScienceDirect Journal of Clinical Neuroscience journal homep...

560KB Sizes 1 Downloads 40 Views

Journal of Clinical Neuroscience 20 (2013) 224–232

Contents lists available at SciVerse ScienceDirect

Journal of Clinical Neuroscience journal homepage: www.elsevier.com/locate/jocn

Clinical Study

Uses and limitations of indocyanine green videoangiography for flow analysis in arteriovenous malformation surgery Yew Poh Ng ⇑, Nicolas KK King, Kai Rui Wan, Ernest Wang, Ivan Ng Department of Neurosurgery, National Neuroscience Institute, 11 Jalan Tan Tock Seng, Singapore 308433, Singapore

a r t i c l e

i n f o

Article history: Received 21 September 2011 Accepted 31 December 2011

Keywords: Arteriovenous malformation Fluorescence intensity analysis Indocyanine green

a b s t r a c t Intra-operative indocyanine green (ICG) videoangiography is a useful addition to cerebrovascular neurosurgery. ICG videoangiography is useful in different phases of arteriovenous malformation (AVM) surgery. Additionally, it can be used to perform semi-quantitative flow analysis. We reviewed our initial assessment of 24 patients who underwent ICG videoangiography during AVM surgery to assess the utility and limitations of the technique as well as to demonstrate semi-quantitative flow analysis, a new capability of ICG videoangiography. Over the course of 3 years, we performed 49 ICG videoangiographies in 24 patients with AVM. In 85% of the pre-resection videos, ICG was useful in localising the arterial feeders, the draining veins and the nidus. Intra-resection ICG videos were recorded for eight of the 23 patients (the ICG from one patient was missing). Post-resection ICG videos were recorded for 14 out of the 23 patients, which were useful in confirming no evidence of nidus in the exposed resection cavity and an absence of flow in the main draining vein. Semi-quantitative flow analysis was performed in eight patients with superficial AVM. The average T½ peak intensities (time to 50% of peak intensity) were 32 s, 33.5 s, and 35.6 s for the arterial feeder, the draining vein and normal cortex, respectively. The arteriovenous T½ peak time was 1.5 s, and the arteriocortex T½ peak time was 3.6 s. The T½ peak fluorescence rates were 84 average intensity of fluorescence (AI)/s, 62.9 AI/s and 28.7 AI/s, for the arterial feeder, the draining vein and normal cortex, respectively. Only one patient of 23 (4.3%) showed residual AVM on post-operative digital subtraction angiography or CT angiography despite negative intra-operative ICG. ICG videoangiography is a useful addition to AVM surgery, but it has some limitations. Flow analysis is a new capability that allows for semi-quantitative AVM perfusion analysis. Ó 2012 Elsevier Ltd. All rights reserved.

1. Introduction Intra-operative indocyanine green (ICG) videoangiography has been a useful addition to cerebrovascular neurosurgery.1 During aneurysm surgery, it can show obliteration of the aneurysm after clipping, and patency of perforating arteries.2,3 In arteriovenous malformation (AVM) surgery, ICG videoangiography qualitatively differentiates the arterial feeders, the nidus and the draining vein, if the AVM nidus is superficially located.4 During the resection, serial intra-operative ICG may be helpful in localising feeders and assessing nidus perfusion.5 At the end of the resection, ICG may show a residual nidus or remnant arteriovenous-shunting.6 As a safe, convenient and easily reproducible technique,7 the uses for ICG videoangiography have expanded rapidly into the realms of tumour surgery,8 spinal arteriovenous fistula surgery,9 bypass surgery10 and carotid endarterectomy.11 Recently, qualitative ICG

⇑ Corresponding author. Tel.: +65 98174772; fax: +65 63577137. E-mail address: [email protected] (Y.P. Ng). 0967-5868/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.jocn.2011.12.038

videoangiography has been further enhanced by the new capability of flow analysis, based on fluorescence intensity measurements.12 Our unit began using ICG videoangiography in 2008. We performed ICG videoangiography on 61 patients, for various reasons, from May 2008 to June 2011. We aim to critically review our initial experience in 24 patients who had ICG videoangiography during AVM surgery. We assess the utility of, and identify the limitations of, ICG videoangiography. In addition, we aim to demonstrate a new capability of ICG videoangiography, semi-quantitative fluorescence intensity analysis, in a specific group of patients with superficial AVM.

2. Patients and methods Our study was performed at two campuses of the Department of Neurosurgery, National Neuroscience Institute, Singapore. Informed consent was obtained from all patients prior to surgery for AVM resection with ICG videoangiography. We conducted a retrospective, descriptive analysis of ICG videoangiography in patients with AVM. Demographic information, case

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

records, radiological images and operative outcome were traced and reviewed. Over the course of 3 years, we operated on 24 consecutive patients with AVM using intra-operative ICG videoangiography.

2.1. Indocyanine green The physical and physiological properties of ICG were first described by Fox and Wood in 1960.13 ICG is a non-toxic tricarbocyanine dye (molecular weight 775 Da) with peak spectral absorption at 805 nm and peak emission at 835 nm. It is used for determining cardiac output, hepatic function, and liver blood flow as well as for ophthalmic angiography; these uses were approved by the United States Food and Drug Administration in 1956 and 1975.14 Upon intravenous (IV) injection, ICG binds rapidly to plasma globulin and is thereby confined to the intravascular compartment. Due to a significant first pass effect in the liver, ICG has a short half-life of 150 s to 180 s. It is not metabolised and is excreted by the liver, limiting any recirculation phenomenon. Its first reported use in neurosurgery occurred in 2001.15 The safety profile of ICG is good. In a large study of more than 240,000 IV injections of ICG, only four patients had adverse reactions. One had an urticarial reaction and three had anaphylactic reactions.16 The methods of ICG videoangiography have been well described by Raabe et al.2 Briefly, ICG is administered IV by the anaesthetist, upon request of the neurosurgeon, at a single bolus of 25 mg in 5 mL of normal saline for adult patients. For paediatric patients, the dosage is 0.2–0.5 mg/kg.

2.2. ICG-integrated microscope and flow analysis All operations were performed using three microscopes available in our department (two Carl Zeiss OPMI PENTERO with INFRARED 800 microscopes (Carl Zeiss, Oberkochen, Germany) and one LEICA M720 OH5 (Leica Microsystems, Singapore) with FL800 unit microscope). All three microscopes had near-infrared video integration and were able to perform ICG videoangiography. After illumination with a near-infrared light source (excitation range of 700–780 nm), real-time images were recorded from the microscope. An optical filter that only allowed fluorescence in the ICG emission range (820– 900 nm) was used. To obtain optimum illumination, the microscope automatically set the diaphragm widely, and the illuminated field diameter was set in the middle position. The illumination intensity was set to 50%. The recommended working distance of the scope was set at 300 mm, and the zoom setting was set at five times for the PENTERO microscope.

2.3. Qualitative ICG video analysis The ICG videos from the 24 patients were retrospectively reviewed by an independent third-party neurosurgeon and were compared to the concurrent, normal, white-light video recording. The ICG video of one patient was missing. The indication for each ICG video during the AVM resection was recorded. As recommended by Hanggi et al.,7 the ICG videos were classified into three categories in relation to the phase of AVM surgery: (i) primary resection (pre-resection), (ii) intra-resection and (iii) post resection. Positive and negative attributes of each ICG video recording were reported. In addition, the suitability of the ICG video for flow analysis was noted. If the video was suboptimal for fluorescent intensity flow analysis, the reason was clearly stated.

225

2.4. Quantitative ICG video fluorescence flow analysis 2.4.1. Delay map and maximum intensity map In 2010, an additional module was installed in one of the PENTERO microscopes with FLOW 800 function. This additional module allowed for a fluorescence intensity analysis of the ICG video. Based on a proprietary algorithm, the microscope was able to compile all the information from the ICG video into one image. Two visual maps could be displayed intra-operatively. The maximum fluorescence attained at each region of interest (ROI) was displayed as different shades of grey. Another map, with different colours, demonstrated the temporal sequence of fluorescence intensity at each image point to determine the direction and sequence of blood flow in the AVM. The intuitive colour scale marked the early start of fluorescence in red and the late start in blue.

2.4.2. Fluorescence intensity graph versus time (flow analysis) In instances where the ICG videos showed superficial AVM with clear demarcation of the arterial feeders, the early draining veins and cortical perfusion (normal capillary phase), the neurosurgeon was able to choose up to eight ROI, based on the maximum intensity map, to calculate up to eight average fluorescence intensity curves against time of recording. This was followed by an automatic calculation of the time to half-peak intensity (T½peak) and the point gradient of fluorescence intensity over time (fluorescence intensity rate) at T½ peak. By placing the ROI at the arterial feeder, the draining vein and the normal cortex, the diagrams and subsequent calculations gave a semi-quantitative transit time of the AVM flow based on the difference in T½ peak time between the arterial feeder and the draining vein. The gradient of the fluorescence intensity rate was likely to be associated with the flow rate of the respective vessel.

2.4.3. Post-resection angiogram All patients who underwent AVM resection in this study underwent post-operative biplanar digital subtraction (DSA) or CT angiography (CTA) to document the completeness of resection. An independent neuroradiologist who was blind to use of ICG videoangiography assessed the completeness of resection. The correlation of post-resection ICG videoangiography with DSA or CTA was evaluated.

3. Results 3.1. Demographics Over a 3-year period, we performed 49 ICG videoangiographies in 24 patients with AVM. The ICG video from one patient was not located. The demographics of our patient population are shown in Table 1. There was an equal distribution of male and female patients with a mean age of 34.5 years (range: 6–56 years). There were 10 Grade 3 Spetzler–Martin AVM, 12 Grade 2 AVM and two Grade 1 AVM. The clinical presentation was haemorrhage in 71% of patients and seizures in 29% of patients. There were no side effects or adverse reactions noted after ICG administration in any of our patients. There was a consistent but transient effect on the patient’s pulse oximetry after ICG administration.

3.2. Qualitative analysis Results of the ICG qualitative analysis are shown in Table 2.

226

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

Table 1 Characteristics of patients who underwent indocyanine green videoangiography for flow analysis in arteriovenous malformation surgery Patient no.

Age

Sex

Presentation

Spetzler–Martin grade

Location

Post-operative DSA/CTA

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

13 44 38 22 11 8 34 47 56 32 50 27 61 46 38 6 35 37 40 31 16 30 44 63

F M M F F M F F M M F F M F M M F F M M M M F F

Hemorrhage Seizure Seizure Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage Seizure Hemorrhage Seizure Seizure Hemorrhage Seizure Hemorrhage Hemorrhage Hemorrhage Seizure Hemorrhage Hemorrhage Hemorrhage Hemorrhage Hemorrhage

1 3 2 2 3 3 3 3 3 2 2 3 3 2 2 2 3 1 2 2 2 2 2 3

Right frontal Left parietal Right occipital Right parietal-occipital Posterior fossa Right parietal-occipital Left occipital Right basal ganglia Right basal ganglia Right temporo-parietal Left cerebellar Left temporal Right parieto-occipital Left occipital Left parietal Left parieto-occipital Left parietal Right frontal Left parieto-occipital Left cerebellopontine angle Right frontal Right frontal Right inferior posterior temporal C1 cervical spine

Complete resection Complete resection Complete resection Complete resection Complete resection Residual nidus Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection Complete resection

CTA = CT angiography, DSA = digital subtraction angiography.

3.2.1. Pre-resection ICG We recorded 20 pre-resection (primary resection) videos from 23 patients. Most (85%) of the pre-resection videos were useful (17 of 20) in localisation of the arterial feeders, the draining vein and the nidus. Three pre-resection ICG videos were not useful. The main reason for failure was secondary to failure of fluorescence of the AVM nidus. This was usually because the AVM nidus was covered with a thin layer of adherent blood in the depth of the clot cavity in those patients who presented with hemorrhage, and there was poor visualisation of the arterial feeders and the draining vein, which limited the exposure. 3.2.2. Intra-resection ICG Intra-resection ICG videos were recorded less frequently. Only eight of the 23 patients had intra-resection ICG videos. However, intra-resection ICG videos, when indicated or requested by the neurosurgeon, were useful. The ICG videos of five patients identified remnants of deep arterial feeders requiring further dissection. The ICG videos of two patients revealed persistent flow in the main draining vein. One ICG video, in a patient with craniocervical AVM, showed no flow in the AVM after temporary clipping of the arterial feeder, with no swelling of the spinal cord. 3.2.3. Post-resection ICG Fourteen out of 23 patients had post-resection ICG videoangiography. The videos were considered useful if there was no evidence of remnant nidus in the resection cavity or if there was an absence of flow in the main draining vein. 3.3. Suitability of ICG video for flow analysis ICG with flow analysis was performed on 12 of 23 patients (52%) under the PENTERO microscope with flow analysis capability. Only 75% of the 12 patients had pre-resection ICG video that was suitable for flow analysis (that is, showed clear demarcation of the arterial feeder, the draining vein and the nidus). The main difficulties in semi-quantitative fluorescent intensity flow analysis on the ICG video were due to: (i) continuous dissection with instrument manipulation and shifting of the focus, zoom or operative

field to achieve greater exposure during video recording, (ii) poor illumination of AVM nidus in the depth of the clot cavity, (iii) cottonoid string and marker fluorescence affecting automatic fluorescence detection, (iv) cerebrospinal fluid (CSF) re-accumulation necessitating suction, (v) blood accumulation with leakage of ICG contrast; and (vi) close interval between ICG recordings resulting in background fluorescence. 3.4. Quantitative ICG video fluorescence intensity flow analysis Suitable pre-resection ICG videos for flow analysis were selected from eight patients with superficial cortical AVM. Overall, three ROI were selected based on the maximum intensity map that demarcated the arterial feeder, the draining vein and the normal cortex. Table 3 shows the time to half-peak intensity (T½ peak) and the point gradient (rate) of average intensity (AI) of fluorescence over time at T½ peak for each of the three ROI. In the eight selected patients, the flow analysis was precise and semi-quantitative. The mean values are shown in Table 4. The average T½ peak intensity was 32 s for the arterial feeder, 33.5 s for the draining vein and 35.6 s for normal cortex. The mean difference in T½ peak intensity time between the arterial feeder and the draining vein was 1.5 s. The mean difference between the arterial feeder of the AVM and cortical perfusion was 3.6 s. The AI of fluorescence over time at T½ peak (AI of fluorescence rate) differ: the arterial feeder has the highest fluorescence intensity rate at 84.1 AI/s, followed by the draining vein at 62.9 AI/s and normal cortex at 28.7 AI/s. The graph profiles of AI of fluorescence over time for the arterial feeder, the draining vein and the normal cortex were different. The arterial feeder tended to have a taller peak with a steeper gradient compared to the draining vein and normal cortex. 3.5. Post-resection angiogram As shown in Table 1, only one of the 23 patients (4.3%) had a small, residual AVM that was revealed by post-operative DSA despite normal post-resection ICG. This paediatric patient underwent stereotactic radiosurgery for treatment of the remnant nidus.

Table 2 Qualitative results of indocyanine green videoangiography for flow analysis in patients who underwent surgery for arteriovenous malformation Patient no.

No. of ICG videos

Preresection

Comments

Intraresection

Comments

Post resection

Comments

Flow analysis

Comments

Utility score

1

1

Useful

Not done



Not done



1

3

Useful

Not done



Useful

Cottonoid string fluorescence

2

3

2

Useful

Not done



Useful

Not available Not available Not available

Instrument manipulation

2

Localisation of AVM nidus and draining vein Localisation of arterial feeder and draining vein Localisation of arterial feeder, AVM nidus and draining vein

2

4

1

Not useful

Dissection during video recording; Cottonoid string fluorescence; Shifting of focus Cottonoid string fluorescence; dissection during recording; blood in cavity

5

3

Useful

Instrument manipulation; shifting of focus

3

6

1

Useful

1

7

1

Instrument manipulation; shifting of focus, blood extravasation staining tissue; AVM nidus hidden by brain parenchyma Instrument manipulation

8 9

AVM nidus in the depth of clot cavity; no fluorescence; blood in cavity Localisation of AVM nidus (delayed fluorescence) Localisation of AVM nidus (required dissection)

Not done



Not done

Absence of nidus and flow in draining vein Presence of nidus, further dissection required –

Useful

No flow in draining vein

Useful

Absence of nidus

Not done



Not done



Not done



Useful

Clear demonstration of deep feeders after dissection

Not done



Not available

Missing 1

Useful

Not done



Not done



1

1

Useful

Not done



Not done



Not available Available

Instrument manipulation; blood in field

10

Shifting of focus

2

11

1

Not done

Localisation of feeder and nidus in depth Localisation of arterial feeders, AVM nidus and draining vein –

Not done



Useful

No flow in main draining vein

Cottonoid string fluorescence; instrument manipulation

1

12

3

Useful

Localisation of arterial feeders, AVM nidus and draining vein

Useful

Useful

No flow in main draining vein

Shifting of focus

4

13

2

Useful

Localisation of arterial feeders, AVM nidus and draining vein

Useful

Shows remnant feeder and draining vein leading to further dissection Shows remnant feeder and draining vein leading to further dissection

Available (poor image) Available

Not done



Available

3

14

1

Not done



Useful

Shows remnant feeder

Not done



15

3

Useful

Useful

Absence of nidus

Not useful

Not done

Useful

Absence of nidus with no flow in vein

Instrument manipulation; irrigation disturbance; remnant contrast seen Cottonoid string fluorescence; shifting of focus

4

2

Shows remnant feeder on further dissection –

Useful

16

Localisation of arterial feeders, AVM nidus and draining vein AVM nidus did not fluoresce

Available (poor image) Available

Instrument manipulation to clear blood clot; remnant contrast in the blood clot; cottonoid string fluorescence Instrument manipulation; shifting of focus

17

2

Not useful

AVM nidus did not fluoresce

Not done



Useful

Absence of nidus

Cottonoid string fluorescence; blood with contrast

1

18

2

Useful

Not done



Useful

Absence of nidus

Cottonoid string fluorescence

3

19

3

Useful

Not done



Useful

Nil

3

3

Useful

Not done



Useful

No flow in draining vein No flow in nidus

Available

20

Localisation of arterial feeders, AVM nidus and draining vein Localisation of arterial feeders, AVM nidus and draining vein Localisation of arterial feeders, and draining vein

Available

Nil

3

Not available

Available (poor image) Available (poor image) Available

1

1

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

Not available Not available

0

1

(continued on next page) 227

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

4 Useful Temporary clipping showed no flow in draining vein Useful

Location of arterial feeder and draining vein

Useful

Useful Not done

Localisation of arterial feeders, AVM nidus and draining vein – Useful

Useful

Shows draining vein, no nidus on 1st video

No flow in draining vein and no remnant nidus Absence of flow in AVM

Available (poor image) Available

Instrument manipulation to clear blood clot; cottonoid string fluorescence

2

3

Cottonoid string fluorescence; instrument manipulation Instrument manipulation; remnant blood Absence of nidus Useful –

Pseudoaneurysm did not fluorescnce due to overlying clot Useful

Not done

Comments

Not done



Not done



Available (poor image) Available

Instrument manipulation

1

Table 3 Quantitative indocyanine green fluorescence intensity analysis in patients who underwent arteriovenous malformation surgery

Preresection

Intraresection

Comments

Post resection

Comments

Flow analysis

Comments

Utility score

228

Patient no.

Arterial feeder

Draining vein

Cortex capillary

T½ peak (s)

Slope (AI/s)

T½ peak (s)

Slope (AI/s)

T½ peak (s)

Slope (AI/s)

10 12 13 15 18 19 22 20

14.69 23.19 24.69 46.2 52.89 36.6 19.73 38.21

26.54 46.9 59.34 35.2 130.13 33.01 135.97 205.33

16.58 23.99 24.31 47.36 57.52 37.88 21.34 39.22

29.76 41.42 51.21 37.74 60.92 26.61 78.51 175.38

16.34 24.98 31.3 51.99 54.42 41.49 21.36 43.38

16.54 19.2 38.3 16.21 48.56 15.57 39.91 35.25

AI = average intensity of fluorescence.

Table 4 Mean T1/2 peak and fluorescent intensity rate for patients who underwent quantitative indocyanine green fluorescence intensity analysis in arteriovenous malformation surgery

Arterial feeder Draining vein Cortex capillary

Mean T1/2 peak(s) (n = 8)

Mean fluorescent intensity rate (AI/s) at T1/2 (n = 8)

Mean

Std error

95% CI

Mean

Std error

95% CI

32.03

4.78

84.05

23.07

33.53

5.05

62.69

17.2

35.66

5.04

20.7– 43.32 21.58– 45.47 23.74– 47.58

36.32– 131.79 27.14– 98.25 19.03– 38.35

28.69

4.67

AI = average intensity of fluorescence, CI = confidence interval, Std = standard.

3.6. Illustrative patient with a left temporal AVM We present a 27-year-old woman with her first onset of generalised tonic–clonic seizure. A pre-operative DSA revealed a left temporal AVM (Spetzler–Martin Grade 3) measuring 2.1  2.5  4.5 cm with feeders from the left anterior temporal middle cerebral artery (MCA) and the posterior temporal posterior cerebral artery (PCA) (Fig. 1). The draining veins were via the sphenoparietal sinus, the superficial middle cerebral vein (Sylvian) into the vein of Labbe to the sigmoid sinus and the vein of Trolard to the superior sagittal sinus. Functional MRI (word task generation) revealed that the AVM was anterior and inferior to both Broca’s and Wernicke’s areas. The patient underwent elective craniotomy and resection of the AVM. Her first ICG video showed clear demarcation of the feeding artery, the main draining vein and the normal cortex perfusion. Delay and maximum intensity maps were generated after flow analysis (Fig. 2a, b). Overall, three ROI were chosen, and the flow fluorescence intensity curves were generated (Figs. 2c and 3). An intra-resection ICG showed persistence of flow in the main draining vein that necessitated further dissection. In the postresection ICG, normal cortical vessels were seen with ICG fluorescence, but the surgical cavity revealed no remnant nidus (Fig. 4). Post-surgery CT angiography revealed no residual AVM (Fig. 5). Post-surgery, her recovery was uneventful, and she developed no neurological deficits.

No. of ICG videos

1

2

5

4

Patient no.

21

22

23

24

Table 2 (continued)

4. Discussion We performed a literature review from 2007-2011 of all the existing publications associated with ICG videoangiography and AVM surgery listed on PUBMED and summarised the results in Supplementary Table 1. Multiple publications have demonstrated the usefulness of ICG videoangiography in AVM surgery.4–7,17–19

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

229

Fig. 1. Digital subtraction angiogram (a) anterior and (b) lateral projection of ICA feeders of the illustrative patient with a left temporal arteriovenous malformation.

Fig. 2. Flow intensity analysis using indocyanine green fluorescent video angiography in the illustrative patient with a left temporal arteriovenous malformation.(a) Maximum intensity map; (b) delay map; and (c) picture showing the three regions of interest for flow analysis (red [1] = arterial feeder; green [2] = draining vein; blue [3] = cortex).

Chen et al.12 in 2011 were the first to report the use of ICG fluorescence intensity in patients with AVM. They concluded that ICG provided high-resolution images that could allow for real-time assessment of blood flow in the surgical field. The colour map and intensity function allowed easy identification of the superficial feeders, the draining veins and the nidus. Our study also to highlights this new capability of ICG semi-quantitative flow analysis in eight patients with AVM. 4.1. ICG videoangiography capture technique and limitations for flow analysis Originally, ICG videoangiography was designed as a qualitative method that was used to demonstrate the flow of ICG through the

arterial, venous and capillary phases of an AVM. The short and long replay function allowed the neurosurgeon to review the recording repetitively and to correlate it with white light video to generate a mental picture of the angio-architecture of the AVM in the operative field. This allowed for accurate identification of the arterial feeders, the draining veins and the nidus. It allowed for the quick formulation of a surgical strategy aimed at controlling the feeders and preserving the draining veins until the end of the surgery. During ICG video recordings, the neurosurgeon performed suction to clear blood and CSF which obscured the fluorescence of ICG in the field. There was a tendency to zoom and focus during recording to see the fluorescence in greater detail. The neurosurgeon could adjust the line of sight of the microscope or perform further dissection to allow for greater delineation of the ICG flow. These

230

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

Fig. 3. Flow fluorescence intensity curves during indocyanine green fluorescent videoangiography of the three regions of interest (red [1] = arterial feeder; green [2] = draining vein; green [3] = cortex).

the field. The structure on which ICG flow is to be demonstrated needs to be in the direct line of vision of the microscope. This is especially important if the AVM nidus is in the depth of a hematoma cavity to ensure maximum penetrance of the near infra-red light. It is necessary to properly adjust the zoom (5 ) and focus before recording begins. Proper coordination with the anaesthetist for the administration of ICG including timing, dosage and dilution is important. Once the recording has started, there should be no further interference (such as, from instrument manipulation, suctioning, shifting of the microscope). Any interference during the recording will lead to erroneous fluorescence intensity analysis. Planning ahead for comparison of pre- and post-excision ICG videos using a specific anatomical landmark for reference may be important. Proper washout of previous ICG fluorescence (around 10–15 minutes) between recordings is necessary. Such stringent standards for ICG videoangiographic recording may be restrictive and can reduce the ease of use of ICG videoangiography. However, this approach was optional for neurosurgeons who wished to perform semi-quantitative videoangiographic flow analysis. It may not be necessary for experienced neurosurgeons looking for qualitatively important information during intraresection. 4.2. Qualitative ICG interpretation

Fig. 4. Post-resection indocyanine green video showing no residual nidus in the surgical cavity in the illustrative patient with a left temporal arteriovenous malformation.

functions demonstrate the flexibility and usefulness of ICG videoangiography for the operating neurosurgeon. However, the added capability of flow analysis, based on the ICG videoangiography, requires additional surgical preparation. Cottonoids with fluorescence markers that are in the operative field should be removed or flipped over. Additionally, complete haemostasis should be secured, and CSF should be cleared from

Our experience with ICG videoangiography was generally similar to that of the published literature. ICG videoangiography was very safe, and there were no adverse events. The largest proportion of ICG videoangiography was performed pre-resection to delineate the angio-architecture of the AVM. ICG was used an intra-operative adjunct technique. Pre-resection ICG was useful in a confirmatory role in combination with various pre-operative imaging modalities and neuro-navigation. Pre-resection ICG was not helpful in 15% of patients, where the AVM nidus was obscured by a thin layer of adherent clot in the depth of a clot cavity. Intra-resection ICG videos, although less frequently recorded, were useful. The indications for intra-resection ICG were varied. Overall, five videos showed remaining, deep arterial feeders that required further dissection. In two videos, persistent flow in the main draining vein was shown. One video showed no flow in the AVM after temporary clipping of the arterial feeder and no swelling of the spinal cord. The varied reasons for intra-resection ICG videoangiography showed flexibility for the surgeon to seek specific answers during surgical resection.

Fig. 5. Post-resection CT angiography showed no residual arteriovenous malformation in the illustrative patient.

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

The post-resection (terminal) ICG were generally useful in showing no flow in the main draining vein and absence of a nidus in the surgical cavity. Of these, it was more important to demonstrate no flow in the main draining vein to indicate complete resection of nidus. Absence of the nidus in the surgical cavity may be indicative of complete resection, but it may also lead to a false of sense of assurance, especially if the nidus was fragmented by clots, diffuse, or covered by overlying brain parenchyma. The noted absence of flow in the main draining vein is not completely reliable because a remnant nidus may be draining into an unexposed deep draining vein. Hence, ICG is not a replacement for the gold-standard technique of intra-operative DSA to confirm completeness of resection.

231

volve the ability to capture fluorescence and flow in the white light view with no loss of anatomical detail. At the time of writing, the authors were aware that such a process was under investigation. Current flow analysis required ICG video to be captured with minimal interference. The algorithm should be enhanced to compensate for interference. Current flow analysis remained a semiquantitative method. Further calibration and quantification of fluorescent analytical capabilities will allow for accurate determination of AVM perfusion. A limitation of our study is that it is a descriptive, retrospective study in a small number of patients in a single institution. No direct comparison with other intra-operative tools such as intra-operative DSA was made.

4.3. Quantitative ICG interpretation Using the new flow analysis capability, we identified eight patients with superficial AVM and further identified the arterial feeder, the main draining vein and the normal cortex. The graphical display of average fluorescence against time for three ROI showed the time delay between the arterial feeder and the draining vein. The mean time difference between the feeder artery and the draining vein was 1.5 s, and it was 3.6 s between the feeder artery and the normal cortex. This function allowed us to measure and clearly differentiate the arterial feeders and the draining veins. It effectively removed any subjectivity in the interpretation of ICG videoangiography and white light video. The ability to abstract the data in a CSV format allowed for further statistical research on modelling predictions of flow. It is yet to be determined what factors are influencing fluorescence intensity. Although each patient was given the same dose of 25 mg of ICG, we noted a wide variance in average intensity. Factors that may determine the variance include intra-and inter-individual patient variation in cardiac output and blood pressure during the surgery, the flow rate of each individual feeder artery, the degree of venous outflow obstruction, the albumin level and lastly, microscope technical factors, such as degree of zoom, focal length and illumination intensity from the light source. Inter-patient variation is likely greater than intra-patient variation. Interestingly, the degree of decay was noted to be exponential and likely subject to hepatic excretory function. At the time of writing, the authors were aware that Kamp et al.20 had presented their findings of FLOW 800 intra-operative quantification of cerebral perfusion in 30 patients of different pathology, measuring parameters such as maximum fluorescence intensity, rise time, time to peak, cerebral blood flow index and transit times from the arteries to the cortex. 4.4. Limitations of ICG Despite improvement in flow analysis, ICG still has significant limitations. ICG signals could only be detected under direct visualisation. As such, a residual nidus may be missed if it is hidden by tissue or by a blood clot. This was evident in the single case we described where a residual nidus was found on post-operative catheter angiography despite the use of ICG intra-operatively. This limitation was significant because resection of AVM is usually most difficult in the deeper portions, which occur near the end of the long surgery. The current ICG videoangiography technique required a switch during surgery from a white light microscopic image to a black and white image for detection of ICG fluorescence and flow analysis. A certain degree of anatomical detail was lost when the screen image was darkened. To overcome such a problem, there is the ability for short and long replay for comparison of images as well as for picture-in-picture comparison. However, an ideal scenario would in-

5. Conclusion ICG videoangiography is a useful addition in AVM surgery. However, there are limitations. Flow analysis is an important new capability that allows for the semi-quantitative analysis of AVM perfusion. Conflicts of Interest/Disclosures The authors declare that they have no financial or other conflicts of interest in relation to this research and its publication. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jocn.2011.12.038. References 1. Dashti R, Laakso A, Niemela M, et al. Microscope integrated indocyanine green video-angiography in cerebrovascular surgery. Acta Neurochir Suppl. 2011;109:247–50. 2. Raabe A, Nakaji P, Beck J, et al. Prospective evaluation of surgical microscopeintegrated intraoperative near-infrared indocyanine green videoangiography during aneurysm surgery. J Neurosurg. 2005;103:982–9. 3. de Oliveira JG, Beck J, Seifert V, et al. Assessment of flow in perforating arteries during intracranial aneurysm surgery using intraoperative near-infrared indocyanine green videoangiography. Neurosurgery 2007;61(3 Suppl.):63–72. discussion 72–63. 4. Killory BD, Nakaji P, Gonzales LF, et al. Prospective evaluation of surgical microscope-integrated intraoperative near-infrared indocyanine green angiography during cerebral arteriovenous malformation surgery. Neurosurgery 2009;65:456–62. discussion 462. 5. Khurana VG, Seow K, Duke D. Intuitiveness, quality and utility of intraoperative fluorescence videoangiography: Australian Neurosurgical Experience. Br J Neurosurg 2010;24:163–72. 6. Takagi Y, Kikuta K, Nozaki K, et al. Detection of a residual nidus by surgical microscope-integrated intraoperative near-infrared indocyanine green videoangiography in a child with a cerebral arteriovenous malformation. J Neurosurg. 2007;107(5 Suppl.):416–8. 7. Hanggi D, Etminan N, Steiger HJ. The impact of microscope-integrated intraoperative near-infrared indocyanine green videoangiography on surgery of arteriovenous malformations and dural arteriovenous fistulae. Neurosurgery 2010;67:1094–103. discussion 1103–1094. 8. Kim EH, Cho JM, Chang JH, et al. Application of intraoperative indocyanine green videoangiography to brain tumor surgery. Acta Neurochir (Wien) 2011;153:1487–95. 9. Oh JK, Shin HC, Kim TY, et al. Intraoperative indocyanine green videoangiography: spinal dural arteriovenous fistula. Spine (Phila Pa 1976) 2011;36:E1578–80. 10. Woitzik J, Horn P, Vajkoczy P, et al. Intraoperative control of extracranialintracranial bypass patency by near-infrared indocyanine green videoangiography. J Neurosurg 2005;102:692–8. 11. Haga S, Nagata S, Uka A, et al. Near-infrared indocyanine green videoangiography for assessment of carotid endarterectomy. Acta Neurochir (Wien) 2011;153:1641–4. 12. Chen SF, Kato Y, Oda J, et al. The application of intraoperative near-infrared indocyanine green videoangiography and analysis of fluorescence intensity in cerebrovascular surgery. Surg Neurol Int 2011;2:42.

232

Y.P. Ng et al. / Journal of Clinical Neuroscience 20 (2013) 224–232

13. Fox IJ, Wood EH. Indocyanine green: physical and physiologic properties. Proc Staff Meet Mayo Clin 1960;35:732–44. 14. Owens SL. Indocyanine green angiography. Br J Ophthalmol 1996;80:263–6. 15. Kuroiwa T, Kajimoto Y, Ohta T. Development and clinical application of nearinfrared surgical microscope: preliminary report. Minim Invasive Neurosurg 2001;44:240–2. 16. Garski TR, Staller BJ, Hepner G, et al. Adverse reactions after administration of indocyanine green. JAMA 1978;240:635. 17. Wang S, Liu L, Zhao YL, et al. Strategy for assisted cerebral arteriovenous malformation surgery. Zhonghua Yi Xue Za Zhi 2010;90:869–73.

18. Dashti R, Laakso A, Niemela M, et al. Application of microscope integrated indocyanine green video-angiography during microneurosurgical treatment of intracranial aneurysms: a review. Acta Neurochir Suppl 2010;107:107–9. 19. Takagi Y, Sawamura K, Hashimoto N, et al. Evaluation of serial intraoperative surgical microscope-integrated intraoperative near-infrared indocyanine green videoangiography in patients with cerebral arteriovenous malformations. Neurosurgery 2012;70:34–42. 20. Kamp MA, Slotty P, Etminan N, et al. Intraoperative quantification of cerebral perfusion by microscope-integrated analysis of near infrared indocyanine green fluorescence angiography. Neurosurgery 2012;70:42–3.