Limit of intraoperative near-infrared spectroscopy monitoring during endovascular thrombectomy in acute ischemic stroke

Abstract

Endovascular thrombectomy is recommended for a persistent ischemic penumbra if recanalization cannot be achieved by the intravenous (IV) administration of recombinant tissue-plasminogen activator (rt-PA) alone. Although endovascular thrombectomy is a powerful treatment for major cerebral artery occlusion, the monitoring of recanalization and reperfusion during acute ischemic stroke presents a therapeutic challenge, and a previous study reported the usefulness of near-infrared spectroscopy (NIRS) for intraoperative monitoring during emergency endovascular thrombectomy for acute large ischemic stroke. Here we present our experience with a relevant case series. We applied NIRS monitoring during endovascular thrombectomy in two patients with large ischemic stroke following carotid artery occlusion and one patient with a non-large ischemic stroke caused by a distal middle cerebral artery (MCA) occlusion. In the patients with large ischemic stroke, complete recanalization of the internal carotid artery was achieved, and NIRS revealed a very good regional oxygen saturation (rSO₂) response. By contrast, in the patient with non-large ischemic stroke, the rSO₂ did not change, despite complete recanalization of the distal MCA. Our findings suggest the limited usefulness of intraoperative NIRS monitoring during emergency endovascular thrombectomy for non-large acute ischemic stroke caused by a distal MCA occlusion. However, intraoperative NIRS monitoring could be used practically to detect recanalization of the major artery during thrombectomy and early IV rt-PA administration in cases involving major artery occlusion.

Introduction

For acute ischemic stroke, the current standard of care comprises the intravenous (IV) administration of recombinant tissue-plasminogen activator (rt-PA) within 4.5 hours after onset.^1,2 In 2015, three randomized controlled trials reported that a combination therapy comprising IV rt-PA with endovascular thrombectomy significantly improved the functional outcomes of patients with major cerebral artery occlusion relative to IV rt-PA alone.^3–5 Endovascular thrombectomy is recommended as a subsequent treatment if recanalization cannot be achieved by IV rt-PA alone and an ischemic penumbra still exists.⁶ Although endovascular thrombectomy is a powerful treatment for major cerebral artery occlusion, it confers the risks of serious surgical complications, such as hemorrhage. Therefore, efforts to reduce surgical complications and thus increase the safety of this procedure are needed. One such effort involves the use of neurophysiological monitoring to detect early surgical complications during endovascular thrombectomy.

Near-infrared spectroscopy (NIRS), a noninvasive method used to measure regional oxygen saturation (rSO₂), could potentially detect changes in cerebral blood flow (CBF). Previous studies have already reported the usefulness of NIRS for the clinical monitoring of CBF,^7–9 and another report described the utility of intraoperative NIRS monitoring during an emergency endovascular thrombectomy for acute large ischemic stroke caused by a carotid artery occlusion.¹⁰ We have applied NIRS monitoring in three cases of acute stroke involving two patients with large ischemic stroke consequent to internal carotid artery (ICA) occlusion and one patient with non-large ischemic stroke caused by a distal middle cerebral artery (MCA) occlusion. All patients achieved complete recanalization via endovascular thrombectomy and stenting. Although NIRS revealed excellent rSO₂ responses to recanalization of the ICA in two patients, it failed to detect a response in the patient with the distal MCA occlusion. Accordingly, intraoperative NIRS monitoring during emergency endovascular thrombectomy, although useful for large acute ischemic stroke caused by carotid artery occlusion, may be limited regarding the detection of recanalization during emergency endovascular thrombectomy for non-large ischemic stroke caused by distal MCA occlusion.

Patients and methods

Three patients were admitted to the hospital with acute ischemic stroke. The baseline National Institutes of Health Stroke Scale (NIHSS) score and multimodal brain magnetic resonance (MR) images were obtained upon admission. One patient received IV rt-PA according to the current guidelines. Mechanical thrombectomy was performed using a SOLITAIRE revascularization device (Covidien) and TREVO revascularization device (Stryker) under local anesthesia. The Thrombolysis in Cerebral Infarction (TICI) score was used to assess recanalization (0 = no perfusion; 1 = penetration but no distal branch filling; 2a = perfusion with incomplete (<50%) distal branch filling; 2b = perfusion with incomplete (>50%) distal branch filling; and 3 = full perfusion with filling of all distal branches). Monitoring of the rSO₂ (INVOS® 5100c cerebral oximeter, Covidien) was initiated immediately before endovascular thrombectomy after placing an adhesive optrode on each side of the patient’s forehead according to the manufacturer’s recommendations.

Case reports

Case 1

A 71-year-old man with pneumonia and alveolar hemorrhage was initially admitted to another institute. A previous prescription of warfarin for atrial fibrillation (Af) was discontinued because he exhibited a prolonged international normalized ratio of prothrombin time. The patient became disoriented at 07:30, and MR images indicated acute ischemic stroke and occlusion of the left ICA. His symptoms were mild, with no aphasia or hemiparesis, and he was initially treated with sodium ozagrel and edaravone. However, he became unconscious and his stroke progressed at 13:45, and he was referred to our institute.

Upon arrival, the patient was somnolent and exhibited aphasia and hemiparesis. His NIHSS score was 22. Diffusion-weighted images (DWI) revealed hyperintensity in the left frontal lobe (Alberta Stroke Program Early Computed Tomography Scale (ASPECTS)-DWI: 10 points, (Figure 1(a)), and MR angiography revealed no signal corresponding to the ICA (Figure 1(b)). IV rt-PA administration was contraindicated because more than 4.5 hours had elapsed since stroke onset. The patient was transferred to the angiography suite for endovascular thrombectomy, where his left rSO₂ was found to be lower than the right rSO₂ (Figure 1(c))_. A massive thrombus was observed in the left common artery and ICA (Figure 1(d)). A suction catheter (Penumbra 5MAX ACE) was used to recanalize the left carotid artery after thrombectomy, and thrombus residue was detected in the MCA. Complete recanalization was achieved after performing a thrombectomy of the MCA using a stent-retriever (Solitaire FR) (TICI score = 3, Figure 1(e)). The left rSO₂ began to improve after carotid artery recanalization and fully recovered after complete recanalization (Figure 1(c)). Post-thrombectomy, the patient’s aphasia and hemiparesis improved but did not completely resolve. The patient was transferred to another institute for rehabilitation on day 54 with a modified Rankin Scale (mRS) score of 4.

Figure 1.

In Case 1, diffusion-weighted images show a hyperintense lesion in the left frontal lobe (a), and a magnetic resonance angiogram reveals a left internal carotid artery occlusion (b). Before thrombectomy, the regional oxygen saturation (rSO₂) (orange curve) was lower on the left side relative to the right (blue curve) (c). Left carotid angiograms reveal thrombi in the internal and common carotid arteries (d). The left middle cerebral artery was recanalized following recanalization of the carotid artery; accordingly, a complete recanalization was achieved (e). The left rSO₂ recovered to the same level as the right rSO₂ after complete recanalization.

Case 2

A 65-year-old man with multiple systemic atrophy was admitted to our institute with a four-day history of acute pyelonephritis. The patient was unconscious, and MR images revealed acute ischemic stroke with an ASPECT-DWI score of nine points (Figure 2(a)); MR angiography showed a left ICA occlusion (Figure 2(b)). The patient exhibited a conjugate deviation, aphasia, and hemiparesis. His NIHSS was 23, and remained unchanged despite the immediate administration of IV rt-PA. The patient subsequently underwent endovascular thrombectomy, before which the left rSO₂ was found to be lower than the right SO₂ (Figure 2(c)). A left carotid angiogram revealed an occlusion of the ICA (Figure 2(d)). Thrombus was retrieved from the ICA using a stent-retriever (Trevo® ProVue Retriever), and complete recanalization was achieved (TICI score = 3, Figure 2(e)). After recanalization, the left rSO₂ rapidly increased to the same level as the right rSO₂ (Figure 2(c)). With assistance, the patient could walk and participate in mutual communication. He was transferred to another institute for further rehabilitation on day 39 with an mRS score of 4.

Figure 2.

In Case 2, diffusion-weighted images show hyperintense lesions in the left caudate head and the corona radiate (a). A magnetic resonance angiogram reveals left internal carotid artery occlusion (b). Before the endovascular intervention, the regional oxygen saturation (rSO₂) was lower on the left side than the right side (c). A left carotid angiogram shows an internal carotid artery occlusion (d). Complete recanalization was achieved after thrombectomy (e). The left rSO₂ was restored to the same level as the right rSO₂ after recanalization.

Case 3

An 80-year-old man with no history of Af was admitted emergently with impaired consciousness upon waking. MR images revealed acute ischemic stroke with an ASPECT-DWI score of eight points (Figure 3(a)), and MR angiography showed a left distal MCA occlusion (Figure 3(b)). His NIHSS score was 10. As the onset was unknown, endovascular thrombectomy without IV rt-PA was performed immediately. Although the thrombus in the distal MCA was retrieved twice using a stent-retriever (Trevo XP), recanalization was not achieved despite the restoration of blood flow. As atherosclerosis was the suspected cause of the distal MCA occlusion, a stent (Enterprise® 2) was deployed to the occluded area. Complete recanalization was achieved following stent deployment (TICI score = 3, Figure 3(c)). However, the bilateral rSO₂ did not fluctuate considerably during endovascular thrombectomy, despite achieving a complete recanalization (Figure 3(d)). The patient was transferred to another institute for further rehabilitation on day 19, with an mRS score of 2.

Figure 3.

In Case 3, diffusion-weighted images show hyperintense lesions in the left temporal lobe and insula cortex (a). A magnetic resonance angiogram reveals a left distal middle cerebral artery (MCA) occlusion (b). A left carotid angiogram also shows a distal MCA occlusion ((c)-1). Although the thrombus in the distal MCA was retrieved twice using a stent-retriever (Trevo XP), recanalization was not achieved ((c)-2). However, stent deployment ((c)-3) led to complete recanalization ((c)-4). The left-right regional oxygen saturation (rSO₂) ratio did not change considerably during endovascular thrombectomy, despite the achievement of a complete recanalization (d).

Discussion

Jöbsis first reported that NIRS could be used to measure cerebral circulation and SO₂.¹¹ Oxy-hemoglobin (oxyHb) and deoxy-hemoglobin (deoxyHb) exhibit different absorption spectra. Therefore, NIRS uses the absorption coefficients of oxyHb and deoxyHb at different wavelengths to calculate changes in the concentrations of these molecules. Previously, NIRS was found to be useful for cerebral circulation evaluations in clinical settings. In cases of subarachnoid hemorrhage, the rSO₂ was found to decrease in a cerebral region containing a vasospasm in a supplying artery. Subsequently, the rSO₂ recovered after vasodilation of the spastic artery via the intra-arterial administration of papaverine hydrochloride or fasudil hydrochloride hydrate.^12,13 Some patients who underwent balloon test occlusion (BTO) experienced a decrease in rSO₂ when the stump pressure decreased to <45 mmHg,⁷ while all cases exhibited a decrease in rSO₂ at a stump pressure of <40 mmHg. Accordingly, a significant linear correlation was observed between the stump pressure and rSO₂ during BTO. Furthermore, NIRS was found to be useful for intraoperatively detecting the deterioration of neurological symptoms after carotid artery clamping during carotid endarterectomy under local anesthesia.⁸ The authors determined that a cut-off value of a 20% decrease in rSO₂ could be used to indicate cerebral ischemia, with a sensitivity and specificity of 83% and 83%, respectively.

In the field of interventional neuroradiology, NIRS was used for cerebral monitoring during aneurysmal embolization.¹⁴ In patients who had procedure-related cerebral vasospasm during embolization, rSO₂ significantly decreased, matching angiographic flow deterioration by vasospasm. It has been stated that NIRS may have a useful role for providing information about trend changes in frontal brain saturation related to angiographic vascular changes during aneurysm embolization. Hernandez-Avila et al. reported on the use of NIRS monitoring for cerebral angiography with and without neuroendovascular therapy.¹⁵ In the case with a carotid cavernous fistula, rSO₂ decreased when a balloon was inflated for occlusion of the fistula in the ICA. In diagnostic angiography, supplementary administration of oxygen increased rSO₂, and displacement of the nasal cannula was detected by a reduction in rSO₂. In a previous case report about embolization for a dural arteriovenous fistula, changes in rSO₂ monitored by NIRS may reflect venous hypertension.¹⁶ In another case with sinus thrombosis, cerebral oximetry involving NIRS was used for intraoperative monitoring of endovascular thrombolysis. Chromophore changes, including oxyHb and deoxyHb, correlated with angiographic flow and symptom resolution.¹⁷ In a study of intraoperative monitoring during carotid artery stenting (CAS), rSO₂ increased by 10% on the affected side after the procedure, and prolonged post-CAS increases in rSO₂ were observed in two patients who experienced clinical hyperperfusion syndrome.⁹ Taken together, these reports demonstrated that rSO₂, as monitored by NIRS, is reflective of CBF.

An earlier report described the usefulness of NIRS monitoring during endovascular thrombectomy for acute large ischemic stroke consequent to proximal MCA occlusion.¹⁰ Similarly, we successfully used NIRS to monitor rSO₂ in our two patients with carotid artery occlusion. However, NIRS monitoring was not useful during an endovascular thrombectomy for acute ischemic stroke caused by distal MCA occlusion. Locally, NIRS can monitor rSO₂ at a maximum depth of 2–3 cm from the sensors affixed to the scalp. Additionally, these NIRS sensors are usually affixed to the bilateral forehead because hair prevents the adhesion of sensors to other areas of the scalp. Therefore, NIRS sensors affixed to the forehead can monitor rSO₂ only in the frontal cortex, wherein blood flow is supplied mainly by the anterior cerebral artery and partly from the MCA. Therefore, NIRS monitoring during endovascular thrombectomy could be useful in cases involving the occlusion of major trunk arteries, such as the ICA and the proximal MCA (M1). When using NIRS to monitor distal MCA occlusions, the sensors could be affixed to the temporal scalp, thus potentially saving the patient’s hair. However, shadows from the sensors on the temporal scalp might impede the operative field. Regarding collaterals, forehead sensors could not sufficiently monitor rSO₂ if the contralateral A1 and anterior communicating artery are well developed and supply blood to the A2 on the occluded side.

Additionally, NIRS monitoring could be useful for the early detection of recanalization in cases of major artery occlusion treated with IV rt-PA. The onset-to-reperfusion time of an endovascular thrombectomy was reported to correlate with the clinical prognosis. Notably, the adjusted odds ratios for each 30-minute time increase were 1.21 for mortality (95% confidence interval, 1.09–1.34; p < 0.001), 0.79 for a favorable outcome (95% confidence interval, 0.72–0.87), 0.78 (95% confidence interval, 0.71–0.86) for excellent outcome, and 1.21 for intracerebral hemorrhage (95% confidence interval, 1.10–1.33).¹⁸ The initiation of NIRS monitoring at the time of IV rt-PA could facilitate rapid decisions regarding endovascular thrombectomy, which should be performed immediately (i.e. without an MR examination) for patients whose symptoms do not change and in whom the ipsilateral rSO₂ remains lower than the contralateral rSO₂ after IV rt-PA. These characteristics indicate that the major arteries remain occluded. However, MR imaging to evaluate recanalization should be performed before endovascular thrombectomy in patients who exhibit improved symptoms and a normalized rSO₂ after IV rt-PA, as in such cases the major arteries have possibly recanalized and angiography might be unnecessary and can therefore be omitted. Angiography after IV rt-PA confers a much higher risk of complications such as hematoma at the puncture site, compared with angiography alone. Finally, the avoidance of an unnecessary angiography could reduce the burdens placed on surgeons and staff, as well as the medical costs.

Conclusion

We have described our experiences with NIRS monitoring during emergency endovascular thrombectomy in two patients with carotid artery occlusion and one patient with distal MCA occlusion. Notably, we found that NIRS monitoring was useful during endovascular thrombectomy for a large stroke consequent to carotid artery occlusion, but not for an acute non-large ischemic stroke caused by distal MCA occlusion. NIRS monitoring might also be useful for the early detection of recanalization after IV rt-PA for an acute large ischemic stroke caused by major artery occlusion.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.