The insular lobe is a common site of intrinsic brain tumor growth (1). Improvements in understanding the insular anatomy and the application of various technologies have made insular glioma operable with acceptable morbidity. Postoperative motor deficits remain a significant concern (2), and ischemia's incidence and risk factors after surgery for insular gliomas need further exploration.

For insular tumors, ischemic strokes can arise by lateral lenticulostriate arteries (LLSAs) from M₁ segment of the middle cerebral artery (MCA), long insular artery from the M₂–M₃ junction at the level of the superior insular sulcus, and even long medullary arteries terminating the M₄ branches (3–5). Although transitory clipping of the latter two types of arteries under motor evoked potentials (MEPs) monitoring may be helpful (3, 6), the characteristics of the thin, long course and ambiguous origin of these two arteries make their reservation still difficult intraoperatively.

LLSAs supply the internal capsule and often pass along the inner side of insular tumors (7). Although the number of LLSAs varies from 1 to 21, even the occlusion of one branch may cause extensive infarction of the subcortical ganglia and internal capsule, resulting in motor and language deficits (8). Immediate postoperative deficits may be caused by resection-induced contusion, edema, and hypoperfusion. However, permanent deficits are mainly caused by infarctions due to disruption of MCA branches on the lateral surface of the insula and LLSAs within the anterior perforated substance on its mesial side (9, 10). In the transsylvian and transcortical approaches for insular glioma resection, the preservation of these critical blood vessels is essential (11). The characteristics of the relatively large diameter and exact origin of LLSAs, make its reservation possible intraoperatively.

This study tried to determine the incidence of ischemic complications and their risk factors, and describe the strategies which focus on LLSAs and MCA preservation employed to prevent the ischemic complications of insular glioma surgery.

We found that ischemic complications were common in insular glioma resection in the study. Additionally, patients whose tumors had the obscure inner edge and whose MEPs declined by >50% intraoperatively were at high risk for developing intraoperative strokes that resulted in motor deficits. Further, we described our strategies for arterial preservation during tumor resection.

In our patients, a transcortical approach was used for better exposure and suitability for glioma resection beyond the insula (20). The transcortical windows in our patients were limited to the pars orbitalis and triangularis of the inferior frontal gyrus. Almost all patients with postoperative speech disorders recovered their capacity for simple daily communication, suggesting that language plasticity might play a role (21). Mostly, the pars triangularis does not always involve the Broca’s area, especially for tumors located in the dominant hemispheres (10). To ensure motor function in patients with insular glioma, preservation of the arteries is critical.

The vascular supply to the insular is mainly derived from the M₂ branches that overlie the insular surface (22). The course of the M₂ segment along the insular surface constitutes a rich arterial network, which is a substantial obstacle in accessing the insular region (20, 23, 24), especially when the arterial network is encased in, distorted, or even wrapped around by the tumor. LLSAs may also be encased by the tumor and serve as a source of vascular supply to the tumor. The long insular artery and long medullary arteries are usually thin, long course, and have ambiguous beginnings. All these characteristics of insular gliomas make maximal resection difficult and increase the risk of ischemic events. Moreover, the insula is largely and frequently supplied by perforating arteries, with no collateral blood supply (25). New evidence suggests that motor deficits often arise from ischemia of the pyramidal tract due to compromise of the blood vessels supplying to the corona radiata and the internal capsule (18).

Moreover, most insular glioma resections affected LLSAs, long insular arteries, and even long medullary arteries terminating the M₄ branches (3–5). Various mechanisms for developing intraoperative brain ischemia have been proposed, including direct vascular damage, vasospasm, and kinking of the arteries by brain retraction (18, 26). Therefore, preservation of these arteries is essential in insular glioma resection.

Ischemic complications of insular glioma resection with a transcortical approach are commonly detected on postoperative MRI, with incidence ranging from 1.4% to 75.4%, and are a significant source of neurological morbidity (12, 21, 26–37). The primary series of procedures for insular glioma resection using the transcortical approach are presented in Table 5. A series of 255 procedures utilizing the transcortical approach had a short-term complication rate of 12.79% and a long-term complication rate of 15.70%, 66.7% of these long-term complications could be attributed to the LLSAs and MCA infarction (30). In 61 of Pallud's 111 cases of insular gliomas who underwent awake transcortical approach, diffusion-weighted hyperintensity was observed in 39 postoperative MRI (75.4%): lenticulostriate artery territory in 5 cases, M₂/M₃ perforator territory in 2, and anterior choroidal artery territory in 1 (1.6%) (34). These eight deep infarcts were systematically associated with a motor deficit. In Mandonnet's series of procedures using transcortical windows for 12 patients, only 1 patient had severe dysarthria associated with a right facial palsy which resolved at 4 months; however, a slightly reduced speech rate was observed (31). Of these 12 patients, none had small areas of ischemia. In a cohort study by Przybylowski et al., which analyzed 48 patients who underwent a transcortical approach, 10 patients (21%) in the transcortical group showed evidence of ischemia on postoperative MR images (27). At the 4- to 6-week clinical evaluation, eight patients (17%) in the transcortical group demonstrated new or worsened neurological deficits. At 1 year postoperatively, four patients (8%) in the transcortical group had persistent neurological deficits. As for insular glioma surgery in this study, ischemic lesions were found in 44 (58.7%) patients using DWI, and core ischemia was noted in 11 of the 44 patients.

Although the incidence of ischemia in our study was slightly higher than that in previous studies (26), it was still within the known range following insular glioma surgery. According to previous research, the insular tumor location is the most potent risk factor for the development of intraoperative ischemia, affecting almost all patients with ischemia (26). In our study, multivariate analysis showed that core ischemia was significantly associated with an obscure inner edge sign. This may be explained by the flat inner edge between the tumor and basal ganglia, wherein the tumor did not invade the basal ganglia; hence, the tumor could be totally resected, and the risk of LLSA damage was low. The presence of an obscure inner edge sign corresponded to a significantly increased risk of LLSA damage. In our study, four patients with less than 90% tumor removal had core ischemia. All the preoperative MRIs of these four patients showed obscure inner edge signs, which may be the main reason for core ischemia.

Tumor location and size may be a significant factor in postoperative risk, while in our study, tumor preoperative volume was not significantly associated with core ischemia, as shown in Table 1. Preoperative tumor volume was a significant predictor of EOR, which is in line with Eseonu’s study (28). Insular tumors localized in zones II and III present significant surgical difficulties, a minor extent of resection, and a higher rate of permanent morbidity (35). Also shown in our study, zone II is a major risk factor for core ischemia and paralysis.

Preservation of the LLSAs during microsurgical resection of insular gliomas is paramount. A few strategies may be effectively employed to enhance the probability of preventing iatrogenic injury to these vessels. Digital subtraction angiography was used to assess the anatomy of the vessels preoperatively (38), while indocyanine green fluorescence helped in identifying and preserving the long perforating branches of the MCA (39). Intraoperative ultrasonography may be facilitated by using contrast agents to improve LLSA visualization (40). Fluorescence-guided surgery with the use of 5-aminolevulinic acid may also help delineate the tumor at the medial edge of dissection for high-grade lesions (41). Intraoperative MRI may help assess the degree of resection and residual tumor at the medial border, which may allow further tumor resection (42). Subcortical motor mapping is performed in the medial plane of the resection to identify the internal capsule (11). However, these methods have several limitations. Notably, the number of LLSAs determined by each method is variable. Moreover, the identification of LLSAs and basal ganglia under a microscope is also essential. In combination with the surgeon’s experience, modern technology enables the safe resection of insular tumors, regardless of their anatomical and functional complexities.

Our continued experience with insular tumor resection has convinced us that the key to the resection of insular lesions is the preservation of the LLSAs and main branches of the MCA. In our experience, different strategies are used, depending on the tumor's location, according to the Berger–Sinai classification. Preservation of the proximal segment of the LLSAs in the anterior location is paramount. If the tumor encases the proximal segment of the LLSAs, it is difficult to completely resect the tumor and the risk of motor deficits increases (43). Given our experience in such situations, we made the “residual triangle” at the proximal segment of the LLSAs to preserve them. We conducted this procedure to avoid distortion, which may cause ischemia (44). In this study, the multivariate analysis did not show a significant association between the tumor encasement of the proximal segment of the LLSAs and core ischemia. Furthermore, the mean volume of the “residual triangle” was <2% of the tumor volume and had little influence on the EOR. This may confirm the effectiveness of this strategy in preserving proximal LLSAs. Moreover, the preservation of distal LLSAs is essential for tumors that are located posteriorly. Since the basal ganglia might have been invaded by the insular gliomas if there was no clear border between the tumor and basal ganglia, the “basal ganglia outline reappearance” was created based on the texture of the basal ganglia and distal branches of the LLSAs, which not only helped prevent direct injury to the posterior limb of the internal capsule but also preserved the distal branches of the LLSAs. When the texture of the basal ganglia and distal branches of the LLSAs was identified to guide the depth of resection, direct injury to the white matter motor fibers was unusual (18), and the tumor could be maximally and safely resected. Finally, in all tumor resection procedures, preservation of the main branches of M₂ is essential. In a subpial fashion, the “sculpting” technique was used to outline the frame of the main branches of M₂. In this way, we could cut off the insular arteries and preserve the arteries that merely passed through the tumor, i.e., the branches of M₂.

Regarding the possible limitations of these strategies, the volume of the “residual triangle” should be left as small as possible due to the possibility of tumor relapse. Skrap et al. also advocated that they would leave a minimal part of the tumor around the LLSAs in case they were encased by the tumor (29). In the process of creating the “basal ganglia outline reappearance,” vasospasm might occur in the distal branches of the LLSAs, especially in patients with arteriosclerosis; in our study, ischemia related to LLSAs occurred in 10 patients, 7 of whose ischemia located in the head of caudate nucleus and anterior limb of internal capsule, which caused by the anterior branch of LLSAs; thus, a thin layer of the tumor or the external capsule left on the surface of basal ganglia was necessary.

While injury to the LLSAs accounts for the majority of postoperative hemiparesis following insular glioma resection (45), injury to the long insular artery is not infrequently implicated. Long insular artery and long medullary arteries terminating the M₄ branches, which are characteristics of the thin, long course, and ambiguous origin, are difficult to be preserved intraoperatively. As in our study, eight deep-seated ischemia were caused by the long insular artery. Besides the subpial resection and transitory clipping of these vessels under MEP monitoring, Ikegaya et al. describe a strategy in which a small piece of gray matter is spared at the bottom of the peri-insular sulcus, where long insular artery and long medullary arteries pass en route to the pyramidal tract (46), which is the same principle as our “residual triangle.”

Intraoperative detection of an impending stroke is essential because a prompt response may make the decline transient instead of causing permanent deficits (26). Although intraoperative MRI is efficient in estimating the EOR, its use in detecting ischemic lesions is not recommended (47). Multiple studies verified that IONM parameters, including MEP and SSEP, are helpful and reliable for predicting and preventing ischemic brain injury during neurosurgery (48). Generally, MEP showed better diagnostic accuracy than SSEP (49). The monitoring of intraoperative MEPs revealed early ischemia and aided in preventing ischemia from becoming permanent through a therapeutic response such as holding surgery, irrigating with warm saline, and releasing retraction of the brain parenchyma (18, 50). In this study, a 50% MEP decline was significantly associated with core ischemia, postoperative paralysis, and 6 months of paralysis but not noncore ischemia. Thus, continuous IONM is essential during insular glioma resection. In our study, 4 of 11 patients with MEP abnormalities had no paralysis. Of these four patients, two with paralysis immediately after surgery recovered fully within 7 days. The other two patients may be false-positive. Abboud et al. reported a false-positive rate of 1.1% of MEPs for outcome prognostication during surgery for supratentorial lesions (51). As to 4 of 11 patients with MEP normal who had core ischemia, three of these four patients were indeed with a decline in MEP amplitude >50% intraoperatively, while withholding surgery, irrigating with warm saline, and releasing retraction of the brain parenchyma, the amplitude of these three patients recovered. Although these three patients had no paralysis, they did have core ischemia. The other patient may be with false negative, which is caused by edema in peri-resectional regions (52). After all, there is no immediate postoperative MRI scan.

Recently, awake craniotomy with intraoperative cortico-subcortical direct electrical stimulation has become an increasingly accepted and recommended technique worldwide, particularly for lesions in the dominant hemisphere (53, 54). In our study, although 4 of the 34 patients with left-sided tumors had long-term aphasia, these patients could still carry out a basic conversation. This study aimed to emphasize skill techniques for vascular preservation. If these vascular preservation strategies are applied in awake craniotomy, intraoperative direct electrical stimulation, and even network-level approaches (55), maximal insular tumor resection may be achieved, and better cognitive and emotional functional outcomes may be preserved.

Insular glioma resection is associated with a high incidence of ischemia, as detected by DWI, as well as new motor deficits that were determined by the treating neurosurgeons. Insular glioma patients with obscure inner edge signs and intraoperative MEP decline >50% had a higher risk of developing core ischemia. With strategies of LLSAs and main branches of MCA preservation, such as “residual triangle” at the proximal segment of the LLSAs, “basal ganglia outline reappearance” at the distal segment of LLSAs, and “sculpting” technique for the main branches of M₂, maximal safe resection of insular gliomas may be achieved. Nevertheless, ischemia and motor deficits occurred, even with artery-preserving strategies intended to minimize such risks. Both function and vessel protection strategies need further exploration.