Staged intervention to enable the resection of a large left temporoinsular cystic glioblastoma with language preservation: illustrative case

Abstract

BACKGROUND

Resection of glioblastoma (GBM) in eloquent regions depends on functional mapping to limit perioperative neurological morbidity. When neurological deficits preclude reliable mapping, neurosurgeons should explore potential mitigation strategies. The authors present the case of a patient with a large left cystic temporoinsular GBM and aphasia, for whom the authors used intraoperative language mapping and a staged approach to enable safe tumor resection.

OBSERVATIONS

A 49-year-old female presented with progressive mixed aphasia for 1 month and new-onset right facial droop. Magnetic resonance imaging (MRI) revealed a large, heterogeneously enhancing, left temporoinsular tumor with a significant cystic component. Her aphasia was profound, and resection without reliable language mapping was deemed unsafe. An initial stereotactic tumoral cyst aspiration was performed, which reduced local mass effect and improved her language function. Cyst decompression thereby enabled both task-based functional MRI and intraoperative awake speech mapping, resulting in a safe resection of her GBM.

LESSONS

Safe resection of eloquently localized GBM is compromised when neurological deficits prohibit intraoperative functional mapping. This case demonstrates a mitigation strategy specific to cystic lesions in which an initial-stage stereotactic cyst aspiration is aimed at generating sufficient interval neurological improvement, such that intraoperative functional mapping can be performed during a second-stage resection.

https://thejns.org/doi/10.3171/CASE24362

Glioblastoma (GBM) is the most common primary brain malignancy in adults and is associated with a median survival of only 14–16 months.¹ The standard of care treatment paradigm involves maximal safe resection followed by concurrent chemotherapy and radiation.^1–3 Extent of resection is a critical predictor of overall survival and can be improved with intraoperative visualization adjuncts such as 5-aminolevulinic acid (5-ALA) or intraoperative magnetic resonance imaging (MRI).^{4, 5} For tumors in eloquent brain regions, additional functional surgical adjuncts, such as preoperative task-based functional MRI (fMRI), tractography, and intraoperative awake functional mapping, are used to better define and visualize the tumor’s relationship to critical functional tissue.⁶ In particular, the resection of tumors near essential language areas benefits from preoperative identification of hemispheric dominance and the localization of key cortical regions for expressive and receptive language function (usually on preoperative fMRI), as well as visualization of the tumor’s relationship to the arcuate fasciculus on tractography.⁷ Additionally, intraoperative localization of these critical language regions can be further refined using direct cortical and subcortical stimulation with continuous language assessments during awake surgery.⁶ However, in the presence of significant peritumoral edema, in which there is infiltration by tumor cells in GBM, tractography-based identification of subcortical fiber bundles becomes obscured. Moreover, the peritumoral edema and brain infiltration by tumor cells can compromise the accuracy of functional imaging and mapping, as neural plasticity in response to the tumor can alter the expected location of functional regions, complicating surgical planning and execution.^{8, 9} Finally, mass effect itself can generate language deficits that preclude reliable fMRI or intraoperative mapping. Recent advancements in resting-state fMRI (rs-fMRI) and tractography offer promising avenues to overcome these obstacles by enhancing the resolution and reliability of functional and structural brain maps, even in the presence of substantial edema or aphasia.^10–14

Although preoperative steroids can sometimes improve edema-related language deficits enough to permit fMRI and intraoperative language mapping, they have little effect on ameliorating deficits due to local tumoral mass effect. We present the case of a patient with a left temporoinsular cystic GBM, who initially presented with mixed aphasia that was severe enough to preclude participation in task-based fMRI and intraoperative language mapping. A staged approach to resection, starting with needle decompression of the tumoral cyst, generated sufficient restoration of language function to enable the patient’s participation in both preoperative task-based fMRI and intraoperative awake language mapping during a second-stage resection. This case exemplifies the efficacy of tumoral cyst aspiration as a mechanism for reducing tumoral mass effect, improving baseline neurological function, and enabling safer GBM resection in eloquent cortical regions.

Illustrative Case

A 49-year-old right-handed female presented with severe mixed aphasia that had rapidly worsened in the previous 24 hours and new-onset right facial droop. She had been having intermittent word-finding difficulty for several weeks, but the sudden worsening to an unrelenting mixed aphasia prompted her presentation to the emergency department. MRI of the brain revealed a 57.2-cm³ heterogeneously enhancing left temporoinsular mass with areas of central necrosis and a large posterior cystic component (approximately 25.5 cm³), concerning for GBM (Fig. 1). There was moderate surrounding vasogenic edema and substantial tumoral mass effect with 8 mm of left to right midline shift. Her profound mixed aphasia, which persisted after 2 days of high-dose dexamethasone, precluded both task-based fMRI and any meaningful attempt at awake resection guided by intraoperative language mapping. Without a reliable modality for functional language mapping and given the tumor’s dominant temporoinsular location, the potential neurological morbidity associated with aggressive tumor resection was thought to be prohibitive.

FIG. 1.

Imaging at initial presentation. Axial T1-weighted postcontrast MRI (A) of the brain in three planes shows a large complex heterogeneously enhancing mass centered in the left insula and extending into the left temporal lobe and frontal operculum. There is a large posterior temporal cystic component. The lesion compresses the left lateral ventricle causing approximately 8 mm of left to right midline shift. Corresponding axial T2-weighted MRI (B) from the same planes demonstrates the large T2-hyperintense posterior temporal cystic component and highlights the moderate edema surrounding the tumor. Coronal (C) and sagittal (D) T1-weighted MRI with contrast depicts the tumor’s involvement of the insula and the anteroposterior extent of the tumoral cyst in the temporal lobe.

The patient’s minimal response to high-dose steroid treatment implied that her language symptoms were more likely related to lesional mass effect rather than perilesional edema. We therefore hypothesized that aspiration of the tumor’s large and surgically accessible temporal cystic component could potentially improve her aphasia. As an initial surgical stage, a stereotactic biopsy of the tumor was performed in conjunction with a needle aspiration of the tumor’s cystic component (Fig. 2). An anterior trajectory was utilized to ensure both a safe biopsy of the nodular portion of the tumor and cyst aspiration. Approximately 20 mL of straw-colored cystic fluid was aspirated, which resulted in an obvious decrease in cyst volume on postoperative imaging (approximately 5.5 cm³; Figs. 2 and 3A–C). Postoperatively, the patient showed clear improvement in her language with reliable object naming, clear comprehension of complex commands, and intact repetition by postoperative day 2. She produced only occasional paraphasic errors. Her neurological recovery allowed for the acquisition of task-based fMRI, which confirmed left-hemisphere language dominance and localized several cortical regions involved in language generation and comprehension with respect to the tumor (Fig. 3B and C). Diffusion tensor imaging (DTI) tractography was also performed. A three-dimensional (3D) overlay of the arcuate fasciculus, the critical language areas identified by fMRI, and a rendering of the tumor are shown in Fig. 3C.

FIG. 2.

3D models generated with Brainlab platforms based on MRI show the total tumor (red) and cyst (purple) volume before (upper) and after (lower) the cyst aspiration procedure. On the right, a 3D model of the trajectory used for stereotactic cyst aspiration is shown superimposed on the 3D rendering of the cyst (purple).

FIG. 3.

Imaging performed after cyst aspiration. A: Axial T1-weighted postcontrast MRI after the stereotactic tumoral cyst aspiration, showing interval decompression of the temporal cyst. The residual heterogeneously enhancing and necrotic mass is still visualized in the left insula, temporal lobe, and frontal operculum. There is visible improvement in the compression of the ventricular system and midline shift. B: Axial (upper) and sagittal (lower) task-based fMRI highlight cortical areas involved in sentence completion, reading comprehension, and verb generation. The fMRI BOLD signal intensity changes corresponding to each language task (circles) are close to the tumor. C: A 3D model, generated with Brainlab, shows the tumor in relation to the arcuate fasciculus and the intersection of fMRI signals from various language tasks. D: Intraoperative photograph demonstrates hits (white squares) identified in language mapping. The sylvian fissure and frontal and temporal opercula are also identified.

With the newly obtained functional imaging and her sustained symptom improvement after cyst aspiration, the patient was able to undergo an awake craniotomy for definitive tumor resection with intraoperative language mapping just 3 days after the first-stage aspiration procedure. Intraoperative language mapping further refined a safe cortical entry zone for resection and more precisely identified the suspected critical language areas seen just anteriorly and posteriorly to the borders of the tumor on fMRI (Fig. 3D). Tumor resection was additionally supported by 5-ALA fluorescence guidance and intraoperative MRI. The insular portion of the tumor was accessed via a tubular retractor (VBAS, Vycor) to minimize retraction and potential injury to surrounding tissue.¹⁵ Resection of the deeper insular portion of the tumor was more conservative, as the patient’s fatigue increased and the reliability of her language assessments declined after 3 hours of awake surgical time. Postoperative MRI demonstrated an extensive subtotal resection, with a rim of residual enhancement at the medial and posterior tumor margins, which abutted an intraoperative language mapping “hit” (Fig. 4).

FIG. 4.

Intraoperative axial T1-weighted MRI without (A) and with (B) contrast from three different planes showing the interval left temporal craniotomy and a substantial subtotal resection of the large left temporoinsular tumor. Precontrast T1-hyperintense signal within the tumor cavity represents postsurgical changes and hemorrhagic blood products. The postcontrast images demonstrate some residual enhancement at the posterior and medial rim of the resection cavity. Coronal (C) and sagittal (D) T1-weighted postcontrast MRI provides additional views of the resection.

Postoperatively, the patient’s neurological examination was unchanged from her preoperative baseline, with intact object naming, comprehension, and repetition. Her language fluency continued to improve throughout the remainder of her hospital course. She was discharged on postoperative day 3 (hospital day 8) with a Karnofsky Performance Status of 80. The pathology report indicated an isocitrate dehydrogenase-wildtype, O-6-methylguanine-DNA methyltransferase–unmethylated GBM.

Approximately 3 weeks later, the patient presented with recurrent aphasia, similar to her initial presentation. Repeat MRI showed re-accumulation of the tumoral cyst along with the stable small area of residual enhancing tumor at the posteromedial edge of the resection cavity. She was trialed on a course of high-dose dexamethasone, with moderate symptom improvement, and soon after received concurrent radiation (60 Gy) and temozolomide. One month after she completed chemoradiotherapy, however, she again developed progressive aphasia due to enlargement of the tumor-associated cyst, which at this time was multiloculated. Twenty weeks after her resection, she underwent an additional surgery for endoscopic lysis of the septations within the cyst, aspiration of the cystic contents, and placement of an Ommaya reservoir in the cyst cavity. Her symptoms initially improved, but over the next few months, while receiving additional cycles of temozolomide, her condition declined again. Her Ommaya reservoir was found to be occluded, and tumor progression was noted on MRI. She underwent an additional craniotomy for cyst fenestration and tumor debulking with improvement in her symptoms.

Three months later, she experienced new-onset focal right arm and hand seizures, and imaging demonstrated rapid tumor regrowth with multifocal recurrence. Despite salvage treatment with laser interstitial thermal therapy and immunotherapy, the patient died approximately 10 months after her initial diagnosis.

Patient Informed Consent

The necessary patient informed consent was obtained in this study.

Discussion

Observations

Resection of brain tumors in eloquent cortical regions continues to require significant preoperative planning and remains a challenge due to the increasingly recognized individual patient variability in functional neuroanatomy.^16–19 Additionally, the characteristics of individual tumors can generate an additional layer of variability that can impact the surgical approach and associated risk. This case involved a cystic GBM in the dominant temporoinsular territory with mass effect that generated significant aphasia, which precluded awake tumor resection and functional mapping. We therefore formulated a staged intervention that exploited the safety and accessible location of the tumoral cyst. After stereotactic cyst aspiration, the patient’s language function improved enough to enable both preoperative task-based fMRI and a meaningful awake tumor resection with language mapping.

Symptomatic cyst accumulation is a well-described entity in GBM, with large cysts (i.e., those with a cystic component of at least 10 mL of volume) reported in about 10% of cases.^20–22 However, the prognostic significance of large cystic components in GBM is unclear. The concept of intratumoral cyst aspiration is not unique to GBM and has been previously described in other contexts as a means to increase the efficacy of subsequent tumor treatment or intervention. For example, initial-stage stereotactic cyst aspiration has been suggested to improve neurological symptoms in 70%–82% of patients with cystic brain metastases and additionally could improve subsequent local tumor control after radiosurgery.²¹ Cyst aspiration, however, usually only provides transient symptomatic relief, as tumoral cysts tend to reaccumulate. Described as a salvage maneuver after recurrence in our case illustration, the placement of an intracyst Ommaya reservoir can allow for serial tapping as needed. Reservoir placement has been shown to significantly reduce and even maintain a reduction in tumor volume, often by more than 50%, for brain metastases and can thereby enable additional localized treatments, including radiosurgery.^{23, 24} Nonetheless, an important limitation of this application of the Ommaya reservoir for intratumoral cysts is the high associated rate of catheter occlusion, as also noted in the context of cystic brain metastases.^{23, 24}

In the surgical management of tumors in eloquent brain areas, surgeons strive for a balance between maximizing the extent of resection and preserving function.²⁵ Advanced functional neuroimaging techniques have demonstrated significant utility during the preoperative planning stage, especially with the increasingly recognized anatomical variability in the cortical localization of eloquence among patients.²⁵ Advanced MRI techniques such as DTI tractography and task-based fMRI are often used in conjunction with the gold-standard intraoperative cortical mapping to minimize surgical neurological morbidity. DTI tractography helps to visualize white matter tracts and functional networks in the vicinity of intra-axial tumors.²⁵ Tractography, however, is known to become inaccurate in the presence of significant peritumoral edema,²⁵ a limitation that is yet to be overcome with ongoing software advances. Additionally, tractography relies on anatomical landmarks for seeding regions of interest, which may not capture the functional variability and tissue displacement introduced by brain tumors.^{26, 27}

Another useful preoperative functional imaging technique is fMRI, which uses changes in blood oxygen level–dependent (BOLD) signals to localize cortical regions of interest for language, motor, and sensory functions.²⁵ Eloquent brain regions identified through task-based fMRI have been shown to closely match the functional regions identified during direct cortical stimulation.²⁵ However, the concordance of fMRI data with intraoperative functional mapping for language is around 80%, a finding that underscores the value of the gold-standard intraoperative mapping and assessment.²⁸ In recent years, additional advanced functional imaging techniques have been developed that can potentially provide meaningful functional localization data despite disabling neurological deficits, and while not yet available at many institutions, they warrant further discussion. Rs-fMRI is a novel fMRI technique that can offer insights into the brain’s functional connectivity, as rs-fMRI is able to identify several networks simultaneously in contrast to task-based fMRI, which only provides data related to individual tasks.^{11, 25} Rs-fMRI does not depend on patient participation, holding great promise as a reliable functional imaging modality in neurologically debilitated patients, especially when its analysis and interpretation become more standardized.

An additional advanced neuroimaging technique that has recently gained traction, especially for motor mapping, is navigated transcranial magnetic stimulation (nTMS). Using a transcranial stimulator coil and peripheral surface electrodes on the patient’s face, arms, and legs to record motor evoked potentials, nTMS can provide noninvasive preoperative functional localization based on real-time responses to focal stimulation that can then be overlaid on stereotactic MRI for intraoperative neuronavigation.^{26, 29} Recent efforts have integrated nTMS and DTI tractography to map language networks, with one study describing gross-total resection in 72.9% of 107 patients with language-eloquent tumors with a relatively low rate of permanent language deficits (5.6%).²⁶

While these novel neuroimaging modalities have significant implications as neuronavigational surgical adjuncts, they are still in their infancy, and intraoperative functional mapping based on cortical and subcortical stimulation remains the gold standard, particularly for language mapping. In the context of eloquent cystic lesions, as occurs in a fraction of GBM cases, staged resection involving initial cyst aspiration followed by intraoperative functional mapping is a viable strategy, as in our case. Further studies are needed to elucidate the reproducibility of this staged intervention paradigm.

Lessons

This case describes a novel staged surgical paradigm aimed at enabling patient participation in intraoperative functional mapping to improve the resectability of cystic GBMs in eloquent cortical regions. The staged operative approach consists of stereotactic needle decompression of the tumoral cyst to restore function, followed by awake tumor debulking with language mapping to safely maximize the extent of resection.

Disclosures

Dr. Placantonakis reported joint ownership of a European Union and Hong Kong patent on the use of GPR133 as a treatment target in glioma, a patent application for anti-CD97 antibodies and antibody-drug conjugates, and consultant fees from Tocagen, Synaptive Medical, Monteris, Robeaute, Advantis, and Servier Pharmaceuticals.

Author Contributions

Conception and design: Placantonakis, Donaldson. Acquisition of data: Placantonakis, Donaldson. Analysis and interpretation of data: all authors. Drafting the article: Donaldson, Golub. Critically revising the article: all authors. Reviewed submitted version of manuscript: all authors. Approved the final version of the manuscript on behalf of all authors: Placantonakis. Administrative/technical/material support: Placantonakis. Study supervision: Placantonakis.

Correspondence

Dimitris G. Placantonakis: NYU Grossman School of Medicine, New York, NY. dimitris.placantonakis@nyulangone.org.