Convergence of the arcuate fasciculus and third branch of the superior longitudinal fasciculus with direct cortical stimulation–induced speech arrest area in the anterior ventral precentral gyrus

J Neurosurg 139:1140–1151, 2023

The objective was to identify the correspondence between the anterior terminations of the arcuate fasciculus (AF) and third branch of the superior longitudinal fasciculus (SLF-III) and the intraoperative direct cortical electrical stimulation (DCS)–induced speech arrest area.

METHODS The authors retrospectively screened 75 glioma patients (group 1) who received intraoperative DCS mapping in the left dominant frontal cortex. To minimize the influence of tumors or edema, we subsequently selected 26 patients (group 2) with glioma or edema not affecting Broca’s area, the ventral precentral gyrus (vPCG), and the subcortical pathways to generate DCS functional maps and to construct the anterior terminations of AF and SLF-III with tractography. Next, a grid-by-grid pairwise comparison was performed between the fiber terminations and the DCS-induced speech arrest sites to calculate Cohen’s kappa coefficient (κ) in both groups 1 and 2. Finally, the authors also demonstrated the distribution of the AF/SLF-III anterior projection maps obtained in 192 healthy participants (group 3) and subsequently correlated these with the speech arrest sites in group 2 to examine their validity in predicting speech output area.

RESULTS The authors found that speech arrest sites were substantially consistent with SLF-III anterior terminations (group 1, κ = 0.64 ± 0.03; group 2, κ = 0.73 ± 0.05) and moderately consistent with AF (group 1, κ = 0.51 ± 0.03; group 2, κ = 0.49 ± 0.05) and AF/SLF-III complex (group 1, κ = 0.54 ± 0.03; group 2, κ = 0.56 ± 0.05) terminations (all p < 0.0001). The DCS speech arrest sites of the group 2 patients mainly (85.1%) emerged at the anterior bank of the vPCG (vPCGa). In group 3, both terminations of AF and SLF-III converged onto the vPCGa, and their terminations well predicted the DCS speech output area of group 2 (AF, area under the curve [AUC] 86.5%; SLF-III, AUC 79.0%; AF/SLF-III complex, AUC 86.7%).

CONCLUSIONS This study supports the key role of the left vPCGa as the speech output node by showing convergence between speech output mapping and anterior AF/SLF-III connectivity in the vPCGa. These findings may contribute to the understanding of speech networks and could have clinical implications in preoperative surgical planning.

A Neurosurgeon’s Guide to Cognitive Dysfunction in Adult Glioma

Neurosurgery 89:1–10, 2021

Cognitive decline is common among patients with low- and high-grade glioma and can significantly impact quality of life. Although cognitive outcomes have been studied after therapeutic interventions such as surgery and radiation, it is important to understand the impact of the disease process itself prior to any interventions.

Neurocognitive domains of interest in this disease context include intellectual function and premorbid ability, executive function, learning and memory, attention, language function, processing speed, visuospatial function, motor function, and emotional function.

Here, we review oncologic factors associated with more neurocognitive impairment, key neurocognitive tasks relevant to glioma patient assessment, as well as the relevance of the human neural connectome in understanding cognitive dysfunction in glioma patients.

A contextual understanding of glioma-functional network disruption and its impact on cognition is critical in the surgical management of eloquent area tumors.

Intraoperative Cognition Mapping and Monitoring “à la carte” in Brain Tumor Patients

Neurosurgery 88:919–930, 2021

The purpose of surgery for brain tumors involving eloquent neural circuits is to maximize the extent of resection while preserving an optimal quality of life. To this end, especially in diffuse glioma, the goal is to remove the cerebral parenchyma invaded by the neoplasm up to the individual cortico-subcortical networks critical for brain functions.

Intraoperative mapping combined with real-time cognitive monitoring throughout the resection in awake patient is thus highly recommended to resume a normal life. Indeed, beyond avoiding hemiplegia or aphasia, enjoying a familial, social, and professional life implies that motor and language mapping is not sufficient.

Identifying and sparing neural networks that subserve cognition (movement control, visuospatial cognition, executive functions, multimodal semantics, metacognition) andmentalizing (theory ofmind, which plays a key role for social cognition) is essential to preserve an adapted behavior.

Here, the aim is to review when and how to map these critical functions, which have nonetheless been neglected for many decades by neurosurgeons. In fact, the disorders generated by surgical injuries of circuits underpinning nonmotor and nonspeech functions are usually not immediately visible on postoperative standard clinical examination, leading the physician to believe that the patient has no deficit. Yet, cognitive or emotional disturbances may subsequently prevent to resume an active life, as to work full time.

Therefore, a systematic neuropsychological assessment should be performed before, during, and after mapping-guided surgery, regardless of the tumor location, to preserve the functional connectome intraoperatively and to plan a postoperative tailored cognitive rehabilitation according to the patient’s needs.

Incidence and linguistic quality of speech errors: a comparison of preoperative transcranial magnetic stimulation and intraoperative direct cortex stimulation

J Neurosurg 134:1409–1418, 2021

Given the interindividual variance of functional language anatomy, risk prediction based merely on anatomical data is insufficient in language area–related brain tumor surgery, suggesting the need for direct cortical and subcortical mapping during awake surgery. Reliable, noninvasive preoperative methods of language localization hold the potential for reducing the necessity for awake procedures and may improve patient counseling and surgical planning. Repetitive navigated transcranial magnetic stimulation (rnTMS) is an evolving tool for localizing language-eloquent areas. The aim of this study was to investigate the reliability of rnTMS in locating cortical language sites.

METHODS Twenty-five patients with brain tumors in speech-related areas were prospectively evaluated with preoperative rnTMS (5 Hz, train of five, average 105% resting motor threshold) and navigated direct cortical stimulation (DCS; bipolar, 50 Hz, 6–8 mA, 200-μsec pulse width) during awake surgeries employing a picture-naming task. Positive and negative stimulation spots within the craniotomy were documented in the same MRI data set. TMS and DCS languagepositive areas were compared with regard to their spatial overlap, their allocation in a cortical parcellation system, and their linguistic qualities.

RESULTS There were over twofold more positive language spots within the exposed area on rnTMS than on DCS. The comparison of positive rnTMS and DCS (ground truth) overlaps revealed low sensitivity (35%) and low positive predictive value (16%) but high specificity (90%) and high negative predictive value (96%). Within the overlaps, there was no correlation in error quality. On DCS, 73% of language-positive spots were located in the pars opercularis and pars triangularis of the frontal operculum and 24% within the supramarginal gyrus and dorsal portion of the superior temporal gyrus, while on rnTMS language positivity was distributed more evenly over a large number of gyri.

CONCLUSIONS The current protocol for rnTMS for language mapping identified language-negative sites with good dependability but was unable to reliably detect language-positive spots. Further refinements of the technique will be needed to establish rnTMS language mapping as a useful clinical tool.

The Glioma-Network Interface: A Review of the Relationship Between Glioma Molecular Subtype and Intratumoral Function


Neurosurgery 87:1078–1084, 2020

Gliomas are a major cause of morbidity. Direct cortical stimulation mapping offers the ability to identify functional areas within the broader neural network both cortically and subcortically. Since the World Health Organization (WHO) 2016 classification categorized gliomas into molecular subgroups with varied molecular signatures and clinical behavior, it is possible that gliomas may demonstrate rates of functional network integration.

We therefore retrospectively reviewed a data registry of 181 patients with dominant hemisphere frontal, parietal, insular, or temporal gliomas. Our goal was to test the hypothesis that WHO glioma histopathology and molecular subtype influences functional language or motor sites identified within the tumor.

Intratumoral function as determined by direct cortical and subcortical stimulation mapping was identified at the highest rate in isocitrate dehydrogenase mutant astrocytomas and oligodendrogliomas. Finally, we reviewed the emerging literature exploring the interface between functional neural networks and gliomas. These data shed light on glioma molecular and histological characteristics most commonly associated within intratumoral function.

5-aminolevulinic acid (5-ALA) fluorescence guided surgery of high-grade gliomas in eloquent areas assisted by functional mapping

5ALA and functional mapping

Acta Neurochir (2013) 155:965–972

Only few data are available on the specific topic of 5-aminolevulinic acid (5-ALA) guided surgery of high-grade gliomas (HGG) located in eloquent areas. Studies focusing specifically on the post-operative clinical outcome of such patients are yet not available, and it has not been so far explored whether such approach could be more suitable for some particular subgroups of patients.

Methods Patients affected by HGG in eloquent areas who underwent surgery assisted by 5-ALA fluorescence and intra-operative monitoring were prospectively recruited in our Department between June 2011 and August 2012. Resection rate was reported as complete resection of enhancing tumor (CRET), gross total resection (GTR) >98 % and GTR>90 %. Clinical outcome was evaluated at 7, 30, and 90 days after surgery.

Results Thirty-one patients were enrolled. Resection was complete (CRET) in 74 % of patients. Tumor removal was stopped to avoid neurological impairment in 26 % of cases. GTR>98 % and GTR>90 % was achieved in 93 % and 100 % of cases, respectively. First surgery and awake surgery had a CRET rate of 80 % and 83 %, respectively. Even though at the first-week assessment 64 % of patients presented neurological impairment, there was a 3 % rate of severe morbidity at the 90th day assessment. Newly diagnosed patients had a significantly lower morbidity (0 %) and post-operative higher median KPS. Both pre-operative neurological condition and improvement after corticosteroids resulted significantly predictive of post-operative functional outcome.

Conclusions 5-ALA surgery assisted by functional mapping makes high HGG resection in eloquent areas feasible , through a reasonable rate of late morbidity. This emerges even more remarkably for selected patients.

Noninvasive Language Mapping in Patients With Epilepsy or Brain Tumors


Neurosurgery 72:555–565, 2013

Functional magnetic resonance imaging (fMRI) has become part of routine brain mapping in patients with epilepsy or tumor undergoing resective surgery. However, robust localization of crucial functional areas is required.

OBJECTIVE: To establish a simple, short fMRI task that reliably localizes crucial language areas in individual patients who undergo respective surgery.

METHODS: fMRI was measured during an 8-minute auditory semantic decision task in 28 healthy controls and 35 consecutive patients who had focal epilepsy or a brain tumor. Nineteen underwent resective surgery. Group and individual analyses were performed. Results in patients were compared with postsurgical language outcome and electrocortical stimulation when available.

RESULTS: fMRI activations concordant with the anterior and posterior language areas were found in 96% and 89% of the controls, respectively. The anterior and posterior language areas were both activated in 93% of the patients. These results were concordant with electrocortical stimulation results in 5 patients. Transient postsurgical language deficits were found in 2 patients in whom surgery was performed in the vicinity of the fMRI activations or who had postsurgical complications implicating areas of fMRI activations.

CONCLUSION: The proposed fast fMRI language protocol reliably localized the most relevant language areas in individual subjects. It appears to be a valuable complementary tool for surgical planning of epileptogenic foci and of brain tumors.