**Preoperative Repetitive Navigated TMS and Functional White Matter Tractography in a Bilingual Patient with a Brain Tumor in Wernike Area**

**Valentina Baro 1, \*, Samuel Caliri 1 , Luca Sartori 1 , Silvia Facchini 2 , Brando Guarrera 1 , Pietro Zangrossi 1 , Mariagiulia Anglani 3 , Luca Denaro 1 , Domenico d'Avella 1 , Florinda Ferreri <sup>4</sup> and Andrea Landi 1**


**Abstract:** Awake surgery and intraoperative neuromonitoring represent the gold standard for surgery

**\*** Correspondence: valentina.baro@unipd.it

of lesion located in language-eloquent areas of the dominant hemisphere, enabling the maximal safe resection while preserving language function. Nevertheless, this functional mapping is invasive; it can be executed only during surgery and in selected patients. Moreover, the number of neurooncological bilingual patients is constantly growing, and performing awake surgery in this group of patients can be difficult. In this scenario, the application of accurate, repeatable and non-invasive preoperative mapping procedures is needed, in order to define the anatomical distribution of both languages. Repetitive navigated transcranial magnetic stimulation (rnTMS) associated with functional subcortical fiber tracking (nTMS-based DTI-FT) represents a promising and comprehensive mapping tool to display language pathway and function reorganization in neurosurgical patients. Herein we report a case of a bilingual patient affected by brain tumor in the left temporal lobe, who underwent rnTMS mapping for both languages (Romanian and Italian), disclosing the true eloquence of the anterior part of the lesion in both tests. After surgery, language abilities were intact at follow-up in both languages. This case represents a preliminary application of nTMS-based DTI-FT in neurosurgery for brain tumor in eloquent areas in a bilingual patient.

**Keywords:** transcranial magnetic stimulation; brain tumor; bilingual; language; preoperative mapping; case report

### **1. Introduction**

Surgical resection of lesions involving the language pathway remains a major challenge for the neurosurgeon, harboring a risk of new functional deficits. Repetitive navigated transcranial magnetic stimulation (rnTMS) has proven to provide a reliable non-invasive preoperatory cortical mapping for language function, showing a good overall correlation with intraoperative direct cortical stimulation (DCS) [1–5]. Nevertheless, its sensitivity, specificity, negative and positive predicting values varies widely among studies. Therefore, rnTMS speech mapping is the only method that can replace DCS when the latter cannot be performed [6–9]. Subcortical tracts can be identified by diffusion tensor imaging-fiber tracking (DTI-FT) based on rnTMS mapping, obtaining an accurate and functionally oriented white matter preoperative study. In fact, it allows planning of the best surgical strategy for resection, improving postoperative outcome, especially in patients who are not eligible for awake surgery [8,10–15]. A detailed preoperative mapping of the language

**Citation:** Baro, V.; Caliri, S.; Sartori, L.; Facchini, S.; Guarrera, B.; Zangrossi, P.; Anglani, M.; Denaro, L.; d'Avella, D.; Ferreri, F.; et al. Preoperative Repetitive Navigated TMS and Functional White Matter Tractography in a Bilingual Patient with a Brain Tumor in Wernike Area. *Brain Sci.* **2021**, *11*, 557. https:// doi.org/10.3390/brainsci11050557

Academic Editors: Nico Sollmann, Petro Julkunen and Nadine Martin

Received: 15 March 2021 Accepted: 26 April 2021 Published: 28 April 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

pathway is mandatory, especially in case of bilingual patients, a peculiar subgroup that can present different patterns of cortical representation of the languages. In fact, the first language (L1) and the second language (L2) are processed both by shared brain areas as well as language-specific areas [16]. Moreover, even in L1 and L2 shared areas distinct language-specific neural population for the different languages have been identified by rnTMS [17]. Furthermore, Tussis et al. studied the cortical distribution of L1 and L2 in the non-dominant hemisphere with rnTMS, disclosing the involvement of dorsal precentral and middle precentral gyrus especially for L1, and triangular inferior frontal gyrus for L2 [18]. Whereby, a comprehensive preoperative understanding of the language pathway may be useful also in patients eligible for awake surgery, enabling a custom tailored craniotomy size and a faster and safer cortical mapping [2]. Herein we present the case of a 54-year-old Romanian woman affected by a primary brain tumor in the left angular gyrus who underwent preoperative rnTMS mapping to explore both Romanian and Italian languages. In the following, neurosurgical planning, surgical intervention and outcome are described and discussed.

#### **2. Case Presentation**

#### *2.1. Patient Information, Clinical and Radiological Findings*

A right-handed, bilingual 54-year-old woman was admitted at the emergency department for a generalized tonic clonic seizure sustained by a primitive brain tumor located between the posterior part of the superior and middle temporal gyri and the anterior part of the angular gyrus in the left hemisphere. The lesion did not enhance after contrast medium administration and it was hypometabolic at 18F-fluorodeoxyglucose PET/MRI. The functional MRI (fMRI) confirmed that the lesion was located in the dominant hemisphere (Figure 1). Due to the anxiety of the patient, mostly related to the diagnosis of brain tumor, the fMRI was performed testing only her mother tongue, i.e., Romanian. Interictal EEG showed an irritative activity in left centro–parietal derivations.

**Figure 1.** (**A**) 3D FLAIR (fluid attenuated inversion recovery) image discloses a primitive brain tumor located between the posterior part of the superior and middle temporal gyri and the anterior part of the angular gyrus in the left hemisphere; (**B**) the lesion does not enhance after contrast medium administration; (**C**) the 18F-fluorodeoxyglucose PET/MRI reveals the hypometabolism of the tumor (\*). (**D**) Axial T2w image fused with the BOLD (blood oxygenation level dependent) signal activation map obtained during word generation task shows a focal cortical activation in the superior–anterior part of the lesion (\*).

−

#### *2.2. Neuropsychological Evaluation*

Concerning the social and work surrounding the patient had been living in Italy for 17 years with her family, perfectly integrated in the social context, working as a housekeeper. Previously, she had 13 years of education, graduating in a vocational school in her home country.

The patient underwent a comprehensive battery of standardized neuropsychological tests performed in Italian, in order to evaluate the impact of the tumor on cognitive functions. A standardized evaluation of Romanian language was not executable because native language versions of the tests were not available and because none of the team spoke Romanian. The assessment was composed of tests covering different cognitive domains. The Oxford Cognitive Screen [19,20], a brief screening instrument composed of tasks on language, visual attention, spatial neglect, praxis abilities, visual and verbal memory, calculation, number reading and executive functions. Specific tests were also administered to better evaluate different cognitive functions. The Prose Memory Test (immediate and delayed recall) and Interference Memory test [21] were used as a measure of verbal memory. Forward and backward digit span and the Corsi block-tapping test were administered to measure short-term memory and working memory both for the verbal and visuospatial components [22]. Selective attention and switching abilities were measured using the Trail-Making-Test, forms A and B [21]. Different components of language abilities were assessed through specific tests: Phonemic Fluency test [21], the Boston Naming Test for visual naming ability [23], verbal comprehension of words and sentences and repetition of words and non-words [24]. Concerning language domain, the baseline preoperative assessment showed an impaired performance in naming and verbal fluency, whereas the other language abilities were normal (Table 1). Furthermore, the patient refused the proposition of an awake surgery. Therefore, we decided to test the patient for both languages by means of rnTMS integrated with DTI-FT. Due to her anxious state only the dominant hemisphere was evaluated, focused on the surgical planning.




**Table 1.** *Cont.*

TCS: correct score (the raw score is adjusted for age and education basing on Italian-normative data from the literature, when appropriate). E.N.P.A.: Esame neuropsicologico per l'afasia (i.e., neuropsychological examination for aphasia). NE: not executable. The impairment of the performance is defined basing on cut-off, from normative data from the literature.

#### *2.3. Patient's Informed Consent*

The patient signed specific informed consent for MRI acquisition, rnTMS tests, neuropsychological evaluation and surgical intervention.

#### *2.4. MRI Acquisition*

The patient underwent brain MRI according to a specific protocol designed for the nTMS and DTI-FT using a 3T scanner (Ingenia 3T, Philips Healthcare) to obtain 3D T1 weighted images (TR/repetition time = 8, TE/echo time = 3.7); 3D FLAIR/fluid attenuated inversion recovery (TR = 4800, TE = 299, TI/inversion time = 1650, flip angle = 40, matrix = 240 × 240 mm<sup>2</sup> , voxel = 1 × 1 × 1 mm<sup>3</sup> , 196 slices, 4.05 min of acquisition time); diffusion weighted sequences (DWI with 32 directions, TR = 8736, TE = 91; single shell, b = 800 s/mm<sup>2</sup> ) for DTI-FT.

#### *2.5. nTMS Language Cortical Mapping and Off-Line Analysis*

The 3D T1-weighted sequence was imported into the nTMS system (NBS system 4.3—Nexstim Oy, Elimäenkatu 9 B, Helsinki, Finland) for language mapping, performed thorough a repetitive stimulation (rnTMS) according to the most update indications [25,26]. The patient's resting motor threshold (RMT) was determined by applying nTMS to the left motor cortex representing the hand, detecting the motor response of the m. abductor pollicis brevis. The patient performed the language assessment (base-line test, rnTMS mapping) first in Romanian (in the presence of an interpreter) and then in Italian. The base-line test was performed twice without stimulation, in order to cross out from the

list the unfamiliar words, possible confounding variables in error analysis. A total of 80 black-and-white drawings of high and low frequency objects were presented on a 17-inch monitor placed 1 m in front of the patient for the picture naming task. Display and inter-picture time were set at 700 ms and 2500 ms, further adjusted to 2 s and 4 s for both languages. The patient was asked to say aloud the initial phrase "this is a . . . " to distinguish between a speech arrest and anomia [27]. At the end of the base-line test, 70 and 67 figures were considered for Romanian and Italian mapping, respectively. The rnTMS stimulation frequency was set at the beginning at 5 pulses at 5 Hz at 110% RMT and then increased to 10 pulses at 10 Hz at 100% RMT because with the previous parameters of stimulation we did not obtain any error. The stimulation coil was randomly moved between the presentation of the images in about 1-cm steps over the perisylvian and peritumoral cortex. The rnTMS pulse train automatically triggered with picture presentation (0 ms) [26]. The entire mapping session was recorded on video for off-line data analysis, performed by an expert neuropsychologist (S.F.), helped by an interpreter for the review of the test performed in Romanian. The errors were classified according to Corina et al.: semantic paraphasias, circumlocutions, phonological paraphasias, neologisms, performance errors and no response errors [28]. We considered a site as language-eloquent if at least two of three stimulations caused an error response [25]. The stimulation sessions were well tolerated with a minimal discomfort reported (Visual Analogue Scale 2/10).

The off-line analysis highlighted 39 performance errors in Romanian (320 spots tested) of which a group of 5 was located in the superior–anterior and posterior–inferior border of the lesion. In Italian, 2 semantic and 15 performance errors were detected (271 sites tested), 3 of them located in the anterior part of the tumor. The language maps showed a convergence of the errors in the anterior middle temporal gyrus, middle middle temporal gyrus, posterior middle temporal gyrus, ventral precentral gyrus and anterior supramarginal gyrus according to the cortical parcellation system as described in Corina et al. [29] (Figure 2). The latest convergence corresponds to the anterior part of the tumor.

**Figure 2.** (**A**) Romanian rnTMS mapping (green spots: performance errors) and (**B**) Italian rnTMS mapping (green spots: performance errors, blue spots: semantic errors). (**C**,**D**) show the anatomical distribution of all errors according to the parcellization system area described by Corina et al., in Romanian and Italian, respectively.
