*2.7. Presurgical Planning*

According to the rnTMS results, the tumor was divided into an eloquent and noneloquent part, the latter identified as our surgical target. Using the planning station, the anterior part was highlighted in red and the target in violet. Then, the final reconstruction of language network was imported into the neuronavigation system to assist surgery (Figure 4).

**Figure 4.** rnTMS-based planning using StealthViz software. (**A**) The anterior and eloquent part of the lesion identified by nTMS is colored in red and the posterior non-eloquent part in violet, (**B**) 3D FLAIR anatomical images for comparison. (**C**,**D**) A subtotal resection was planned and guided by neuronavigation (in violet the non-eloquent part).

#### *2.8. Surgical Intervention and Neuropsychological Follow Up*

Surgical resection of the posterior non-eloquent part was achieved by neuronavigation because the lesion was not clearly distinguished from normal brain parenchyma. Integrated histological and molecular diagnosis disclosed a WHO-grade IV gliomas [35]. She received perioperative antiepileptic drugs prophylaxis. Moreover, the patient received dexamethasone 4 mg four times daily for one week followed by gradual tapering. Postoperative neuropsychological assessment, performed after one week, showed a global worsening of the performance in language tasks (reading, number writing) and in other cognitive functions (praxical function in right hand, short- and long-term verbal memory, verbal working memory). However, the follow-up evaluations performed at 1 and 4 months after surgery, revealed a restoration of functions through the time. The performance at four months after the surgery was comparable with the baseline (Table 1). Relatives reported intact native language performance as well. The patient underwent whole brain radiotherapy (60 Gy/30 fractions) and medicated with Temozolomide (two cycles). The patient did not present seizures at last follow-up (10 months).

#### **3. Discussion**

Despite awake surgery associated with DCS still represents the gold standard for language mapping, an accurate preoperative assessment of language pathway is required to establish the best surgical strategy, for the risk–benefit balance and for the patient's counselling [36–38]. This is mandatory especially in case of patients with lesions located in eloquent areas who are not eligible for awake surgery [39]. Commonly, fMRI is the most accessible and applied preoperative mapping technique, providing the identification of eloquent cortical areas for different types of functions. Nevertheless, the indirect signal of area activation provided near a brain lesion could be undermined by a metabolic uncoupling induced by the lesion itself, determining a reduced fMRI signal in perilesional eloquent cortex [40–42]. This phenomenon, associated to a normal or increased activity in homologous brain regions, can simulate a reorganization of the function [43]. Moreover, previous studies have not clarified the reliability of fMRI for preoperative language

mapping in tumors located in language-eloquent areas [42,44,45] thus, the use of fMRI in adjunct to other mapping methods is suggested [3,6,46].

nTMS is a recent and promising preoperative mapping technique for cortical functions localization and the development of nTMS-based DTI-FT allows a functionally oriented white matter reconstruction. In fact, the white matter reconstruction based on the rnTMS mapping showed a more accurate and reliable reconstruction of the subcortical language pathway compared to the standard anatomical technique [10,13]. Nonetheless, few centers have a broad experience with this technique and the language mapping has been less investigated compared to the motor nTMS mapping [6]. This may reflect the fact that language function is the result of a complex cortical and subcortical network which is more difficult to localize and challenge to map [28,47–49]. Currently, rnTMS combined with nTMS-based DTI-FT could be remarkably useful for patients who are not eligible for awake surgery, providing information concerning the true eloquence of the lesion with a high specificity of rnTMS in localizing language-negative areas. Furthermore, this technique can identify the presence of intra-hemispheric tumor-induced plasticity [14] or inter-hemispheric function reorganization and/or migration involving the non-dominant hemisphere [50–52].

For bilingual and multilingual neurosurgical patients affected by lesion located in language-eloquent areas, the preoperative mapping and languages preservation represent an additional major goal. At present, bilingual (and multilingual) neurosurgical patients have been investigated mostly with DCS and fMRI, as highlighted in a very recent review by Polczynska and Bookheimer [16]. This review suggests several principles concerning languages organization in bilingual patients, which may be useful in predicting the likelihood of separate versus converging representation of languages (i.e., age of L2 acquisition, proficiency level of L2 and linguistic distance between L1 and L2). Nonetheless, fMRI may falsely identify certain brain regions as potentially eloquent as above mentioned. Moreover, DCS evaluates a restricted coverage within language areas, mostly focused on sites in the frontal or posterior languages eloquent pathway. Unexpectedly, rnTMS has not been applied to study neurosurgical bilingual patients so far and late bilingual population has not been investigated.

We described the case of a bilingual patient affected by brain tumor located in language-eloquent region. The patient refused to undergo an awake craniotomy. Therefore, in order to deal with the aim of a safe resection, we applied an alternative method that could offer an accurate mapping of both languages. Preoperative languages mapping was obtained by rnTMS and nTMS-based DTI-FT language assessment according to the protocol for rnTMS language mapping used at our institution and established in the literature [11,25,53]. Language mapping for Romanian (L1) and Italian (L2) showed a convergence in the posterior areas of language pathways [48]. This overlapping may be explained by the high proficiency of L2 identified by extensive neuropsychological assessment and by the common derivation from Romance language as previously described. Furthermore, the language mapping disclosed the true eloquence of the anterior part of the lesion for both L1 and L2, limiting the surgical target to the non-eloquent region. Despite the worsening of language tasks and in some cognitive functions, the short-term follow-up highlighted a restoring of functions, comparable with the baseline. Probably this transient worsening was imputable to the surgical manipulation of subcortical fibers producing a functional, rather than anatomical damage. When performing language assessment in a bilingual patient, the setting should consider the presence of an interpreter both during rnTMS and off-line analysis, if the L1 is not properly known by one of the clinical staff members. Ideally, the native language version of the neuropsychological tests should be available and administered by a properly educated interpreter, to achieve a greater accuracy. This can be considered as an intrinsic limitation and possible bias when analyzing different languages, but it can be overcome by the advantages of an accurate mapping of the currently speaking languages, which requires to be preserved. Moreover, rnTMS still presents other pitfalls that need to be assessed. In fact, the likelihood of detecting language-positive

spots is still low, drawing attention to the necessity of a revision of current stimulation protocols [49]. Furthermore, the interpretation of the hesitation errors varies among authors, constituting a matter of debate [49,54,55] and, in addition, the pre-existing moderate aphasia or severe cognitive impairment could undermine the reliability of the examination, entailing an accurate patients selection [56]. Nonetheless, the use of the initial sentence during picture naming, helping to distinguish between speech arrest and anomia is not routinely applied [25,27]. Another point is the influence of the antiepileptic drugs on the cortical excitability, which may influence the stimulation threshold as reported for the motor cortex but not investigated for extra-motor cortex [57]. Regarding the functional tractography obtained from the language mapping, a meaningful and debatable protocol should be assessed [10,11,13].

#### **4. Conclusions**

Our experience showed the reliability of rnTMS mapping in a bilingual patient who required surgery for a language-eloquent lesion for both languages. The potentials of this technique are different. First of all, the clinical application in safe neurosurgical practice is clear, because it represents a good tool for pre-surgical mapping, when awake surgery is not applicable for different reasons and rnTMS may allow filling of this gap. Furthermore, in the specific case of brain tumor, the preoperative mapping with nTMS-based DTI allows a better comprehension of language pathway reorganization and plasticity. A second important application concerns the neural basis of language, and bilingualism in particular, which remain still unclear in the literature. In this context, further studies with rnTMS on bilingual patients and healthy subjects are advocated to a comprehensive study of languages organization and plasticity.

**Author Contributions:** V.B.: methodology, data curation, formal analysis, writing—original draft, writing—review and editing, S.C. and L.S.: methodology, data curation, formal analysis, software, investigation, S.F.: data curation, formal analysis, investigation, writing—review and editing, B.G.: resources, software, writing—review and editing, P.Z.: resources, software, M.A.: supervision, validation, L.D., D.d.: supervision, F.F., A.L.: supervision, validation, writing—review and editing. All authors have read and agreed to the published version of the manuscript.

**Funding:** The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

**Informed Consent Statement:** The patient signed her specific informed consent for MRI acquisition, nTMS tests, neuropsychological evaluation and surgical intervention. Moreover, she authorized the publication of this case report after anonymization of data and images.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author.

**Acknowledgments:** The authors thank Francesca Baro for the linguistic revision of the manuscript and Gianluigi De Nardo, neurophysiology technician, for his assistance during the tests.

**Conflicts of Interest:** The authors declare no conflict of interest.

**Ethical Approval:** For this type of paper, a case report, no approval of the local ethical committee was deemed necessary. The work is in accordance with the declaration of Helsinki and its later amendments, for as far applicable.

#### **Abbreviations**

AF: arcuate fascicle; DCS: direct cortical stimulation; DTI-FT: diffusion tensor imaging-fiber tracking; FAT: frontal aslant tract; IFOF: inferior fronto-occipital fascicle; ILF: inferior longitudinal fascicle; fMRI: functional MRI; L1: first language; L2: second language; nTMS: navigated navigate transcranial magnetic stimulation; RMT: resting motor threshold; rTMS: repetitive transcranial magnetic stimulation; rnTMS: repetitive navigate transcranial magnetic stimulation; SLF: superior longitudinal fascicle; UF: uncinate fascicle.
