Abbreviations:

SLF — superior longitudinal fasciculus

AF — arcuate fasciculus

FAT — frontal aslant tract

The first awake surgeries of the brain were performed in the first half of the XX century by an outstanding Canadian neurosurgeon W. Penfield on patients with epilepsia. During the surgery, he stimulated the surface of the cerebral cortex with electrical discharges and recorded his observations. After gaining experience and performing more than 750 awake surgeries, W. Penfield formulated the modern description of the topography of cortical representations of motor and sensory functions. Later G. Ojemann used awake surgeries to map the white matter pathways associated with the cortical speech zones. In particular, he described nominative (amnesic) aphasia after subcortical stimulation of the parietal lobe and development of sudden neurological deficit when performing memory tasks during the stimulation of thalamic structures [1].

Direct stimulation of the cortical zones and pathways is currently the standard in resection of intracerebral tumors. According to the European recommendations [1], an expansion of tumor resection significantly prolongs the overall survival of patients with gliomas of both low and high malignancy. On the other hand, brain functions must be mapped to preserve patient’s quality of life. Therefore, in case of diffuse tumors, such as gliomas, the main task is to remove the part of the brain affected by tumor cells, based on individual functional boundaries rather than "oncological" boundaries, which are essentially absent in gliomas, which always have invasive growth [1].

In other words, it is extremely important to map the cortex and the white matter pathways responsible for sensorimotor, visual-spatial and linguistic functions. To confirm this new concept, a recent meta-analysis, which presented the results of surgical treatment of 8091 patients with brain gliomas, demonstrated that the use of intraoperative mapping made it possible to achieve statistically significant reduction in permanent neurological deficit, despite the increase in the frequency of resections in functionally important areas; in addition, the degree of resection has been increased [1].

Gliomas are common tumors of the CNS and most often (in 40% of cases) they are localized in the frontal lobes [2], where various functionally important zones are located as well, including those responsible for motor functions and speech.

Surgical treatment of gliomas in functionally significant areas of the brain is a very challenging task. In addition to detailed knowledge of the functional anatomy of the cerebral cortex by a neurosurgeon, he or she also need to take into account the anatomy of the pathways during the surgery. Historically, the neurosurgeons and neurophysiologists focused on the functional anatomy of the cerebral cortex and only to a lesser extent on the pathways, of which the pyramidal tract is the most often mapped during surgeries [3]. The long association fibers involved in ensuring speech function have been less studied in the surgery of intracerebral tumors.

Despite significant progress in understanding the structure of the long association pathways of the brain, which has been achieved over the past 30 years due to the invention of diffusion-tensor imagine (DTI) and the improvement of mathematical models for receiving and processing the signal, such as HARDI, HARDI Q-ball, HARDI-CSD, their accuracy remains insufficient to definitely resolve the issue of anatomical structure of the associative fibers [4, 5]. Nevertheless, modern capabilities of the HARDI-tractography make it possible to identify the terminals of the tracts, the intersection of various bundles, and the course of fibers in the zone of tumor infiltration and edema [6]. Based on these data, it is possible to develop a preoperative representation of the topographic-anatomical relationship between the tract and the tumor (variants of displacement, infiltration, destruction, intact bundle).

Studied of the tract fibers course using Klingler dissection on autopsied brain specimen is important for understanding the microsurgical anatomy of the tracts [7].

Currently, awake craniotomy is the standard for intraoperative mapping of functional speech zones [8, 9].

The literature describes numerous cases of transient postoperative aphasia during surgical resection of gliomas with awakening, if gliomas are located in the dominant speech hemisphere. According to H. Duffau et al. [8], nearly 80% of patients experience speech disorders immediately after awake surgeries on the dominant hemisphere. After 3 months, up to 95% of patients no longer have neurological deficits. According to the results of other authors [10], permanent speech disorders can persist in almost 10% of patients. Intraoperative mapping is used to reduce the likelihood of postoperative speech disorders not only for cortical speech zones, but also for long association fibers.

The study purpose was to compare the results of intraoperative mapping and the postoperative speech function in patients with gliomas of the premotor area of the speech-dominant frontal lobe, which involved the superior longitudinal, arcuate, and frontal aslant tracts, who were operated on using awake craniotomy.

1. The SLF/AF and FAT topography

The topography and segmentation of long association fibers (SLF, AF, FAT) and their functional significance [11–13] are presented in Table 1 (they were described in more detail in our previous papers [6, 14]).

The superior longitudinal fasciculus (SLF) is a complex tract consisting of three segments: SLF I, SLF II, SLF III. The data on the structure, functions and symptoms detected in case of damage to these tracts are presented in Table 1. This segmentation may seem convention-based, especially for SLF I, which is anatomically separated from the rest of the segments. This approach is explained by the data of autoradiography on primates and the common function of all segments: establishing the connection between the frontal, parietal lobes and temporal lobes.

Currently the arcuate fasciculus (AF) is divided into two segments: dorsal and ventral. The dorsal segment connects the middle and inferior temporal gyri with the inferior frontal one; damage to it leads to disruption of the lexical and semantic aspects of speech and transcortical motor aphasia. The ventral segment connects the superior and middle temporal gyri with the lower frontal one (pars triangularis). The damage to this segment causes phonemic (literal) paraphasias, as described by Karl Wernicke in the XIX century. Some authors [11] describe the AF as part of the SLF.

The SLF and AF fibers run in parallel, however the AF, unlike the SLF, does not switch in the parietal lobe. The anatomy, segmentation and symptoms of damage to the main long association fibers have been described in more detail earlier in both Russian and foreign studies [11—13, 15, 16].

The frontal aslant tract (FAT) has been discovered later than all other pathways, using MRI tractography. It runs aslant from the inferior frontal gyrus (pars opercularis) to the medial surface of the additional motor cortex. Normally, excessive motor speech activity is suppressed through the frontal aslant tract and articulation is initiated. Partial damage or intraoperative stimulation of this tract causes stuttering or speech arrest, and full damage leads to the development of the specific Foix-Chavany-Marie syndrome (paresis of the face, larynx and jaw). It is assumed that FAT plays an important role in the initiation of spontaneous speech, linking Broca's area and the SFG [17].

The exact structure of the bundles described above remains unclear, their segmentation is not fully understood and mainly based on the difference in neurological deficits observed during intraoperative mapping of these fibers, which in itself is a rather controversial decision, since association pathways exhibit the greatest variability. Their structure and segmentation are described in detail in our previous work [16].

Topographically, the SLF II, SLF III and both AF segments run in parallel in the frontal lobe, closely adjoining each other, and the FAT intersects them at a right angle, as a result of which most clinical cases involve all the above-described bundles, and the intraoperative differential diagnosis of the pathway lesion is difficult. It is made even more challenging in the presence of preoperative or intraoperative aphasia, which blur the pattern of speech monitoring and makes it difficult to monitor the patient's neurological status. The layout of the main association tracts is shown in Fig. 1.

2. Materials and Methods

2.1. Patients' series

The inclusion criteria for this study were over 18 years of age, supratentorial glioma in the dominant hemisphere (speech), awake surgery, and the mandatory identification of the functionally significant pathways during the surgery (SLF, AF, FAT). In this article, we described a series of 12 clinical observations (6 men and 6 women, mean age of 45 years (29—67)) of patients with gliomas located in the left frontal lobe near the speech zones, of whom 6 patients had diffuse astrocytomas (Grade II), 1, anaplastic astrocytoma (Grade III), 1, glioblastoma (Grade IV), 1, oligodendroglioma (Grade II) and 3 had anaplastic oligodendroglioma (Grade III). In 6 patients clinical presentation included structural epilepsy with focal seizures, in 2 it was represented by speech disorders before the surgery, and in 4 by general cerebral symptoms (headache).

2.2. Pre- and postoperative contrast-enhanced MRI and MR-tractography

All 12 patients underwent MRI with contrast enhancement and re-construction of long association fibers according to the HARDI (High Angular Resolution Diffusion Weighted Imaging) method with establishing the topographic-anatomical relationship between the fibers and the tumor before the surgery and within the first 72 hours after its completion.

2.3. Neurophysiological control and awake craniotomy technique

All patients underwent surgical interventions with awakening according to the asleep-awake-asleep protocol using cortical and subcortical stimulation in order to localize functionally significant structures and clarify the possible amount of resection. The average current for direct electrical stimulation of the cortex and association tracts was 3 mA (1.9–6.5 mA). Direct electrostimulation was performed using a bipolar electrode. In all cases, electrocorticography was used during the surgery to control the epileptic activity of the cerebral cortex. Intraoperative ultrasound and fluorescent navigation with 5-aminolevulinic acid (5-ALA, Alasens) were used for the intraoperative identification of tumor boundaries in 6 cases.

2.4. Pre-, intra- and postoperative speech assessment

The state of speech functions was assessed before, during and after surgery by a neuropsychologist.

A comprehensive neuropsychological study was carried out according to the method of A.R. Luria before the surgery and before the discharge [18]. This method allows for a detailed qualitative analysis of the detected disorders, and it also allows to establish the topical affiliation of the identified symptoms. Different types of praxis, qualitative features of speech functions (including writing and counting), spatial functions, auditory and visual gnosis, as well as cognition were examined. The particular emphasis was placed on the study of speech function. Spontaneous speech, naming, understanding, repetition, and dictation were evaluated. Vocabulary and inertia of speech functions were assessed using a fluency of speech test with naming of words with a given character for 1 minute (“red” or “green” objects, “nouns starting with letter C or S”). In addition, a computerized naming test, which was used for intraoperative testing, was used for all patients before and after the surgery.

The leading hand was determined by the questionnaire of M. Annett [19], according to which 11 patients were right-handed, and 1 was re-trained at an early age left-hander, who completely switched to the right hand. The dominance of the hemisphere in terms of speech was determined using dichotic listening with the determination of the corresponding coefficient.

A computerized naming test [20] was used for intraoperative speech monitoring with naming of nouns or verbs using presentation of simple black-and-white pictures (a total of 30 pictures representing actions or objects); automated series were also evaluated (counting from 1 to 10, listing of months, days of the week). During the entire awake period, the patient was engaged in free dialogue in the absence of electrostimulation during the resection of the tumor.

3. Results

3.1. Preoperative neuropsychological study of patients

Prior to the surgery, 10 out of 12 patients had normal speech. One patient (No 3) had mild efferent motor aphasia (according to A.R. Luria) or Broca's aphasia, which manifested as individual perseverations in spontaneous speech and writing. Another patient (No 11) had normal speech, but had isolated perseverations of letters and syllables when writing. All patients had varying degrees of impairment of aural-speech memory and dynamic praxis.

3.2. Intraoperative mapping of functional speech zones of the cortex

The cortical area of motor speech was identified intraoperatively in 4 out of 12 patients. Its detection during the electrical stimulation was accompanied by speech arrest or perseverations of previous words, which was characteristic of efferent motor aphasia according to A.R. Luria (Broca's aphasia). In 2 cases, the cortical motor speech zone was located in the posterior regions of the inferior frontal gyrus, which coincided with the generally accepted anatomical borders of the Broca's are; in 2 other cases, it was localized in the middle and posterior parts of the middle frontal gyrus, respectively. In the remaining 8 observations, no cortical speech zones were detected during the surgery.

3.3. Intraoperative fluorescence diagnostics

Visible fluorescence was observed in 3 cases out of 6 patients operated on with the use of 5-ALA: One observation had a bright character (a patient with glioblastoma) and in 2 cases it was moderate (patients with Grade III gliomas). In 1 case, the fluorescence of the tumor was observed on the cerebral cortex. The fluorescence was absent in 3 patients with diffuse astrocytomas. Fluorescence was not used in 6 remaining cases.

3.4. Intraoperative mapping of the long association fibers

During the intraoperative mapping of the speech zones, the detected speech disorders were similar to those in efferent motor aphasia (Broca's), e.g., perseveration, speech arrest, as well as acoustic-mnestic aphasia (nominative, forgetting words). Less commonly, speech disorders occurred within the framework of the subcortical injury, e.g. dysarthria, slowdown. Speech disorders occurred for the first time outside of stimulation during free dialogue in 6 of 12 observations, as the tumor was being removed (paraphasia of a different nature, perseveration, pronouncing of words by syllables, dysarthria and slowdown, forgetting words). Of these 6 observations, direct electrostimulation was performed only in 2 cases, in which the speech disorders occurred again. In the remaining 6 observations, speech disorders occurred for the first time directly during the direct electrical stimulation. Therefore, speech disorders during the direct electrical stimulation were observed only in 8 out of the 12 observations: speech arrest, verbal and literal paraphasias when naming pictures, perseveration of previous words, forgetting words, slowing down and impaired speech articulation (Table 2). In 11 cases, the complex of the superior longitudinal and arcuate fascicula (SLF/AF) was localized in the bed of the tumor being removed; in one patient the frontal aslant tract (FAT) was located in the area. In our study we considered the complex of the SLF and the AF as a whole, without distinguishing these tracts from one another and segmentation of the SLF into individual fascicles. The detailed data on speech disorders before the surgery, during the surgery and after it at different time points are presented in Table 2. Without presenting pictures, we didn’t catch the naming disorders (temporal component); after the surgery it was the one that could have become predominant.

As can be seen from the Table 2, after the surgery certain speech disorders were identified in 11 of 12 patients. In one of them, speech disorders appeared 1 day after the surgery, which can be associated with an increase in postoperative cerebral edema. One patient developed moderate right-sided hemiparesis, mainly in the arm, after the surgery.

Table 2 also demonstrates that aphasia, detected after the resection of the frontal lobe tumor of the left hemisphere, had a complex character in 10 patients in the studied series. Only two patients did not have aphasia after the surgery. One of them (No 1) had subcortical speech disorders (slowing down, non-gross dysarthria) and writing (“disrupted” writing, non-flowing, micrography). His tumor was small, located mainly in the dorsomedial prefrontal area, during testing there were speech disorders similar to postoperative ones. In another patient (No 7), speech remained normal during the observation period. Her tumor was small, located near the Broca's area (according to functional MRI). During the intraoperative testing, perseveration and paraphasia were detected when naming actions, which made it possible to identify the Broca's area.

The analysis of complex speech syndrome in 10 patients is of particular interest. Eight of them had disorders typical for injuries to the temporal lobe, in addition to the perseverating syndrome of varying severity (efferent motor aphasia according to Luria, Broca's aphasia) typical for injuries to the left frontal lobe. These were primarily naming disorders, very similar to those in acoustic-mnestic aphasia (according to Luria). They were pronounced, accompanied by a description of the functional purpose of the presented object, in nearly all 8 patients. Verbal paraphasia, which the patients used instead of the real names of the objects, were also very characteristic; the paraphasias were often very far in meaning from the real name. In one case, such paraphasias were noted already during the intraoperative mapping (presentation of pictures with actions). An alienation of the meaning of words was observed In 4 out of these 7 patients, sometimes even in pronounced form (when asked to show parts of the face, objects in the room, pictures in an album, etc.). Such features of speech disorders made one think about the onset of conductive aphasia due to the severance of connections between the speech zones of the temporal and frontal lobes of the left hemisphere caused by damage to the long association fibers (SLF/AF complex).

After the surgery, patient No 2 had clear transcortical motor aphasia (dynamic). During the testing of the SLF/AF tracts, speech arrest and perseveration were noted. However, the postoperative picture better corresponded to the FAT involvement, which can be attributed to close spatial arrangement of the SLF/AF and FAT tracts in the area of the medial frontal gyrus. Another patient (No 5) had complicated motor aphasia (efferent motor - Broca’s aphasia) and afferent motor aphasia after the surgery, accompanied by impaired oral praxis and blurred speech. This component is characteristic of a parietal injury. The patient had a tumor predominantly of the median frontal gyrus, to a lesser extent of the lower frontal gyrus. Perseverations and speech blurring were also detected during the intraoperative testing. We could associate the onset of the parietal component in aphasia with the involvement of the parietal component of the superior longitudinal fasciculus.

The average Karnofsky index before surgery was 90 points, by the time of discharge, 70 points.

3.5. Magnetic Resonance Imaging

MRI conducted in all patients within the first 48—72 hours after the surgery showed that total resection was performed in 7 (more than 90% of the tumor) cases, subtotal, in 2, partial, in 2, and open biopsy was performed in 1 case. According to the postoperative MR tractography, the bed of the removed tumor was immediate adjacent to the SLF/AF complex in 7 cases, was located near the SLF/AF complex in 3 cases, and was immediately adjacent to FAT in 2 cases.

3.6. Clinical examples

Clinical example 1

Patient A, female, 29-year-old (observation No 7) The disease manifested in the form of attacks of speech disorders with subsequent short-term loss of consciousness. MRI of the brain revealed a tumor in the left frontal lobe (posterior parts of the inferior frontal gyrus - near Broca's area) (see Fig. 2). Examination by a neuropsychologist before the surgery: speech functions normal; auditory and visual memory at the lower limit of the norm. The removal of the left frontal lobe tumor was performed with electrophysiological monitoring and awakening. The brain cortex was mapped intraoperatively: the Broca's area was detected upward from the tumor (speech arrest was detected). Immediately towards the end of the tumor resection, during the electrical stimulation, in the region of the lower lateral surface of the surgical bed at a depth of about 3.5 cm from the cortex, speech disorders appeared in the form of incorrect naming of actions, and isolated perseverations; there were no visible tumor residues in this area. Topographically, this zone in the bed of the removed tumor corresponded to the course of the SLF/AF complex (which also corresponded to the data of both pre- and postoperative MR tractography). After the surgery, MRI with contrast enhancement showed total resection of the tumor; speech disorders were not detected during the examination by a neuropsychologist. Histological diagnosis: anaplastic oligodendroglioma.

Clinical example 2

Patient M, female, 60-year-old (observation No 4). Was admitted to the hospital with general cerebral symptoms. MRI before the surgery revealed glioma of the posterior portions of the inferior frontal gyrus on the left. In a preoperative neuropsychological study speech and writing were normal; the only disorders were clear defects of aural-speech memory with impaired selectivity of traces and disorders of dynamic praxis. The syndrome corresponded to the lesions of the posterior parts of the left frontal lobe.

The cortical motor zone of the arm was identified during the surgery (Fig. 3).

During subcortical electrostimulation at a depth of about 4 cm from the cortex, specific errors were identified in the naming of the actions shown in the pictures. Specificity consisted in the fact that the patient used verbs that are very far in semantics from the correct ones to refer to the actions shown in the pictures. For example, when a patient was shown a picture where a girl strokes a cat, she said: “Scratching the ground”, when she was shown a picture where the girl was brushing her teeth, she said: "Strokes the sand."

On Day 2 after the surgery, the efferent motor aphasia with verbal perseverations, naming disorders which were SLF/AF complex-specific and verbal paraphasias of similar nature were observed. Speech disorders had clear signs of damage to both the frontal and temporal lobes, which could be associated with the damage to the SLF/AF complex. Control MRI with contrast enhancement after the surgery, showed subtotal removal of the tumor.

Final histological diagnosis: diffuse astrocytoma Grade II.

Clinical example 3

Patient K, female, 48-year-old (observation No 9) The diagnosis of a tumor of the left frontal lobe was established 9 years ago after a single convulsive attack, which did not reoccur. All this time the patient abstained from the surgery and was under dynamic observation. At the next brain MRI, a significant increase in the size of the tumor was identified. The tumor had the following localization: the prefrontal regions and the poles of the left frontal lobe, the middle and anterior sections of the middle frontal and partially superior frontal gyri. Examination by a neuropsychologist before the surgery: speech is normal, other cognitive functions are preserved. The removal of the left frontal lobe tumor was performed with electrophysiological monitoring and awakening. During the surgery, spontaneous speech disorders in the form of isolated literal paraphasias appeared towards the end of the tumor resection in the posterior surface of the operating bed at a depth of about 3 cm from the cortex. Electrical stimulation was performed in the immediate area, during which speech disorders in the form of multiple perseverations and paraphasias were observed, and therefore a fragment of the tumor was left in this area. The identified zone could topographically correspond to the course of the frontal aslant tract (FAT), connecting the cortex of the medial surface of the frontal lobe and Broca's area. This is supported by the data of both pre- and postoperative MR tractography. However, on the 1^st day, the patient had more complex syndrome of speech disorders, consisting of efferent motor and acoustic-mnestic (nominative) aphasia, which was most pronounced at an arbitrary level (during tests). The patient did not name either objects or actions properly and she constantly perseverated. Characteristically, during the perseverations, she used the same verbs as during stimulation during the surgery. The temporal component of aphasia during the surgery in the region of the middle and superior frontal gyri of the cortical-subcortical level of the left frontal lobe indicates dissociation of the frontal and temporal lobes due to damage to the SLF/AF complex. This may be attributed to the large size of the tumor and the development of postoperative edema in the area of the not only the FAT, but also the SLF/AF. The results of the examination and identified disorders are presented in Fig. 4.

In this clinical example, the FAT mapping is interesting. Intraoperatively, the patient developed perseverations and paraphasias, these speech disorders could have been caused by electrical stimulation of the FAT, which links the extra motor cortex and Broca's area. The path of the fibers of the tract identified during the surgery corresponded anatomically to the FAT's course along the posterior contour of the tumor, obtained by MR tractography. However, in the postoperative period, due to the large size of the tumor and the development of perifocal edema, the onset of clinic presentation associated with injury to the SLF/AF complex was observed, which was dominated by temporal aphasia.

Discussion

The modern understanding of the neuroanatomical basis of linguistic functions has been established through models involving many areas of the cerebral cortex, functioning as part of a large network, including those consisting of the superior frontal gyrus, inferior parietal lobe, middle temporal gyrus, inferior temporal lobe region and other white matter tracts such as the SLFs, IFOFs, and fibers in the deep parts of the frontal lobe [21].

It should be noted that in general the function of the white matter pathways is not as well studied as that of the cortical structures of the brain. One of the reasons is the limitations of such research since it is usually difficult to find a patient with selective lesions of a particular tract. In addition, although the cortex can be mapped using surface electrodes, for example, as a method of preoperative study in patients with epilepsy or brain tumors, subcortical fibers are not as easy to evaluate this way [22].

The stimulation zone is considered functionally significant when speech disorders are noted three times in a row after subsequent stimuli, and speech function is recovered after the end of the stimulation. The type of speech disorders is verified by a neuropsychologist, and the nature and severity of speech disorders are assessed based on the tests used. The next stage is the resection of the tumor, taking into account the mapping data; periodic subcortical stimulation is performed to search for functionally significant white matter pathways.

Therefore, the use of intraoperative mapping and electrophysiological monitoring allows surgeons to remove the maximum possible volume of the pathological focus while maintaining neurological functions and minimizing the rate of postoperative complications [23—25].

The frontal lobe is the largest lobe of the human brain, its volume is up to 40% of the whole brain, and in terms of gliomas, it ranks first ahead of other brain lobes [10]. The main association fibers of the frontal lobe include the superior longitudinal fasciculus (SLF), which is divided into three segments (SLF I, II, III), the frontal aslant tract (FAT) and the lower frontal-occipital fasciculus (IFOF), which we did not review in our work. The variable anatomy of these tracts has been shown in the earlier works [14, 16].

In our series of 12 patients, all tumors were located in the left frontal lobe. The group included 7 patients with low-grade gliomas (Grade II) and 5 with high-grade gliomas (Grade III — IV). In 6 cases, speech disorders were observed without direct electrical stimulation during the resection of the tumor in the subcortical parts. In 8 patients, speech disorders developed during the electrical stimulation. The nature of speech disorders was diverse. In one patient, speech disorder was delayed (on Day 2) and reversible. In our opinion, this may be attributed to transient perifocal edema caused by the proximity of the SLF/AF fibers. The symptoms fully regressed after anti-edema therapy.

In this study, we observed reproducible speech symptoms in 8 patients during the electrical stimulation of the white matter of the frontal lobe. Our results showed that 11 patients had speech disorders without motor disorders of the extremities or tongue (except for 1 observation with the development of right-sided hemiparesis in the postoperative period). It means that aphasic disorders were not associated with damage to the motor component (the cortex and pyramidal tract).

After the surgery, an increase in speech disorders was noted in all patients (except 1 patient, observation No 7). It is noteworthy that in most cases aphasia was complex and combined focal symptoms of injury to both the frontal and temporal lobes, and in one patient (observation No 5) to the frontal and parietal lobes. It indicates partial damage to the long association fibers, leading to symptoms of dissociation of these brain lobes. The speech disorders described by us are fully aligned with the descriptions of conductive aphasia presented in the literature [26—28].

As a rule, if an intraoperative damage is not gross and incomplete, speech disorders are transient in nature and regress in a period from several days to several weeks. In these situations, the work of a speech therapist in the postoperative period is of particular importance.

Thus, 11 out of 12 patients had various variants of speech disorders in the postoperative period, despite the fact that the studied tracts were identified during the awake surgery and this information served as the basis for stopping the resection of the tumor. Our results correspond to the data presented in international literature, according to which patients have speech disorders after awake surgery with an incidence of up to 95%. (Such surgeries are, in principle, are performed when the speech cortical zones or vocal tracts are close). When the resection is methodically performed with mapping of the tract, speech disorders regress within 3 months after the surgery and remain permanent in only 5% of patients.

In one of the observations, it was impossible to reconstruct a part of the SLF/AF fibers in the tumor itself with the help of MR-tractography. When measuring the distance from the surface of the cortex to the fibers of the tract in the pre-operative images, we see a discrepancy with the distance from the tract to the cortex measured during the surgery. The tract was detected at a closer distance to the cortex, which indicates the impossibility of reconstructing fibers of the long association tract in the tumor itself. Any preoperative tractography is provisional (it is a mathematical model), because it has significant limitations in the reconstruction of fibers passing inside the tumor.

A number of researchers use different current strengths (from 2 to 10 mA) and different pulse amplitudes during the intraoperative stimulation. K. Seidel et al. [29] suggest to stop tumor resection in case of response to stimulation with a current strength of 2 mA. H. Duffau et al. [30] also use a current of 2 mA and considers the method of direct electrostimulation safe, accurate and reliable mean to identify pathways. It is well known that the strength of a current is interdependent with the distance over which it spreads; with an approximate ratio of 1 mm to 1 mA (the “golden rule” of neurophysiology) [31]. The parameters and the strength of the current are of great importance when stimulating the association fibers. In our series of observations, the current strength was 2–5 mA, which is in agreement with the data of foreign authors.

In patients with malignant gliomas, focal neurological symptoms are significantly more frequent before the surgery. In case of localization of long association fibers or the functional cortical speech zone near the tumor, it usually manifests as aphasia. In case of gross aphasia before the surgery, the awake craniotomy is contraindicated. At the same time, in case of the infiltrative nature of tumor growth, long association fibers can be involved in its structure, and their function can be preserved, especially in benign gliomas [32]. In case of malignant gliomas there is a destruction zone with an already formed focal neurological deficit in the white matter. In low-grade gliomas, the main symptoms before the surgery are focal epileptic seizures, aphasic disorders are extremely rare, which increases the complexity of the surgery and requires mandatory mapping of not only cortical zones, but also long association fibers.

In our series, the SLF/AF complex was found in 9 out of 12 patients during the surgery. We consider it unreasonable to differentiate these tracts during the surgery due to their close topographic-anatomical location and parallel course of the fibers. The difference between these tracts is that the SLF fibers switch in the parietal zone of Geschwind's zone, while the AF fibers go directly from the frontal lobe to the temporal lobe. In addition, during the surgery, it is difficult to identify the SLF segments due to their close proximity to each other and the similarity of the neurological symptoms of a lesion. However, anatomical separation of the SLF and AF fibers and the possibility of distinguishing between individual SLF segments have been shown in a number of papers [33]. In our opinion, studying the course of the fibers on anatomical preparations is necessary for a neurosurgeon-neuro-oncologist, however, distinguishing individual closely located tracts, and especially their segments, from each other is hardly possible during a real surgery. Therefore, we can agree with the opinion of a number of colleagues that intraoperatively it is reasonable to identify the SLF and AF fibers as a single complex. This is due not only to the proximity and contours of their course, but also to the similar symptoms in case of damage [32].

Therefore, clinical significance of such detailed intraoperative differential diagnosis of lesions to specific segments of association tracts remains controversial, since damage to any of the above described fibers is an indication for stopping tumor resection regardless of the specifically affected segment. However, understanding of the topography of the above fibers and their functions can provide valuable intraoperative information about the degree of involvement of each particular segment and influence the decision to continue with the tumor resection.

Conclusions

1. In case of intracerebral tumors of the frontal lobe of the dominant speech hemisphere, it is advisable to perform MR tractography with the reconstruction of long association fibers (SLF/AF complex and FAT) to assess their relationship with the tumor and their electrophysiological identification in conditions of awake craniotomy.

2. In surgery of tumors of the frontal lobe of the dominant hemisphere in awake conditions, after identifying functionally significant cortical structures at all stages of tumor resection, it is advisable to maintain continuous speech contact with the patient, complementing it with electrical stimulation, since the likelihood of speech disorders is high. The segmentation of the SLF fibers and distinguishing the SLF from the AF is difficult in real operating conditions.

3. In the majority of observations, patients experience an increase in speech disorders in the early postoperative period (11 out of 12 patients). These speech disorders eventually regress in the vast majority of patients.

4. In case of damage to the long association fibers of the dominant hemisphere during the surgery, a syndrome of separation of the frontal, temporal, and less often the parietal lobes may develop followed by development of complex types of aphasias, which may increase in the early postoperative period.

Acknowledgments:

The authors are grateful for the help in preparing the manuscript to the students V.А. Tyurin and S.N. Belyaev ("N.I. Pirogov Russian National Medical Research University" of the Ministry of Health of Russia, Moscow) and the intern A.V. Kondrashov ("I.M. Sechenov First Moscow State Medical University" of the Ministry of Health of Russia, Moscow).

The article was supported by RFBR grants No 16-04-01419 “Dynamic studies of the microstructure and plasticity of corticospinal tracts on a model of traumatic and neoplastic lesions of the human brain using diffusion-kurtosis magnetic resonance imaging, diffusion with high angular resolution and non-contrast magnetic resonance perfusion”, grant RFBR No 18-29-01-032

"Study of the individual variability of the functional integration of brain areas in glial tumors for preoperative non-invasive mapping" and grant RFBR No 17-00-00158 KOMFI "Study of clinical and molecular genetic patterns of gliomas of the human brain with a long period of overall survival".

Limitations:

The follow-up interval was insufficient in patients operated on in 2017 (less than 12 months). The study did not include any left-handed persons. We did not consider patients with bilateral speech representation in tumors of the subdominant hemisphere and cross aphasia.

Authors declare no conflict of interest.

Commentary

One of the important problems in neurooncology is surgical tactics in the treatment of gliomas of functionally significant areas of the brain. Neurosurgery focuses on the mapping of the cerebral cortex. Methods for mapping pathways, especially the long association fibers of the brain, have been studied to a lesser extent and are not implemented into practice to the same degree. These conductors support the most important functions of the human brain, which was confirmed by a number of studies (H. Duffau et al., M. Berger et al., S. Sarubbo et al., 2015).

The article summarizes the experience of mapping the complex of the superior longitudinal and arcuate tracts in 12 patients with gliomas of the left frontal lobe. The authors continue to summarize the clinical observations after 2 previous publications on the anatomy of the long association fibers (S.А. Goryainov, V.Yu. Zhukov, A.A. Potapov, 2014 and 2017).

A review of the literature and clinical examples are provided, with particular attention on the details of mapping of long association fibers, features of neuropsychological tests and postoperative speech disorders. The work is illustrated with clinical examples of patients with gliomas of the left frontal lobe with intraoperative identification of the complex of the superior longitudinal and arcuate tracts.

The authors dataset is rather small and they should continue this important study; our recommendation would be to include patients with gliomas of the brain that involve other long association tracts.

This is the first work in the Russian literature on this particular topic and it will be useful for practicing neurosurgeons who treat patients with intracerebral brain tumors, which will improve the safety and effectiveness of surgical interventions.

V.L. Puchkov (Moscow, Russia)