Radiology
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

DOI: 10.1148/radiol.2433060068
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stippich, C.
Right arrow Articles by Sartor, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stippich, C.
Right arrow Articles by Sartor, K.
(Radiology 2007;243:828-836.)
© RSNA, 2007

Neuroradiology

Localizing and Lateralizing Language in Patients with Brain Tumors: Feasibility of Routine Preoperative Functional MR Imaging in 81 Consecutive Patients1

Christoph Stippich, MD, Nora Rapps, MD, Jens Dreyhaupt, PhD, Anita Durst, Bodo Kress, MD, Ernst Nennig, PhD, Volker M. Tronnier, MD, and Klaus Sartor, MD

1 From the Division of Neuroradiology, Department of Neurology (C.S., N.R., A.D., B.K., E.N., K.S.), Department of Neurosurgery (V.M.T.), and Institute of Medical Biometry and Applied Informatics (J.D.), University of Heidelberg Medical School, Im Neuenheimer Feld 400, 69120 Heidelberg, Germany. Received January 13, 2006; revision requested March 10; revision received April 27; accepted May 5; final version accepted October 12. Address correspondence to C.S. (e-mail: christoph_stippich{at}med.uni-heidelberg.de).


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Purpose: To prospectively assess the feasibility of standardized presurgical functional magnetic resonance (MR) imaging for localizing the Broca and Wernicke areas and for lateralizing language function.

Materials and Methods: The study was approved by the responsible ethics commission, and patients gave written informed consent. Eighty-one patients (36 female and 45 male patients; age range, 7–75 years) with different brain tumors underwent blood oxygen level–dependent functional MR imaging at 1.5 T with two paradigms: sentence generation (SG) and word generation (WG). Functional MR imaging measurements, data processing, and evaluation were fully standardized by using dedicated software. Four regions of interest were evaluated in each patient: the Broca and Wernicke areas and their anatomic homologues in the right hemisphere. Statistics were calculated.

Results: The SG and WG paradigms were successfully completed by all (100%) and 70 (86%) patients, respectively. Success rates in localizing and lateralizing language were 96% for the Broca and Wernicke areas with the SG paradigm, 81% for the Broca area and 80% for the Wernicke area with the WG paradigm, and 98% for both areas when the SG and WG paradigms were used in combination. Functional localizations were consistent for SG and WG paradigms in the inferior frontal gyrus (Broca area) and the superior temporal, supramarginal, and angular gyri (Wernicke area). Surgery was not performed in seven patients (9%) and was modified in two patients (2%) because of functional MR imaging findings.

Conclusion: Functional MR imaging proved to be feasible during routine diagnostic neuroimaging for localizing and lateralizing language function preoperatively.

© RSNA, 2007


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Neurosurgical interventions in patients with brain tumors located in or close to language areas can cause intraoperative neuronal damage, leading to language deficits and thus to a further reduction in quality of life. Therefore, in patients with frontal and temporal lobe tumors, neurosurgery should be carefully considered.

Imaging techniques—in particular, magnetic resonance (MR) imaging—provide detailed morphologic information on brain disease. However, morphologic imaging is of limited diagnostic value before surgery for tumors in close spatial relationship to the functional cortex, because such imaging is unable to measure and visualize brain function. The "hand knob" of the precentral gyrus, responsible for motor hand function, is the only reliable morphologic landmark for identifying a functionally important brain area (1). Brain sites essential for language cannot be localized by using morphologic criteria alone (2).

Preoperative functional MR imaging can provide such diagnostic information: With functional MR imaging, it is possible to determine the language-dominant hemisphere (3,4) and the spatial relationships between brain tumors and language areas before surgery (5,6). This facilitates the selection of candidates and the strategy for neurosurgery, leading to function-preserving resections. Because of its noninvasiveness and preoperative availability, functional MR imaging offers major advantages over the traditional reference procedures that are based on intraoperative electrocorticography (2) or intraarterial administration of barbiturates (the Wada test) (7). There is broad evidence from combined functional MR imaging studies with electrocorticography (5,6,8) or Wada testing (3,4,9) that confirms the spatial accuracy and validity of preoperative functional MR imaging for localizing and lateralizing language function. However, the results of language lateralization are not perfectly consistent across different studies owing to differences in experimental design and limitations intrinsic to the method.

Because no guidelines exist for clinical functional MR imaging examinations, including guidelines for analysis and medical interpretation, each investigator must establish his or her own standards. Important prerequisites for the clinical use of functional MR imaging are optimization, standardization, and evaluation in larger cohorts of patients.

The objective of this study, therefore, was to prospectively assess the feasibility of standardized presurgical functional MR imaging for localizing the Broca and Wernicke areas and for lateralizing language function.


   MATERIALS AND METHODS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Patients
After approval by the responsible ethics commission and individual written informed consent were obtained, a total of 81 consecutive patients were enrolled in the study from October 1999 to January 2004—36 female patients and 45 male patients ranging in age from 7 to 75 years (mean age, 42 years). Hand preference was determined by using a modified questionnaire according to Annett (10).

The 81 frontal and temporoparietal lobe tumors included 58 gliomas (30 World Health Organization [WHO] grade II, 12 WHO grade III, and 16 WHO grade IV tumors), seven metastases (two from adenocarcinoma, two from lung cancer, two from malignant melanoma, and one from renal clear cell carcinoma), and one malignant pleomorphic cell tumor. Seven patients had cavernous angiomas, three patients had cortical dysplasia, and one patient had a brain abscess. Four histologic diagnoses were not available.

Presurgical functional MR imaging was performed in 70 patients with left-sided tumors—60 right-handed, seven left-handed, and three ambidextrous persons. Eleven patients with right-sided tumors—seven left-handers and four right-handers—were also included in the study on the basis of clinical signs and symptoms (language deficits).

At the time of functional MR imaging, 32 patients had aphasic symptoms; none of the patients was globally aphasic. Forty-eight patients presented with other tumor-associated neurologic deficits such as epilepsy (n = 34), paresis (n = 8), impaired sensibility (n = 3), and disturbance in vision (n = 3). All clinical data were provided by one author (V.M.T.) and compiled by another (N.R.).

Standardized Functional and Morphologic MR Imaging
Functional and morphologic whole-head imaging was performed by using a clinical 1.5-T MR system (EDGE; Marconi, Cleveland, Ohio) and a conventional head coil. Preformed foam cushions were used for biparietotemporal head fixation that spared the temporomandibular joint.

Blood oxygen level–dependent (BOLD) contrast functional images were acquired by using a T2*-weighted single-shot gradient-echo echo planar imaging sequence (repetition time msec/echo time msec, 3000/80; field of view, 256 x 256 mm; matrix, 128 x 128 voxels; pulse angle, 90°; 22 transverse sections; section thickness, 5 mm; intersection gap, 1 mm). After functional MR imaging measurements were obtained, a contrast agent (gadopentetate dimeglumine [Magnevist; Schering, Berlin, Germany] or gadodiamide [Omniscan; Amersham Büchler, Braunschweig, Germany]) was administered at 0.2 mL per kilogram of body weight, and additional T1-weighted morphologic three-dimensional data were acquired in all patients (with their heads in an unchanged position) for the overlay of activation maps on morphologic images (radiofrequency-spoiled fast low-angle shot [FLASH] sequence; 30/4.4; with 135 transverse sections and a section thickness of 1.3 mm in patients undergoing neuronavigation and 120 sagittal sections and a section thickness of 1.5 mm in all other patients).

The functional MR imaging protocol was based on visually triggered nonvocalized sentence generation (SG) and word generation (WG) (11) (Fig 1). During the training that preceded all functional MR imaging examinations, both paradigms were adapted at the stimulation computer to each patient's individual linguistic abilities in an attempt to increase both the technical success rate and the localization-achieved success rate (LSR), along with high BOLD signal intensity and high correlation of the measured BOLD signals with the applied hemodynamic reference function (HRF). Therefore, the trigger frequency was varied systematically within the standardized asymmetric block design (four 36-second periods of stimulation alternating with five 18-second periods of rest) (12). Both paradigms were adjusted and then trained until there were no learning effects.


Figure 1
View larger version (17K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 1: Examples of visually presented triggers for SG (left) and WG (right) paradigms. All triggers are listed in Tables 1 (SG paradigm) and 2 (WG paradigm).

 
SG was triggered by 12 different comiclike pictures, each of which coded for a standardized simple sentence (Table 1). Depending on the patient's individual linguistic abilities, each picture (trigger) was presented to the patient for 1–6 seconds, which resulted in six to 36 triggers per stimulation period (12). The standard presentation time was 2 seconds. In WG, the patients were given 12 different generic terms as triggers that referred to various categories (eg, colors) (Table 2). The patients were asked to generate as many words (eg, red, yellow, green) as possible that they associated with each generic term. As a standard, each term was presented to the patients for 6 seconds (six triggers per stimulation period). In four patients in whom severe aphasic symptoms precluded proper task performance, a longer presentation time was needed (9 seconds, four triggers per stimulation period). All triggers (for both the SG and WG paradigm) were pseudorandomized to rule out direct stimulus repetitions and to ensure that identical paradigms were used in the functional MR imaging measurements. During rest, the patients were instructed to focus on a black crosshair on a white background. They were asked to discontinue the covert language production and to relax. An offset of 9 seconds was used before every functional MR imaging measurement to eliminate T1 effects. The total measuring time per paradigm was thus 4 minutes 3 seconds.


View this table:
[in this window]
[in a new window]

  		Table 1. Set of Visual Stimuli (Triggers) for SG Paradigm

 

View this table:
[in this window]
[in a new window]

  		Table 2. Set of Visual Stimuli (Triggers) for WG Paradigm

 
Special nonmagnetic mirror glasses that could be fitted with optically corrected lenses (Fig 2) were used for visual backprojection. The visual stimulation helped to prevent an overlap of auditory and language-associated brain activations in the region of the superior temporal gyrus that can occur with auditory stimuli.


Figure 2
View larger version (92K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 2: Nonmagnetic mirror glasses with slide-in module for optical correction lenses that were fitted before functional MR imaging in the ophthalmology outpatient clinic. The adjustable mirrors can be removed from the spectacle frame for head coils with mirror-holding devices.

 
In all patients, the training and individual adaptation of paradigms and all MR imaging measurements were performed by one author (C.S.).

Standardized Processing and Evaluation of Individual Morphologic and Functional MR Imaging Data
The processing of all morphologic and functional MR imaging data was performed by one of two authors (C.S., with 9 years of experience, or E.N., with 4 years of experience). All morphologic and functional MR images were evaluated by one author (C.S.).

Software (BrainVoyager; BrainInnovation, Maastricht, the Netherlands) was used for standardized processing and evaluation of all morphologic and functional MR imaging data. The software's functions included motion correction, spatial and temporal smoothing, and a voxelwise calculation of BOLD activations with linear cross correlations (see Statistical Analysis section for details). Data processing was automated, except for the overlay of functional on morphologic MR images, which was performed manually. The time needed for offline processing was approximately 70 minutes for each patient, including 20 minutes personal expenditure time (the time needed to perform all manual steps in data processing and evaluation).

On morphologic three-dimensional MR images, anatomic brain structures affected by the tumors were assessed according to criteria and landmarks established by Naidich et al (13).

Four regions of interest were evaluated for language activation: (a) the Broca area in the left inferior frontal gyrus (Brodmann areas BA44, 45); (b) the Wernicke area in the left superior temporal gyrus (BA22), the adjacent superior temporal sulcus, the middle temporal gyrus (BA21, 22), the supramarginal gyrus (BA40), and/or the angular gyrus (BA39); (c) the anatomically homologous areas to the Broca area in the right hemisphere; and (d) the anatomically homologous areas to the Wernicke area in the right hemisphere.

After the functional activation maps were overlaid on the corresponding morphologic three-dimensional data, the functional MR imaging data were evaluated on an individual basis by using a standardized approach (11,12): To achieve a precise determination of the anatomic correlates of the different functional activations by also eliminating very small clusters in the activation maps, the empirically proved cluster size of 36 mm3 was applied as standard. At first, a very high statistical threshold value for the correlation (r) between the measured BOLD signals and the HRF was selected so that no functional activation was displayed (empty map). This threshold was then continually reduced. As a result, the activation with the highest correlation with the HRF that exceeded the cluster size of 36 mm3 was displayed first. Through further reduction of the threshold, activations in other functional areas with a lower correlation between the measured BOLD signals and the HRF were displayed in a hierarchic order. This procedure was continued until activations were identified in all regions of interest (Fig 3).


Figure 3
View larger version (84K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 3: Standardized evaluation of individual preoperative functional MR imaging data through continuous reduction of statistical threshold (red arrowheads). Functional activation maps overlaid on transverse contrast material–enhanced T1-weighted MR images (three-dimensional FLASH sequence, 30/4.4) at level of Sylvian fissure show that dynamic thresholding results in a hierarchic order of functional activations and enables adaptation to different levels of activation between patients or functional MR imaging measurements. Default parameters were a cluster size of more than 36 mm3 and a lower-limit correlation of BOLD signals with the HRF of r > 0.4 and P < .05 (as shown by BOLD signal time courses). Image 0, an empty map (EM), shows the statistical threshold r value of 1.0. In image 1 (r = 0.82) Wernicke area (W) activation (black arrowhead) correlates best with the HRF. In image 2 (r = 0.68), Broca area (B) activation (black arrowhead) has second-best correlation with the HRF. In image 3 (r = 0.62), right hemisphere Broca area homologue (BR) activation (black arrowhead) has third-best correlation with the HRF, and the Broca area has increased in volume. Cluster sizes of Broca area and right hemisphere homologue were measured to calculate lateralization index for Broca area versus right hemisphere homologue (arrow). In image 4 (r = 0.60), right hemisphere Wernicke area homologue (WR) activation (black arrowhead) has lowest correlation with the HRF. Lateralization index for Wernicke area versus right hemisphere homologue was calculated (arrow). Standardized measurements of BOLD signals were performed in each region of interest (Broca area, Wernicke area, and right homologues of Broca and Wernicke areas).

 
For each region of interest separately, a threshold of r = 0.4 with P < .05 was established as a lower limit to ensure that BOLD signals were clearly distinguishable from background noise. If no BOLD activation was displayed in a region of interest within the lower limit, this was classified as "no activation." Likewise, BOLD signals with a relative change greater than 5% were not included in the evaluation because such high-level activation is likely to originate from draining veins rather than from capillaries. Where anatomic structures remained preserved, the correlates of functional activations were assessed precisely, verifying the correct identification of the Broca and Wernicke areas and their right hemisphere homologues.

Language dominance was determined on the basis of regional lateralization indexes that were calculated separately for the Broca area and right hemisphere Broca area homologue and the Wernicke area and right hemisphere Wernicke area homologue according to the following established equation (11): LI = (nVxL – nVxR)/(nVxL + nVxR), where LI is the lateralization index, nVxL is the number of activated voxels at a particular threshold in the left hemisphere, and nVxR is the number of activated voxels at a particular threshold in the right hemisphere.

Validity of Functional Language Localization and Lateralization
The validity of functional MR imaging language localization and lateralization with the SG and WG paradigms was assessed in eight patients in comparison to established reference procedures: electrocorticography (n = 2, V.M.T.), Wada testing (n = 3, C.S.), or both (n = 3).

Statistical Analysis
In the BrainVoyager software, the functional activations were calculated by correlating the measured data in each voxel with the time course of the applied HRF, which resulted in the correlation coefficient r indicating the strength of covariation. An r value of 1.0 indicates perfect agreement of the data time course and the model (HRF); an r value of 0, no covariation; and an r value of –1, perfect opposition. Covariation of measured data and HRF was considered for P < .05. All calculations of lateralization index and descriptive statistics (absolute and relative frequencies) were performed by two authors (N.R. and J.D.) with software (SAS, version 8.2; SAS Institute, Cary, NC). Technical success rates and LSRs were calculated for either task (SG and WG) and for the combination of both tasks (SG plus WG).


   RESULTS
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 
Functional Localization of Broca and Wernicke Areas
Presurgical functional MR imaging was successfully performed in all 81 patients with SG (technical success rate, 100%) and in 70 patients with WG (technical success rate, 86%). The SG paradigm enabled functional localization of the Broca and Wernicke areas in 78 patients (LSR, 96%). In two cases, BOLD activation was subthreshold (r < 0.4). In one case, motion artifacts precluded evaluation. The WG paradigm enabled localization of the Broca area in 66 patients (LSR, 81%) and the Wernicke area in 65 patients (LSR, 80%). In addition to the 11 patients who could not perform the cognitively more demanding WG during the training period, it was impossible to obtain robust BOLD signals in four (Broca area) and five (Wernicke area) other patients (r < 0.4).

The SG paradigm enabled localization of the right hemisphere Broca area homologue in 68 patients (LSR, 84%) and the right hemisphere Wernicke area homologue in 65 patients (LSR, 80%). The WG paradigm enabled localization of the right hemisphere Broca area homologue in 60 patients (LSR, 74%) and the right hemisphere Wernicke area homologue in 53 patients (LSR, 65%). In all cases, subthreshold BOLD activation (r < 0.4) was the reason for the lower LSR in the right than in the left hemisphere. When both paradigms were used in combination, the Broca and Wernicke areas could be localized in 79 patients (LSR, 98%), the right hemisphere Broca area homologue could be localized in 73 patients (LSR, 90%), and the right hemisphere Wernicke area homologue could be localized in 72 patients (LSR, 89%) (Tables 3, 4; Fig 4).


View this table:
[in this window]
[in a new window]

  		Table 3. Anatomic Structures Affected by Space-occupying Masses at Morphologic MR Imaging

 

View this table:
[in this window]
[in a new window]

  		Table 4. Anatomic Correlates of Functional Activation Sites

 

Figure 4
View larger version (49K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 4: Preoperative standardized functional MR imaging with SG paradigm in 43-year-old man with recurrent large left frontal astrocytoma (WHO II) in a critical spatial relationship to the inferior frontal gyrus (Broca area). Anatomy is not preserved. Functional activation maps overlaid on transverse (left) and sagittal (right) contrast-enhanced T1-weighted MR images (three-dimensional FLASH sequence, 30/4.4) show frontal BOLD activation, indicating the Broca area to be localized at the dorsal end of the inferior frontal gyrus, presumably in the pars opercularis (white arrowheads) in or at least directly adjacent to the T1-hypointense, nonenhancing tumor. Temporal lobe BOLD activation indicated the Wernicke area to be in its classic location of the dorsal superior temporal gyrus (black arrowhead). With respect to the functional MR imaging findings, a partial resection was performed. Note the stimulus-associated activation of the visual cortex, indicating good visual input.

 
Lateralization of Language Function
Regional lateralization indexes for the Broca and Wernicke areas versus their right hemisphere homologues were calculated whenever BOLD activation was detected in one hemisphere. With the SG paradigm, language dominance was determined for the Broca and Wernicke areas versus their right hemisphere homologues in 78 patients (LSR, 96%), which corresponded to the LSR for localizing the Broca and Wernicke areas. With the WG paradigm, language dominance was determined for the Broca area versus its right hemisphere homologue in 67 patients (LSR, 83%) and for the Wernicke area versus its right hemisphere homologue in 65 patients (LSR, 80%). Whereas the success rate of determining language dominance for the Wernicke area versus its right hemisphere homologue (LSR, 80%) corresponded to the success of localizing the Wernicke area, it was slightly higher for the Broca area versus its right hemisphere homologue (LSR, 83%) than for the localization of the Broca area (LSR, 81%). The reason was that one patient showed activation in the right hemisphere Broca area homologue, but no corresponding activation in the Broca area. With the SG and WG paradigms in combination, language lateralization was possible for the Broca and Wernicke areas versus their right hemisphere homologues in 79 patients (LSR, 98%).

As a result of preoperative functional MR imaging results, surgery was not performed in seven patients (9%), who underwent radiation therapy instead. Surgery was modified to partial resections in two other patients (2%) who had large malignant gliomas.

Validation of Preoperative Functional MR Imaging Protocol
Electrocorticographic results confirmed the correct preoperative functional localization of the Broca area in two patients and the Wernicke area in three patients. Results of Wada testing confirmed correct preoperative language lateralization in six patients (Figs 5, 6).


Figure 5
View larger version (139K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 5: Validation of preoperative functional MR imaging language localization with intraoperative electrocorticography. Functional activation maps (white clusters) overlaid on coronal (top left), sagittal (top right), and transverse (bottom left) contrast-enhanced T1-weighted MR images (three-dimensional FLASH sequence, 30/4.4) and intraoperative photograph (bottom right) in 52-year-old right-handed man with right parietal oligoastrocytoma (WHO II) and aphasic symptoms. Functional MR imaging and Wada testing indicated an atypical right language dominance. During craniotomy with the patient awake, direct cortical stimulation of the area of the BOLD activation in the superior temporal gyrus (right hemisphere Wernicke area homologue) interrupted language production and confirmed the preoperative functional MR imaging localization. Yellow lines indicate planned resection borders during neuronavigation.

 

Figure 6
View larger version (67K):
[in this window]
[in a new window]
[Download PPT slide]
  		Figure 6: Validation of preoperative functional MR imaging language lateralization by using selective Wada test in 23-year-old woman with left temporal cavernoma. Left: Functional activation map overlaid on sagittal contrast-enhanced T1-weighted MR image (three-dimensional FLASH sequence, 30/4.4). Middle and right: Lateral views from selective digital subtraction angiography. Sodium amytal was injected into the posterior temporal artery (middle) and into the artery of the angular gyrus (right). Left dominance, as determined with preoperative functional MR imaging with the SG (results shown) and WG paradigms, was confirmed. Language-associated BOLD activation in left superior temporal gyrus localizes Wernicke area directly adjacent to the tumor (black arrowhead). The inferior frontal gyrus shows an anatomic variant, with a fusion of its pars triangularis with the middle frontal gyrus, where BOLD activation indicates the Broca area to be located (white arrowhead).

 

   DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
ADVANCE IN KNOWLEDGE
 References
 
With optimized clinical imaging protocols and standardized data processing and evaluation routines (11), functional MR imaging proved to be feasible during the presurgical diagnostic routine MR work-up in patients with brain tumors for localizing and lateralizing language function: In our series of 81 consecutive patients, a technical success rate of 100% and a localization and lateralization success rate of 98% was achieved.

Because the imaging protocol requires a measuring time of only 8 minutes for two different language paradigms, functional MR imaging can be implemented easily into routine diagnostic neuroimaging. Important prerequisites for successful examinations were as follows: First, standardized imaging procedures based on nonvocalized language paradigms with stable head fixation that helped reduce motion artifacts (14); second, the adaptation of the stimulation to each patient's individual linguistic abilities during the training prior to functional MR imaging; third, visual stimulation with optical correction, which made examination of patients with impaired vision possible; and finally, the evaluation of data with a dynamic statistical threshold.

In presurgical functional MR imaging, the application of two different paradigms providing consistent functional localizations (11) proved to be useful. Most aphasic patients found it easier to perform SG than WG, because SG required only reproduction of standardized sentences. In two patients, however, functional MR imaging failed to depict language areas because the BOLD signals were subthreshold. The fact that diagnostic evaluation of one data set (with the SG paradigm) was impossible because of motion artifacts also emphasizes the importance of multiple measurements in clinical functional MR imaging. In this particular case, functional MR imaging for language localization could be performed successfully with the second paradigm (WG). The use of multiple paradigms has the further advantages of allowing consideration of several linguistically important components of human language in presurgical functional neuroimaging (15) and of improving the assessment of hemispheric dominance (16).

However, owing to limitations intrinsic to the method, functional MR imaging does not provide a quantitative measure of brain activation and hemispheric dominance. Functional MR imaging language lateralization depends on the paradigms applied (17), on the brain's workload associated with a particular language task (12), and on the statistical threshold used for data evaluation. Accordingly, in the absence of factors that influence language lateralization such as left-handedness (18), any "atypical" language dominance (equal, mixed, or right dominance) suggested by presurgical functional MR imaging cannot be attributed solely to tumor-associated "neuroplastic" changes (19). In such patients, additional validation with established reference procedures should be considered (Fig 5).

The validity of presurgical functional MR imaging for language localization has been confirmed by others (3–6,8,9,20–22) with electrocorticography or Wada testing and was therefore not the topic of this work. However, the agreement between functional MR imaging and reference procedures was not perfect in all of these previous studies and depended on differences in experimental setup, functional MR imaging protocols, data processing, and evaluation procedures. Considering the invasiveness of both reference procedures, the validity of our specific presurgical functional MR imaging procedure was assessed in only a limited number of patients, but there was broad agreement between the results obtained.

Real-time functional MR imaging (23,24) enables one to assess the success of functional MR imaging examinations during each measurement, such that erroneous measurements can be detected immediately and repeated. This may further increase the technical success of presurgical functional MR imaging. Until now, the functionality of the available real-time software is still limited compared with that of established offline-processing tools, justifying the additional time needed for semiautomated data processing.

Methodologic limitations of functional MR imaging that may lead to altered BOLD signals or localization errors should be taken into consideration in patients with malignant gliomas, highly vascularized metastases, or cerebral arteriovenous malformations (25–28). These alterations result from neovascularization, vascular compression, or altered vascular autoregulation. When functional images are superimposed on morphologic three-dimensional images or are used for neuronavigation, localization inaccuracies, which are usually less than 5 mm in size, can occur (29,30). Preoperatively acquired images do not correctly reflect intraoperative conditions. After the dura is opened, loss of cerebrospinal fluid and tissue removal cause brain shift (31,32).

It is important to note that resection borders cannot be determined reliably from functional MR images because the spatial extent of the BOLD clusters and the euclidean coordinates of the centers of gravity change with the statistical threshold. However, functional MR imaging is capable of precisely localizing the center of a functional area within the gyrus that needs to be protected during surgery. Multiple linguistically different language paradigms—which should be used in presurgical functional MR imaging to better account for the complexity of human language function—may activate slightly different locations within the same functional area (eg, Broca or Wernicke), indicating that all these different locations should be preserved.

This study had limitations regarding the evaluation of the effect of presurgical functional MR imaging on therapeutic decisions, neurosurgical procedures, and clinical outcome. To this end, controlled prospective trials based on standardized functional MR imaging need to be conducted in the future—trials that should include a detailed documentation of therapeutic decisions and of functionally guided neurosurgery, along with pre- and postoperative neuropsychologic assessment of language function.

In conclusion, presurgical functional MR imaging can be considered feasible for language localization and lateralization during routine clinical neuroimaging, if methodologic limitations and potential sources of error are accounted for and examinations are standardized. In that case, the method has great potential to assist function-preserving treatment in patients with brain tumors and to substantially reduce the number of invasive measures needed during the diagnostic work-up and during surgery. To further establish functional MR imaging as a diagnostic neuroimaging procedure, a consensus on stimulation paradigms, imaging protocols, data processing, and medical interpretation of functional MR imaging data is imperative.


   ADVANCE IN KNOWLEDGE
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
References
 


   ACKNOWLEDGMENTS
 
The authors thank Maria Blatow, MD, for proofreading and correcting English.


   FOOTNOTES
 

Abbreviations: BOLD = blood oxygen level dependent • FLASH = fast low-angle shot • HRF = hemodynamic reference function • LSR = localization-achieved success rate • SG = sentence generation • WG = word generation • WHO = World Health Organization

Authors stated no financial relationship to disclose.

Author contributions: Guarantor of integrity of entire study, C.S.; study concepts/study design or data acquisition or data analysis/interpretation, all authors; manuscript drafting or manuscript revision for important intellectual content, all authors; manuscript final version approval, all authors; literature research, C.S., N.R., A.D.; clinical studies, C.S., B.K., E.N., V.M.T.; statistical analysis, N.R., J.D., E.N.; and manuscript editing, C.S., N.R., A.D., B.K., V.M.T., K.S.


   References
TOP
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 ADVANCE IN KNOWLEDGE
 References
 

  1. Yousry TA, Schmid UD, Alkadhi H, et al. Localization of the motor hand area to a knob on the precentral gyrus: a new landmark. Brain 1997;120(pt 1):141–157.[Abstract/Free Full Text]
  2. Ojemann G, Ojemann J, Lettich E, Berger M. Cortical language localization in left, dominant hemisphere: an electrical stimulation mapping investigation in 117 patients. J Neurosurg 1989;71:316–326.[Medline]
  3. Binder JR, Swanson SJ, Hammeke TA, et al. Determination of language dominance using functional MRI: a comparison with the Wada test. Neurology 1996;46:978–984.[Abstract/Free Full Text]
  4. Desmond JE, Sum JM, Wagner AD, et al. Functional MRI measurement of language lateralization in Wada-tested patients. Brain 1995;118(pt 6):1411–1419.[Abstract/Free Full Text]
  5. FitzGerald DB, Cosgrove GR, Ronner S, et al. Location of language in the cortex: a comparison between functional MR imaging and electrocortical stimulation. AJNR Am J Neuroradiol 1997;18:1529–1539.[Abstract]
  6. Rutten GJ, van Rijen PC, van Veelen CW, Ramsey NF. Language area localization with three-dimensional functional magnetic resonance imaging matches intrasulcal electrostimulation in Broca's area. Ann Neurol 1999;46:405–408.[CrossRef][Medline]
  7. Wada J, Rasmussen T. Intracarotid injection of sodium amytal for the lateralization of cerebral speech dominance. J Neurosurg 1960;17:266–282.
  8. Hirsch J, Ruge MI, Kim KH, et al. An integrated functional magnetic resonance imaging procedure for preoperative mapping of cortical areas associated with tactile, motor, language, and visual functions. Neurosurgery 2000;47:711–721.[CrossRef][Medline]
  9. Yetkin FZ, Swanson S, Fischer M, et al. Functional MR of frontal lobe activation: comparison with Wada language results. AJNR Am J Neuroradiol 1998;19:1095–1098.[Abstract]
  10. Annett M. A classification of hand preference by association analysis. Br J Psychol 1970;61:303–321.[Medline]
  11. Stippich C, Mohammed J, Kress B, et al. Robust localization and lateralization of human language function: an optimized clinical functional magnetic resonance imaging protocol. Neurosci Lett 2003;346:109–113.[CrossRef][Medline]
  12. Konrad F, Nennig E, Ochmann H, Kress B, Sartor K, Stippich C. Does the individual adaptation of standardized speech paradigmas for clinical functional magnetic resonance imaging (fMRI) effect the localization of the language-dominant hemisphere and of Broca's and Wernicke's areas [in German]. Rofo 2005;177:381–385.[Medline]
  13. Naidich TP, Hof PR, Gannon PJ, Yousry TA, Yousry I. Anatomic substrates of language: emphasizing speech. Neuroimaging Clin N Am 2001;11:305–341.[Medline]
  14. Hajnal JV, Myers R, Oatridge A, Schwieso JE, Young IR, Bydder GM. Artifacts due to stimulus correlated motion in functional imaging of the brain. Magn Reson Med 1994;31:283–291.[Medline]
  15. Ojemann G, Mateer C. Human language cortex: localization of memory, syntax, and sequential motor-phoneme identification systems. Science 1979;205:1401–1403.[Abstract/Free Full Text]
  16. Ramsey NF, Sommer IE, Rutten GJ, Kahn RS. Combined analysis of language tasks in fMRI improves assessment of hemispheric dominance for language functions in individual subjects. Neuroimage 2001;13:719–733.[Medline]
  17. Lehericy S, Cohen L, Bazin B, et al. Functional MR evaluation of temporal and frontal language dominance compared with the Wada test. Neurology 2000;54:1625–1633.[Abstract/Free Full Text]
  18. Szaflarski JP, Binder JR, Possing ET, McKiernan KA, Ward BD, Hammeke TA. Language lateralization in left-handed and ambidextrous people: fMRI data. Neurology 2002;59:238–244.[Abstract/Free Full Text]
  19. Ulmer JL, Hacein-Bey L, Mathews VP, et al. Lesion-induced pseudo-dominance at functional magnetic resonance imaging: implications for preoperative assessments. Neurosurgery 2004;55:569–579.[Medline]
  20. Lurito JT, Lowe MJ, Sartorius C, Mathews VP. Comparison of fMRI and intraoperative direct cortical stimulation in localization of receptive language areas. J Comput Assist Tomogr 2000;24:99–105.[CrossRef][Medline]
  21. Pouratian N, Bookheimer SY, Rex DE, Martin NA, Toga AW. Utility of preoperative functional magnetic resonance imaging for identifying language cortices in patients with vascular malformations. J Neurosurg 2002;97:21–32.[Medline]
  22. Yetkin FZ, Mueller WM, Morris GL, et al. Functional MR activation correlated with intraoperative cortical mapping. AJNR Am J Neuroradiol 1997;18:1311–1315.[Abstract]
  23. Fernandez G, de Greiff A, von Oertzen J, et al. Language mapping in less than 15 minutes: real-time functional MRI during routine clinical investigation. Neuroimage 2001;14:585–594.[CrossRef][Medline]
  24. Weiskopf N, Veit R, Erb M, et al. Physiological self-regulation of regional brain activity using real-time functional magnetic resonance imaging (fMRI): methodology and exemplary data. Neuroimage 2003;19:577–586.[CrossRef][Medline]
  25. Lehericy S, Biondi A, Sourour N, et al. Arteriovenous brain malformations: is functional MR imaging reliable for studying language reorganization in patients? initial observations. Radiology 2002;223:672–682.[Abstract/Free Full Text]
  26. Holodny AI, Schulder M, Liu WC, Maldjian JA, Kalnin AJ. Decreased BOLD functional MR activation of the motor and sensory cortices adjacent to a glioblastoma multiforme: implications for image-guided neurosurgery. AJNR Am J Neuroradiol 1999;20:609–612.[Abstract/Free Full Text]
  27. Schreiber A, Hubbe U, Ziyeh S, Hennig J. The influence of gliomas and nonglial space-occupying lesions on blood-oxygen-level-dependent contrast enhancement. AJNR Am J Neuroradiol 2000;21:1055–1063.[Abstract/Free Full Text]
  28. Lazar RM, Marshall RS, Pile-Spellman J, et al. Anterior translocation of language in patients with left cerebral arteriovenous malformation. Neurology 1997;49:802–808.[Abstract/Free Full Text]
  29. Stippich C, Heiland S, Tronnier V, Mohr A, Sartor K. Functional magnetic resonance imaging: physiological background, technical aspects and prerequisites for clinical use [in German]. Rofo 2002;174:43–49.[Medline]
  30. Stippich C, Kress B, Ochmann H, Tronnier V, Sartor K. Preoperative functional magnetic resonance tomography (FMRI) in patients with rolandic brain tumors: indication, investigation strategy, possibilities and limitations of clinical application [in German]. Rofo 2003;175:1042–1050.[Medline]
  31. Maurer CR Jr, Hill DL, Martin AJ, et al. Investigation of intraoperative brain deformation using a 1.5-T interventional MR system: preliminary results. IEEE Trans Med Imaging 1998;17:817–825.[CrossRef][Medline]
  32. Roberts DW, Hartov A, Kennedy FE, Miga MI, Paulsen KD. Intraoperative brain shift and deformation: a quantitative analysis of cortical displacement in 28 cases. Neurosurgery 1998;43:749–758.[CrossRef][Medline]



This article has been cited by other articles:


Home page
 RadiologyHome page
A. Bizzi, V. Blasi, A. Falini, P. Ferroli, M. Cadioli, U. Danesi, D. Aquino, C. Marras, D. Caldiroli, and G. Broggi
Presurgical Functional MR Imaging of Language and Motor Functions: Validation with Intraoperative Electrocortical Mapping
Radiology, August 1, 2008; 248(2): 579 - 589.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF)
Right arrow Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Stippich, C.
Right arrow Articles by Sartor, K.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Stippich, C.
Right arrow Articles by Sartor, K.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
RADIOLOGY RADIOGRAPHICS RSNA JOURNALS ONLINE