Task evoked fMRI has been used successfully for localization of eloquent brain regions prior to the performance of brain surgery; however, conventional fMRI requires patient co-operation and is not often suitable for patients with physical impairment and/or cognitive dysfunction. An alternative method to overcome these challenges of conventional fMRI could be the use of resting state fMRI (rsfMRI) for pre-surgical planning¹. In this study, we have investigated the integration of the information from rsfMRI and dynamic contrast enhanced (DCE) MRI derived perfusion maps to enhance the capabilities of rsfMRI to understand brain tumors. Contrast enhanced imaging shows only the enhancing part of the tumor while conventional FLAIR images shows tumor and edema. Perfusion imaging helps in identifying the enhancing tumor boundary from non-enhancing tumor and edema. By combining rsMRI with perfusion we can identify the relationship of the functionally connected eloquent areas of the brain with the tumor boundaries that will help in defining the surgical resection.

Data Acquisition: A total of 22 patients diagnosed with brain tumor were included in this retrospective study in accordance with local ethical committee approval. All the imaging were done on 3.0 T scanner (Inginia, Philips Healthcare). Imaging protocol of rsfMRI echo-planar imaging (EPI) with TR/TE:3000/35ms, 90° Flip Angle, 96×94 matrix, 230×230 FOV, 4mm slice, no gap, 30 section and 120 frames. DCE-MRI data were acquired with TR/TE:4.4/2.1ms, 12 slice with 6mm thickness, 32 dynamic with 3.9s temporality. Conventional MR imaging includes T1-weighted, T1 weighted post contrast (TR/TE=700/25ms, 0.9 mm slice thickness, 280×278mm² matrix, 250×250mm² FOV ) and FLAIR images ( TR/TE=4700/280ms, 1650ms inversion time, 1.2mm slice thickness, 90˚ flip angle, 220×220 matrix, 247×247mm² FOV).

Image processing: Using SPM8 software, the functional images were slice-time and motion corrected, co-registered to the anatomical template and smoothed with a kernel of 8 mm (full-width-at-half-maximum). To correct for physiological fluctuations, the time-series from the cerebrospinal fluid (CSF) and white matter were included as co-variates in the linear regression analysis. All the time series were band pass filtered between 0.01 to 0.1 mHz. Gray matter, white matter, and CSF voxels were segmented from the T1-weighted images using Freesurfer. Further left and right motor and language regions were also segmented using Freesurfer for analysis. Cerebral blood volume (CBV) map was estimated from DCE-MRI data using in-house developed software².

Functional connectivity: Six seed time series were computed by taking the mean of the time series in left and right motor, left and right Broca; and left and right Wernicke regions. All the voxels in brain were correlated with these time series to extract six functional connectivity maps each. All the maps were combined to generate one map by taking maximum correlation for each voxel and finally this map was superimposed on FLAIR and CBV maps that may guide pre-surgical planning.

Representative figure 1 shows the results on one patient; functional connectivity (thresholded) results super imposed on FLAIR and CBV (gray scale) in (a) motor area, (b) broca, (c) Wernicke and (d) rCBV. Figure 2 shows the functional connectivity map superimposed on FLAIR and CBV on a patient dataset, functional connectivity on all the maps are combined in one map. Similar results are obtained on overall 22 patients.Resting state fMRI based functional connectivity results on perfusion maps and structural images can potentially help clinicians in pre-surgical planning. The regions near tumor that show high correlation are seen as active regions that might contribute to motor/language activities, and perfusion maps help clinicians in demarking the tumor region. An integration of both the techniques may help clinicians in better surgical planning.