Background: Functional magnetic resonance imaging (fMRI) is a widely used method for research and visualization of the brain function. However, its clinical use is still limited. Our objective was to study fMRI reliability in localizing the primary hand motor cortex (M1) under pathological conditions caused by the proximity of a brain tumour. The results were then compared with standard technique of cortical function mapping-electric cortical stimulation (ECS).
Method: We compared M1 areas localized with the fMRI and ECS in 18 patients with brain tumours in fronto-parietal regions. The 1.5 T blood oxygenation-level dependent (BOLD) fMRI was performed preoperatively using a motor task involving rhythmic touching of the thumb consecutively with other fingers on the same hand contralateral to the affected hemisphere. Each individual fMRI result was displayed at the P < 0.05 significance level corrected for family wise error (more conservative approach) or at the P < 0.001 level uncorrected (less conservative approach) and projected on the T1-weighted image used for neuronavigation.
Findings: In 12 patients (66.6%) we found full agreement between the fMRI and ECS. In 3 patients (16.6%) the overlap was only partial, with one ECS testing position on motor response found outside the BOLD signal cluster. In another 3 cases (16.6%) there was a discrepancy between the two methods. The fMRI sensitivity for localizing the ECS reactive M1 cortex was 71%. The fMRI/ECS consistency was within a 5-mm range in 77% of the testing positions used for ECS which complies with the inherent accuracy of the navigation system.
Conclusions: Because the overlap between the two methods never exceeded 10-mm, we found that the fMRI method correctly guided the ECS to the M1 cortex in 83% of patients. Infiltrative growth of the tumour and collateral oedema were the reasons for the BOLD signal suppression in three patients. Our results support using ECS as a more reliable tool for M1 cortical mapping than fMRI.