Objective To evaluate the application of combination of BOLD-fMRI and diffusion tensor tractography (DTT) in pre-operative evaluation of cerebral tumors located at the left language related areas.Methods A non-vocal button pressing semantic judging paradigm was developed and validated in 10 right-handed volunteers at 3 T.After validation, this protocol combined with DTI were applied to 15 patients with left cerebral tumor prior to surgical resection, and 3 of them had aphasia fMRI data analysis was on subject-specific basis by one-sampled t-test.The distance from the tumor to Broca area and pre-central "hand-knot" area were measured separately.Functional language laterality index (LI) was calculated by taking out Broca area and Wernicke area.Three dimensional architecture of frontal lobe white matter fibers, especially arcuate fasciculus, were visualized using diffusion tensor tractography on Volume-one software.The images demonstrating relationship among tumor, language activation areas and white matter fibers were reviewed by neurosurgeons as part of pre-operative planning.One year after the operation, patients were followed up with MRI and language function test Results The non-vocal semantic judging paradigm successfully detect Broca area, Wernicke area and pre-central "hand-knot" area In 12 of 15 patients, the relationship of Broca area and pre-central motor area to the left brain tumor in language related areas was identified, which make the pre-operative neurosurgical plan applicable to minimize the disruption of language and motor.8 patients had the left language dominant hemisphere, 3 patients with the right language dominant hemisphere and 1 patient with bilateral dominance.The other 3 patients' fMRI data were corrupted by patients' motion.Diffusion tensor images were corrupted by motion in 1 patient but demonstrated the impact of tumor on left arcuate fasciculus in 14 patients.Diffusion tensor tractography showed disruption of left arcuate fasciculus in 3 patients, deviation/deformation in 6 cases and unaffected by tumor in the remaining 5 cases.Pre-operative review of language activation maps and diffusion tensor tractography of arcuate fasciculus influenced the surgical approach in all 12 patients.Post-operatively, one patient died accidentally, the remaining 14 patients showed no tumor recurrence after one year follow-up.No patient showed deterioration of language function, though 3 patients who had aphasia pre-operation improved slightly.Conclusion The combination of fMRI and diffusion tensor tractography is helpful for identifying the relationship of brain tumors with language cortex and white matter fibers, which is important for sparing language function during individually surgical plan.

Objective To study the CT and MRI findings of adenoid cystic carcinoma(ACC) of the head and neck.Methods All 16 cases of ACC were verified by histopathology.Imaging data were analyzed retrospectively.Results The lesion mainly occurred in maxillary sinus in 5 cases,in middle ear and in the floor of mouth in 3 cases respectively,in nasopharynx and parotid in 2 cases respectively,in nasal cavity in 1 case.On CT,the lesion showed ill defined and irregular-shaped soft mass in 6 cases and associated with osteolytic bony destruction in 3 cases.On MR T1WI,the lesion showed isointensity compared to the muscle in 14 cases and slightly hypointensity in 2 cases.On T2WI,the lesion showed isointensity in 6 cases,hyperintensity in 10 cases,heterogeneous intensity in 9 cases.Postcontrast MR imagings demonstrated obvious enhancement in 15 cases,moderate enhancement in 1 case.Lesion with short T2 septum in 7 cases,all of them showed no enhancement.Conclusion CT is the optimal modality in showing osteolytic bony destruction.MRI can demonstrate the invaded extent of the lesions clearly.Combined imaging modalities can provide more comprehensive information in diagnosis and therapy of adenoid cystic carcinoma of the head and neck.