• Differentiation between brain tumors and tumefactive lesion by DWI
• Differential diagnosis of brain tumors
• WHO grading of brain tumors and DWI
• Novel finding in DWI for differential diagnosis

Radiologists, neurologists, neurosurgeons and clinicians interested in diagnosis of neuroradiologyThe presentation aims to provide tips for diagnosis of brain tumors using DWI and ADC.Diffusion-weighted MR imaging produces unique contrast images based on differences in mobility of water protons between tissues. Highly cellular tissues have reduced water diffusivity, appearing hyper-intense on DWI. As the apparent diffusion coefficient (ADC) calculated from diffusion-weighted images (DWI) is associated with tumor cellularity(1,2), and is considered an important biomarker of cancer(3), we sought to determine if ADC can be used to differentiate and grade brain tumors. Furthermore, high b-value DWI is more sensitive to diffusion, and is diagnostically useful in various clinical settings (4). Therefore, we also compared the power of high- and standard b-value (b-4000, b-1000) imaging on a 3-Tesla (3T) MR instrument. Preoperative MR images including DWI (b=1000 s/mm2 and b=4000 s/mm2) of patients with histologically confirmed brain tumors admitted in our institution were analyzed retrospectively. ADC of tumor was measured by placing multiple regions of interest (ROI) on ADC maps. ROIs in contrast-enhanced tumors were placed at the enhanced site confirmed on contrast-enhanced T1-weighted MR images. In weakly enhancing or non-enhancing tumors, areas of hyper-intensity on FLAIR images were identified as tumor and ROIs were carefully placed in the solid tumor components after comparing ADC maps with other MR images. Hemorrhagic and cystic lesions were avoided using T1-, T2-, FLAIR-, and T2* MR images. All brain tumor subgroups were analyzed. We also evaluated the relationship between ADCs and histology including tumor cellularity in glioblastoma and malignant lymphoma. Presence of non-enhancing peritumoral DW high intensity lesion (NePDHL) was confirmed in both DWI sequences, and were termed “Definite” (D-NePDHL) if present as hyper-intensity in both sequences.ADC at b-4000 was associated with tumor cellularity more significantly than ADC at b-1000 (5). A significant negative correlation existed between ADC and astrocytic tumors of World Health Organization grades 2–4 (grade 2 vs grades 3 and 4, accuracy 91.3% [P < 0.01]; grade 3 vs 4, accuracy 82.4% [P < 0.01]) (6). ADC of dysembryoplastic neuroepithelial tumors (DNTs) was higher than that of astrocytic grade 2 tumors (accuracy, 100%) and other glioneuronal tumors (6). ADC of malignant lymphomas was lower than that of glioblastomas and metastatic tumors (accuracy, 83.6%; P < 0.01) (5,6). ADC of embryonal tumors including medulloblastoma was lower than that of ependymomas (accuracy, 100%)(6). ADC of meningiomas was lower than that of schwannomas (accuracy, 92.4%; P < 0.01). ADC of craniopharyngiomas was higher than that of pituitary adenomas (accuracy, 85.2%; P < 0.05) and germ cell tumors (accuracy 91.7%: P < 0.01)(7). ADC of epidermoid tumors was lower than that of chordomas (accuracy, 100%)(6). ADC of hemorrhagic pituitary adenomas was lower than of the other lesions with similar appearance on conventional MRI (non-hemorrhagic pituitary adenomas, craniopharyngiomas, Rathke’s cleft cysts; accuracy100%)(8). Although ADCs of meningiomas were inversely associated with the histological grade (P < 0.01), there was considerable overlap in grading (9). In 25% of glioblastoma patients D-NePDHL was present while this feature was conspicuously absent in other brain tumors including malignant lymphoma and metastatic brain tumors. Cases with D-NePDHL had significantly early distant/dissemination recurrence (p < 0.0001) and carried poor prognosis (p = 0.0007).DWI studies at higher diffusion gradient strength (b-values) have been used for the diagnosis of acute stroke, assessment of lesion-to-normal contrast in neurodegenerative diseases, prediction of the glioma grade, and for differentiation of some brain tumors (4,10-12). Studies of multi-component diffusion in brain tissue have demonstrated that the slow component on high b-value is more sensitive than on b-1000 DWI, an observation which supports our finding that ADC based on higher b-values reflects changes in tumor cellularity more accurately(5). Our findings further confirm that DWI and ADC calculations based on high b-value yield more reliable results. Increasing the b-value up to 4000 s/mm2, converts mono-exponential diffusion within the brain tissue to bi-exponential. Diffusion consists of slow- and fast-components corresponding with intra- and extra-cellular diffusions, respectively (13,14). As the fast-component dominates at relatively low b-values (<1000 s/mm2), ADC values primarily reflect fast diffusion and thus the amount of extracellular space in tumor. On the contrary, the slow-component dominates at relatively high b-values (~4000 s/mm2) and ADC values primarily reflect slow diffusion and thus the amount of intracellular space in tumor(13). The slower diffusion component is supposed to reflect the concentration of water-bound macromolecules, cell size, and tissue architecture (tortuosity) (15).ADC is useful for differentiation of some human brain tumors, particularly DNT, malignant lymphomas versus glioblastomas and metastatic tumors, and ependymomas versus embryonal tumors. Presence of D-NePDHL is very specific in glioblastoma and is associated with poor prognosis. D-NePDHL is a significant indicator for early distant/dissemination recurrence in glioblastoma. Study of DWI of peritumoral region is useful for differentiation of glioblastoma.