Search

The main goal of this study is to use image registration techniques to combine functional and structural information from different MR-derived images (ADC and rCBV maps, contrast enhanced T1-weighted MR images) in patients diagnosed with brain gliomas. This is a particularly challenging registration problem due to the diverse image characteristics and to the inconsistency between structural and functional image features. Specific registration protocols have been designed, including the selection of critical parameters, for the alignment of the multi-modal information. In addition, we have developed a visual assessment protocol to evaluate the accuracy of registration results obtained with the proposed methods. Preliminary results suggest that misregistration errors are below the functional image resolutions, thus achieving sub-voxel accuracy

SUMMARY: We assessed the influence of inclusion (method 1) and exclusion (method 2) of intratumoral vessels when determining maximum relative cerebral blood volume (rCBVmax) in 3 types of low-grade gliomas (LGGs): astrocytomas, oligoastrocytomas, and oligodendrogliomas. Method 1 yielded significantly higher mean rCBVmax than method 2. However, only method 2 demonstrated a significant (P = .026) association between rCBVmax and membership of a differently ranked histologic category. Exclusion of intratumoral vessels appears, therefore, preferable when determining rCBVmax in LGGs.