Frans Gerritsen

OBJECT: This prospective study was conducted to quantify brain shifts during open cranial surgery, to determine correlations between these shifts and image characteristics, and to assess the impact of postimaging brain distortion on neuronavigation. METHODS: During 48 operations, movements of the cortex on opening, the deep tumor margin, and the cortex at completion were measured relative to the preoperative image position with the aid of an image-guidance system. Bone surface offset was used to assess system accuracy and correct for registration errors. Preoperative images were examined for the presence of edema and to determine tumor volume, midline shift, and depth of the lesion below the skin surface. Results were analyzed for all cases together and separately for four tumor groups: 13 meningiomas, 18 gliomas, 11 nonglial intraaxial lesions, and six skull base lesions. For all 48 cases the mean shift of the cortex after dural opening was 4.6 mm, shift of the deep tumor margin was 5.1 mm, and shift of the cortex at completion was 6.7 mm. Each tumor group displayed unique patterns of shift, with significantly greater shift at depth in meningiomas than gliomas (p = 0.007) and significantly less shift in skull base cases than other groups (p = 0.003). Whereas the preoperative image characteristics correlating with shift of the cortex on opening were the presence of edema and depth of the tumor below skin surface, predictors of shift at depth were the presence of edema, the lesion volume, midline shift, and magnitude of shift of the cortex on opening. CONCLUSIONS: This study quantified intraoperative brain distortion, determined the different behavior of tumors in four pathological groups, and identified preoperative predictors of shift with which the reliability of neuronavigation may be estimated.

Existing theories of the relationship between tidal range, tidal prism, and inlet cross section are reviewed and compared. Existing data on inlets are introduced, as well as the effect of the quantity and type of littoral material on the inlet action.

A crucial step in volume rendering is the design of transfer functions that will highlight those aspects of the volume data that are of interest to the user. For many applications, boundaries carry most of the relevant information. Reliable detection of boundaries is often hampered by limitations of the imaging process, such as blurring and noise. We present a method to identify the materials that form the boundaries. These materials are then used in a new domain that facilitates interactive and semiautomatic design of appropriate transfer functions. We also show how the obtained boundary information can be used in region-growing-based segmentation.