In this paper we analyze volumes of the cerebrospinal fluid spaces for the diagnosis of hydrocephalus, which are served as reference values for future studies. We first present an automatic method to estimate those volumes from a new three-dimensional whole body magnetic resonance imaging sequence. This enables us to statistically analyze the fluid volumes, and to show that the ratio of subarachnoid volume to ventricular one is a proportionality constant for healthy adults (= 10.73), while in range [0.63, 4.61] for hydrocephalus patients. This indicates that a robust distinction between pathological and healthy cases can be achieved by using this ratio as an index.

• MRI assessment of CSF volume is feasible in healthy volunteers/hydrocephalus patients. • CSF volume can be obtained on a whole-body scale. • The ratio of subarachnoid and ventricular CSF is constant in healthy volunteers. • CSF linear discriminant analysis can distinguish between patients and healthy volunteers. • Entire CSF volume imaging is useful for diagnosing and following hydrocephalus.

• MRI can provide helpful information about cerebrospinal fluid volumes. • 3D CSF mapping allows quantitative follow-up in communicating and non-communicating hydrocephalus. • Following intervention, fissures and cisterns readjust in both forms of hydrocephalus. • These findings support the hypothesis of suprasylvian block in communicating hydrocephalus. • 3D mapping may improve shunt dysfunction detection and guide valve pressure settings.

International audienceOur work aims to develop a biomechanical model of hydrocephalus both intended to perform clinical research and to assist the neurosurgeon in diagnosis decisions. Recently, we have defined a new MR imaging sequence based on SPACE (Sampling Perfection with Application optimized Contrast using different flip-angle Evolution). On these images, the cerebrospinal fluid (CSF) appears as a homogeneous hypersignal. Therefore such images are suitable for segmentation and for volume assessment of the CSF. In this paper we present a fully automatic 3D segmentation of such SPACE MRI sequences. We choose a topological approach considering that CSF can be modeled as a simply connected object (i.e. a filled sphere). First an initial object which must be strictly included in the CSF and homotopic to a filled sphere, is determined by using a moment-preserving thresholding. Then a priority function based on an Euclidean distance map is computed in order to control the thickening process that adds "simple points" to the initial thresholded object. A point is called simple if its addition or its suppression does not result in change of topology neither for the object, nor for the background. The method is validated by measuring fluid volume of brain phantoms and by comparing our volume assessments on clinical data to those derived from a segmentation controlled by expert physicians. Then we show that a distinction between pathological cases and healthy adult people can be achieved by a linear discriminant analysis on volumes of the ventricular and intracranial subarachnoid spaces

Cerebrospinal fluid imaging plays a significant role in the clinical diagnosis of brain disorders, such as hydrocephalus and Alzheimer's disease. While three-dimensional images of cerebrospinal fluid are very detailed, the complex structures they contain can be time-consuming and laborious to interpret. This paper presents a simple technique that represents the intracranial cerebrospinal fluid distribution as a two-dimensional image in such a way that the total fluid volume is preserved. We call this a volumetric relief map, and show its effectiveness in a characterization and analysis of fluid distributions and networks in hydrocephalus patients and healthy adults.