Routine controls for the geometric accuracy assessment of MR images used in neurosurgery

This paper describes the methodology we set up in our hospital to investigate whether or not the MR images from our MRI device are geometrically reliable for use as reference images in stereotactic neurosurgery procedures. In fact, in these clinical procedures geometric accuracy is a prime concern. MRI is often the only diagnostic methodology by which the pathology of interest is detectable. On the other hand, the physical measurements on which this technology lies do non guarantee geometrical accuracy.

So we set up a measurement protocol to assess the geometric accuracy of an MR image. To reach this goal we use a self-designed phantom. On this simple phantom, simulating a human head, the stereotactic ring used for neurosurgical procedures was mounted. We made acquisitions in CT and MRI, and compared the images found.

Our findings were that, with our MRI device, the geometric accuracy obtained depends only on the acquisition matrix used, and is not affected by other technological factors.

The method here described can be used in any site where stereotactic neurosurgical procedures are applied on the basis of MR images, as it is simple and cheap, both from an economic point of view, and from a machine-time and personnel-time point of view.