Alignment correction algorithm for transformation of stereotactic anterior commissure/posterior commissure-based coordinates for image-guided functional neurosurgery

Abstract

To the Editor: I would like to add some comments to the article of Krauss et al. (6). We routinely use a trigonometric method of calculation of the stereotactic coordinates of invisible targets for functional neurosurgery, based only on axial CT scans. From December 1996 to November 1998, we performed 39 implantations for deep brain stimulation (28 subthalamic nuclei, 10 globus pallidus internus, 1 ventralis intermedius, in 20 Parkinsonian patients), with postoperative CT/MRI and clinical (improvement in UPDRS, Schwab-England, Hoehn-Yahr scales) confirmation of the accuracy of the procedure. This trigonometric method was presented at the 12th Congress of the European Society for Stereotactic and Functional Neurosurgery (Milan, June 12-15, 1996) and published in Acta Neurochirurgica (Wien) (138:615, 1996) and Rivista di Neuroradiologia (11:43-50, 1998) . We emphasize the points in which our method differs from that of Krauss et al., and we described how our method enlarges the applicability of the method itself. Besides the anterior and posterior commissures (AC and PC), only on axial CT scans, we calculate a third point, I, the center of the infundibular recess of the third ventricle. This enables us to fix a reference plane (the midsagittal plane of the third ventricle), to which we refer the distances of the "invisible targets" (subthalamic nucleus, inferior parietal gyrus, ventralis intermedius) codified by the Stereotactic Atlases(2,9). In this way, we have our own "internal" reference system that enables us not "to fix the base line (y axis) as parallel to the intercommissural (IC) line as possible," even though it is reassuring to aim at a parallelism between the y axis and the IC line. For this reason, we position our stereotactic frame (Leksell G; Elekta Instruments, Stockholm, Sweden) with a +12-degree angulation to the O-M line (11). Sagittal and coronal reformatted CT images are not necessary to identify AC and P. We use only axial CT scans to obtain the coordinates of AC and PC, which usually are found on different slices and the coordinates of the third point I. The exact IC length is obtained by the "double target method" (1), avoiding sagittal MRI scans and/or CT reformatted images, which are less reliable. It is then unnecessary "to check with the length measured on the MRI scans." Obviously proportional corrections are made in comparison with the standard values (IC length, distances of the invisible targets from the midpoint of IC, MP, on axial, sagittal, and coronal planes) of the Stereotactic Atlases, and the whole procedure is repeated, for a double check, according to the atlas of Andrew-Watkins (2), referring to the posteroinferior border of the foramen of Monro (FM) to the anterior border of PC and to the point I (FM-PC-I plane). As we refer to AC-PC-I and/or FM-PC-I planes instead of the IC line, it is unnecessary to correct the tilt of IC separately (roll, pitch, yaw) on the different planes. The target's coordinates (Tx, Ty, Tz) are directly calculated applying the formulas of the Appendix, and it is unnecessary to transfer the values of the CT coordinates to the center of the frame. They are directly referred to the frame. Tx and Ty values are projected on the axial CT slice corresponding with Tz. In this way, the calculated target is visualized before the time of surgery. All the values are calculated in real time through a simple QBASIC PC program. Our conclusions are as follows: 1) using a third point, I, and referring to a plane (AC-PC-I or FM-PC-I), the meticulous positioning of the frame with IC parallel to the y axis is not critical; 2) the third point, I, enables us to use our method, in contrast with Krauss, even in the case of "evident roll (tilting of the head relative to the frame along the y axis)"; 3) we do not agree with Krauss et al. that their method "could be used with stereotactic MRI," because we consider MRI not fully reliable (3-5,7,8,10,12); 4) neurophysiological and clinical intraoperative confirmation through microrecording or semi-microrecording and stimulation is mandatory in each case. Franco Ammannati Lorenzo Bordi Paolo Gronchi Florence, Italy