Stereotactic Method for Determining Anatomical Localization in Physiological Brain Images
Peter T. Fox, Joel S. Perlmutter, Marcus E. Raichle · Journal of Cerebral Blood Flow & Metabolism · 1984
To the Editor: Physiological brain images pro duced by positron emission tomography (PET) or single photon emission computed tomography (SPECT) are seductive, appearing self-sufficient and inviting evaluation by visual inspection. The eye readily makes the assumption that physiological demarcations correspond directly to familiar ana tomical boundaries. Unfortunately, anatomy is only indirectly and indistinctly portrayed in such images. Regional analysis of a physiological measurement, however, has meaning only when the anatomical structure corresponding to each area within the to mographic image is known. Moreover, the termi nology with which anatomical location is expressed must be precise and universally comprehensible. Therefore, we echo the recommendation in Dr. Mazziotta's editorial in this issue (p. 481) that a method for determining anatomical location within physiological brain images that is itself independent of the physiological image be developed, standard ized, and accepted into general use. We have developed and validated a stereotactic method for determining anatomical localization within physiological brain images that meets many of Dr. Mazziotta's proposed criteria (Fox et ai., 1985). The relation of the planes of section of the physiological image to the cranial vault must be per manently recorded. We accomplish this with a lat eral skull radiograph obtained after head positioning and immobilization. A series of well-defined mea surements, followed by computer-programmed tri gonometric coordinate transformations ascribe a stereotactic coordinate (bicommissural coordinate system) to each point within the tomographic brain image. Both the underlying anatomical assumptions and the overall accuracy of this method have been extensively validated. Thus, anatomy is determined and described objectively, accurately, and unambiguously. Comparisons can be made between sub jects, between laboratories, between imaging mo dalities, and even with postmortem data. Total in dependence from the physiological image makes this method equally applicable to all forms of phys iological imaging, including measurements of he modynamics, metabolism, pharmacokinetics, and biochemistry. In fact, this method can be readily applied to anatomical brain imaging techniques, to facilitate research applications of structural tomog raphy and to allow correlations between PET and proton nuclear magnetic resonance images or x-ray computed tomography. It should be noted that although every pixel within the tomographic image is ascribed a stereo tactic coordinate, the grouping of these coordinates into regions of interest remains at the discretion of the investigator. This approach allows the investi gator to select the angle and level of the planes of section most appropriate for each study, without the necessity of scanning in planes parallel to a refer ence plane. We anticipate that this method will fulfill the needs of many laboratories for the determination and description of specific anatomical locations within physiological images.