Comparison of the precision of two standardized co-ordinate systems for the quantitation of brain anatomy: preliminary results.

We assessed reproducible definition of two standardized co-ordinate systems for intersubject analysis of brain images. The baselines in the two co-ordinate systems were a modification of the canthomeatal (mCM) line and the anterior-posterior commissural (AC-PC) line. Axial spin-echo MR images of four subjects at 1.5T were used. Operator error was computed from the replicate analyses of two operators. The mCM line was determined by the lens of the eye and the internal auditory canal, and the AC-PC line was determined by the intersection of AC and PC with the interhemispheric fissure. Reproducibility of the mCM markers (SD = 0.59 mm) did not differ significantly from that of the AC-PC line (SD = 0.68 mm). The measurement error of the angle of the baseline (delta alpha), however, was more than 7 times as large for the AC-PC line as for the mCM line. An additional error affecting the rostrocaudal rotation of the co-ordinate systems, attributable to the distance between the anatomic markers, was 2.1 and 3.6 degrees (3 mm and 5 mm slice thickness) for the mCM co-ordinate system and 8.2 and 11.0 degrees (3 mm and 5 mm slice thickness) for the AC-PC system. The AC-PC line based co-ordinate system is therefore, less reproducible than the mCM line based system. This could be improved if a combination of axial and sagittal images were used for the definition of the AC-PC line.