PURPOSE: The three-dimensional stereotactic region of interest template (3DSRT) is computer software, which enables the automatic measurement of regional cerebral blood flow (rCBF). This study was undertaken to compare the rCBF values obtained using the 3DSRT method and the conventional manual tracing method. MATERIALS AND METHODS: Twelve patients with normal brains who underwent technetium-99m L,L-ethyl cysteinate dimer single-photon emission computed tomography studies were enrolled in this study. The brains were divided into 12 segments in each hemisphere, and rCBF was measured in each segment. The regions of interest were automatically placed in the segments in the 3DSRT method and were manually traced by five nuclear medicine technicians in the manual tracing method. The rCBF values obtained were compared between the two methods. Interoperator reliability was evaluated in the manual tracing method. RESULTS: The rCBF values were significantly higher in the manual tracing method than in the 3DSRT method in all segments except for the angular segment. A good correlation was seen between the two methods in the rCBF values in 10 (83.3%) of the 12 brain segments (range of coefficient of determinations: 0.73-0.94). A poor correlation, however, was seen in the pericallosal (0.50) and hippocampal (0.53) segments. Interoperator reliability was lower in these two segments than in the other segments in the manual tracing method. CONCLUSION: Although the rCBF values obtained using the manual tracing method and the 3DSRT method show good correlation in most segments, care is required when comparing the results for the pericallosal and hippocampal segments.
PURPOSE: The three-dimensional stereotactic region of interest template (3DSRT) is computer software, which enables the automatic measurement of regional cerebral blood flow (rCBF). This study was undertaken to compare the rCBF values obtained using the 3DSRT method and the conventional manual tracing method. MATERIALS AND METHODS: Twelve patients with normal brains who underwent technetium-99m L,L-ethyl cysteinate dimer single-photon emission computed tomography studies were enrolled in this study. The brains were divided into 12 segments in each hemisphere, and rCBF was measured in each segment. The regions of interest were automatically placed in the segments in the 3DSRT method and were manually traced by five nuclear medicine technicians in the manual tracing method. The rCBF values obtained were compared between the two methods. Interoperator reliability was evaluated in the manual tracing method. RESULTS: The rCBF values were significantly higher in the manual tracing method than in the 3DSRT method in all segments except for the angular segment. A good correlation was seen between the two methods in the rCBF values in 10 (83.3%) of the 12 brain segments (range of coefficient of determinations: 0.73-0.94). A poor correlation, however, was seen in the pericallosal (0.50) and hippocampal (0.53) segments. Interoperator reliability was lower in these two segments than in the other segments in the manual tracing method. CONCLUSION: Although the rCBF values obtained using the manual tracing method and the 3DSRT method show good correlation in most segments, care is required when comparing the results for the pericallosal and hippocampal segments.