Two-dimensional SPECT display and three methods for integrated visualizatio
n of SPECT and MRI patient data are evaluated in a multiobserver study to d
etermine whether localization of functional data can be improved by adding
anatomical information to the display. Methods: SPECT and MRI data of 30 pa
tients were gathered and presented using four types of display: one of SPEC
T in isolation, two integrated two-dimensional displays and one integrated
three-dimensional display. Cold and hot spots in the peripheral cortex were
preselected and indicated on black-and-white hard copies of the image data
. Nuclear medicine physicians were asked to assign the corresponding spots
in the image data on the computer screen to a lobe and a gyrus and give a c
onfidence rating for both localizations. Interobserver agreement using kapp
a statistics and average confidence ratings were assessed to interpret the
reported observations. Results: Both the interobserver agreement and the co
nfidence of the observers were greater for the integrated two-dimensional d
isplays than for the two-dimensional SPECT display. An additional increase
in agreement and confidence was seen with the integrated three-dimensional
display. Conclusion: Integrated display of SPECT and MR brain images provid
es better localization of cerebral blood perfusion abnormalities in the per
ipheral cortex in relation to the anatomy of the brain than single-modality
display and increases the confidence of the observer.