[Brain perfusion SPECT in the prognostic assessment of stroke]

To compare the diagnostic yield of CT scan and perfusion SPECT on admission and its prognostic value in clinical outcome.

25 ischemic stroke cases were studied on admission (<24 h) and at 30-60 days by CT scan, 99mTc-HMPAO-SPECT and neurological scales. Infarct size and severity on SPECT were assessed: visually "Total Weighted Score," added value in 22 areas, and by several semiquantitative count-based indices.

Sensitivity: the first CT scan was positive in 24% patients, initial SPECT in 75% (73% of pure subcortical infarcts and 91% of those with cortical involvement). Localization: kappa: 0.725 between SPECT findings on admission and those in control-CT at 5 days. Extent and severity: correlations between count-based and visual indices (r: >0.719), the latter correlated slightly better with clinical scales. Both predicted similarly (Rho>0.739) infarct size in CT diagnostic scan. Early Outcome: There were statistical differences between deceased and survivors in SPECT (<24h) indices and CT-infarct size (mean 5 days), but not in neurological scores on admission. Long term Outcome: Correlation of initial SPECT indices with follow-up functional scores (SNSLP, Barthel index; mean 37 days) was only significant for visual SPECT indices (Rho:0.560 to 0.620). Nevertheless the best predictor of functional status on discharge was the Barthel Index on admission.

1) Early SPECT has good sensitivity and accurate infarct size prediction so it can be a useful tool for deciding thrombolytic therapy; 2) Visual scores perform as well as more complex indices; 3) Infarct volume seems to be a critical determinant in vital outcome; other factors (strategic localization, etc.) might influence long term functional status.