Brain single photon emission computed tomography at term age for predicting cerebral palsy after preterm birth

In order to assess the predictive value of neonatal brain perfusion with single photon emission computed tomography (SPET) with regard to neuromotor outcome at a corrected age of 18 months, 34 infants with birth weight <1,500 g and gestation age <34 weeks underwent brain technetium-99m ethylcysteinate dimer (99Tc(m)-ECD) SPET at term age. The perfusion defects were estimated by visual interpretation. Consecutive semiquantitative assessment was made in 26 cases and reference values for the tracer were collected from images of 17 preterm infants with normal outcome after the follow-up period. Relative regional cortical (frontal, sensorimotor, parietal and occipital), cerebellar and thalamic perfusion levels were evaluated in middle sagittal slices and hemispheric asymmetries in transaxial slices. Perfusion defects predicted cerebral palsy (CP) (n = 11) with 82% sensitivity, 70% specificity and 74% accuracy, the corresponding figures for ultrasound (US) being 73, 83 and 79%, respectively. The sensitivity of SPET in predicting moderate or severe CP (n = 7) was 100% and the specificity 67%, the corresponding figures for US being 71% and 74%, respectively. Brain SPET seems to identify the most severe forms of CP in preterm infants very well at term age, but cannot identify all mild ones. In addition to a low specificity, the radiation exposure restricts usefulness of the method for clinical purposes.