OBJECTIVE: The involvement of globus pallidus has been found in neonates with acute bilirubin encephalopathy (ABE), but so far the relationship between the severity of hyperbilirubinemia and changes of globus pallidus has not been studied further. The present study was conducted to understand possible relationship between the MRI signal changes of globus pallidus and severe hyperbilirubinemia in the neonates to provide evidences for diagnosis of bilirubin encephalopathy and prediction of outcome. METHODS: Thirty-six neonates with severe hyperbilirubinemia (total serum bilirubin, i.e., TSB > 342 micromol/L) received MRI examination in a magnetic field with the strength of 1.5 - 3.0 Tesla at ages of 10 +/- 6 (2 - 34) days of life, of whom 15 were assessed as acute bilirubin encephalopathy (ABE). Routine T1WI with three kinds of sequences (SE, IR and FFE), T2WI and DWI with two kinds of sequences (EPI-SE and DW SSh SENSE) were applied. Two neuroradiologists who knew nothing about the clinical history analyzed MRI findings. TSB, unconjugated bilirubin (UCB), conjugated bilirubin (CB) and albumin (ALB) in all these neonates were measured with the same methods and analyzer. RESULTS: Symmetric hyperintense globus pallidus was shown on T1WI in 20 neonates, of whom three had symmetric hyperintense globus pallidus on T2WI (TSB = 745.3 micromol/L, 735.7 micromol/L, 707.6 micromol/L, respectively) at the same time. Remarkably higher TSB, molar ratio of TSB and ALB (B/A) and UCB were found in 20 neonates with hyperintense globus pallidus as compared to 16 cases without changes of globus pallidus (P = 0.000). No abnormal signal changes were found on DWI for all neonates; 9 neonates with TSB ranged from 342.0 micromol/L to 427.5 micromol/L did not show any abnormalities of globus pallidus, 3 of 7 neonates with TSB ranged from 427.5 micromol/L to 513.0 micromol/L, but 17 of 20 neonates with TSB more than 513.0 micromol/L showed distinct hyperintense globus pallidus. Hyperintense globus pallidus on T1WI was related to the severity of hyperbilirubinemia (Fisher's Exact Test, P = 0.000). Hyperintense globus pallidus on T1WI was found in all the 15 neonates with ABE (Fisher's Exact Test, P = 0.000), at the same time, hyperintense globus pallidus on T2WI was found in 3 cases with severe ABE. Six infants received the second MRI at ages from 2 months to 12 months, two of three infants with ABE in neonatal period showed the signal changes of globes pallidus from T1WI hyperintensity to T2WI hyperintensity and they developed cerebral palsy. The other one with normal signal showed hearing dysfunction. Another three infants without history of ABE did not show abnormal signals of globus pallidus and developed normally. CONCLUSIONS: The symmetric involvement of globus pallidus with hyperintense signals on MRI T1WI indicates the severity of hyperbilirubinemia, which relates to time of exposure to hyperbilirubinemia and is a characteristic finding of ABE. The signal switch of hyperintense globus pallidus from T1WI to T2WI predicts poor outcome.
OBJECTIVE: The involvement of globus pallidus has been found in neonates with acute bilirubinencephalopathy (ABE), but so far the relationship between the severity of hyperbilirubinemia and changes of globus pallidus has not been studied further. The present study was conducted to understand possible relationship between the MRI signal changes of globus pallidus and severe hyperbilirubinemia in the neonates to provide evidences for diagnosis of bilirubinencephalopathy and prediction of outcome. METHODS: Thirty-six neonates with severe hyperbilirubinemia (total serum bilirubin, i.e., TSB > 342 micromol/L) received MRI examination in a magnetic field with the strength of 1.5 - 3.0 Tesla at ages of 10 +/- 6 (2 - 34) days of life, of whom 15 were assessed as acute bilirubinencephalopathy (ABE). Routine T1WI with three kinds of sequences (SE, IR and FFE), T2WI and DWI with two kinds of sequences (EPI-SE and DW SSh SENSE) were applied. Two neuroradiologists who knew nothing about the clinical history analyzed MRI findings. TSB, unconjugated bilirubin (UCB), conjugated bilirubin (CB) and albumin (ALB) in all these neonates were measured with the same methods and analyzer. RESULTS: Symmetric hyperintense globus pallidus was shown on T1WI in 20 neonates, of whom three had symmetric hyperintense globus pallidus on T2WI (TSB = 745.3 micromol/L, 735.7 micromol/L, 707.6 micromol/L, respectively) at the same time. Remarkably higher TSB, molar ratio of TSB and ALB (B/A) and UCB were found in 20 neonates with hyperintense globus pallidus as compared to 16 cases without changes of globus pallidus (P = 0.000). No abnormal signal changes were found on DWI for all neonates; 9 neonates with TSB ranged from 342.0 micromol/L to 427.5 micromol/L did not show any abnormalities of globus pallidus, 3 of 7 neonates with TSB ranged from 427.5 micromol/L to 513.0 micromol/L, but 17 of 20 neonates with TSB more than 513.0 micromol/L showed distinct hyperintense globus pallidus. Hyperintense globus pallidus on T1WI was related to the severity of hyperbilirubinemia (Fisher's Exact Test, P = 0.000). Hyperintense globus pallidus on T1WI was found in all the 15 neonates with ABE (Fisher's Exact Test, P = 0.000), at the same time, hyperintense globus pallidus on T2WI was found in 3 cases with severe ABE. Six infants received the second MRI at ages from 2 months to 12 months, two of three infants with ABE in neonatal period showed the signal changes of globes pallidus from T1WI hyperintensity to T2WI hyperintensity and they developed cerebral palsy. The other one with normal signal showed hearing dysfunction. Another three infants without history of ABE did not show abnormal signals of globus pallidus and developed normally. CONCLUSIONS: The symmetric involvement of globus pallidus with hyperintense signals on MRI T1WI indicates the severity of hyperbilirubinemia, which relates to time of exposure to hyperbilirubinemia and is a characteristic finding of ABE. The signal switch of hyperintense globus pallidus from T1WI to T2WI predicts poor outcome.