Quantitative ADC in bilirubin encephalopathy

Abstract

It was with great interest that we read the paper by Cece and colleagues [1] entitled ‘Diffusion-weighted imaging of patients with neonatal bilirubin encephalopathy.’ The authors conclude that quantitative apparent diffusion coefficient (ADC) measurement could be a promising marker of neonatal bilirubin encephalopathy. The article, however, raises several questions that merit clarification. First, there is little doubt that several in this cohort of infants with peak bilirubin levels in the range of 25–58 mg/dl developed permanent bilirubin-induced neuronal injury and adverse long-term neurodevelopmental sequelae of chronic bilirubin encephalopathy (CBE). Indeed, most demonstrated overt signs of intermediate to advanced stages of acute bilirubin encephalopathy (ABE) and/or peak TSB levels [35 mg/dl, a threshold in the US Pilot Kernicterus registry that was more often than not associated with moderate to severe chronic neurodevelopmental impairment [2, 3]. Some, however, may have had reversible ABE and escaped their hazardous bilirubin exposures unscathed given the timely use of exchange transfusion [4]. Neurological findings at the time of hospitalization discharge and long-term neurodevelopmental evaluations, however, are not reported. Similarly, auditory brainstem evoked responses are not detailed. It is important to distinguish infants who developed CBE from those who did not and compare ADC values between these two groups to ascertain whether ADC is a potential marker of CBE or a marker of ABE alone and not its progression to CBE. Follow-up MRI studies to confirm neuroradiographic evidence of CBE would be of interest in this regard but were not reported. Second, although quantitative ADC values were statistically increased relative to controls, the numeric difference was modest and its clinical relevance uncertain, particularly since ADC is sensitive to multiple technical and pathologic factors. In addition, the region of interest (ROI) depicted in Figure 1 does not selectively localize to the globus pallidus, but also includes a portion of the posterior limb of the internal capsule. There is no inter-rater/intra-rater reliability measurement given for the manual drawing of the ROI, which is prone to error. The subthalamic area is also known to be involved in kernicterus, and this region was not measured in the report. More details about the controls would be of interest as well, as there are very few neonatal cases that have a spinal disorder and a normal brain MRI. Third, Figure 3 demonstrates a statistically significant relationship between total bilirubin levels and ADC values. However, the reported r value of 0.41 indicates that only 16 % (r = 0.16) of the variability in ADC can be accounted for by total bilirubin, i.e., 86 % is accounted for by other factors. This comment refers to the article available at doi:10.1007/s11604-012-0166-4.