White matter lesions in phenylketonuria: Evaluation with magnetic resonance imaging and magnetic resonance spectroscopy

Phenylketonuria (PKU) is a congenital metabolic autosomic recessive disease, caused by a deficit in the liver of phenyl-alanine hydroxylase, the enzyme responsible for conversion of phenyl-alanine (PHE) into tyrosine. Reduction of this enzymatic activity is responsible for increased phenyl-alanine in blood and tissues and, above all, in brain. Accumulation of PHE causes neural damage which produces a typical clinical picture with mental retardation, psychiatric symptoms and epilepsy. It is now possible to diagnose this disease early (with neonatal screening), before irreversible clinical symptoms reflecting central nervous system injury appear. Early diagnosis allows timely onset of therapy (the only possible) consisting of a special diet with reduced intake of PHE (integrated with a mix of aminoacids) whose objective is to keep levels of PHE low in the blood (3-6 mg/dl). Magnetic Resonance Imaging (MRI) is the elective diagnostic tool to evaluate in vivo the involvement of the brain in PKU. Previous MRI morphological studies in patients with PKU have reported various focal symmetrical lesions in periventricular white matter (especially parieto-occipital) of patients with PKU with PHE blood values higher than 10 mg/dl. These lesions, whose importance is not yet clear, seem to represent a reversible structural alteration of myelin, since they regress if blood PHE decreases. Proton magnetic resonance spectroscopy ( ¹H-MRS) can measure in vivo brain metabolites which could help determine the nature of white matter lesions. In particular, changes in NAA (a marker of neuronal integrity) or mI (a potential astrocytic marker) could point to possible neurochemical dysfunction, whereas Cho levels may parallel the degree of the tissue myelination. The purpose of the present study was to evaluate morphologically and biochemically the regional specificity of white matter lesions with structural MRI and with ¹H-MRSI. The study included 12 patients with PKU ten to 42 years of age. All patients underwent structural MRI scans while eight of them were also studied with ¹H-MRSI. Structural MRI lesions in white matter were analyzed both qualitatively (signal intensity) and quantitatively (location and extension). ¹H-MRSI metabolites were measured as the ratio of the area under each peak: NAA/Cr, NAA/Cho, Cho/Cr. Analysis of location and extension of the lesion on structural MRI data showed limited involvement of parieto-occipital white matter in three cases (with isointense or vaguely hypointense lesions in T1, and moderately hyperintense lesions in T2); medium involvement in six cases (with fairly hypointense or isointense lesions in T1, fairly or moderately hyperintense lesions in T2); serious involvement in three cases (with isointense or fairly hypointense lesions in T1, and fairly hyperintense lesions in T2). As for ¹H-MRSI data, ANOVA showed a significant reduction of NAA/Cho and increase in Cho/Cr in white matter lesions, but no change in NAA/Cr. No correlation was found between clinical parameters and morphological or spectroscopic data. In conclusion, our morphological MRI data confirmed the presence of multiple signal alterations, focal and symmetrical, in deep periventricular white matter (especially posterior), with occasional involvement of subcortical white matter. However, these lesions do not seem to be strongly predictive of clinical outcome. ¹H-MRSI data suggest increased Cho levels in white matter lesions. Since Cho is thought to reflect membrane turnover, these data may support the demyelinating nature of lesions, consistent with earlier post mortem studies.