With regard to MRI diagnosis of PSP, all the attention has been given to atrophy of the midbrain where pathologic changes are known to occur (43). Signal changes in this region have also been noted. Less attention has been paid to basal ganglia abnormalities and cerebral atrophy, which are more variable and less useful for the diagnosis.
MRI simply confirmed what pneumoencephalography first, and later CT (44,45) have shown: demonstration of midbrain atrophy is the cardinal neuroradiologic feature of PSP diagnosis. This demonstration can be achieved by observing the midbrain size either in axial sections or in the midline sagittal section (17,23,25). The latter has the advantage of showing a direct comparison between the size of the midbrain and that of the pons, not only in the sagittal diameter, but also in "height," or vertical or caudo-cranial diameter. The midbrain size is, in fact, reduced in all the directions. The upper profile of the normal midbrain in sagittal section is characterized by a slight superior convexity with the highest point approximately midway between the aqueduct and the mamillary bodies. In PSP, this profile flattens and may even become concave in its posterior part, because of atrophy and reduced height of the midbrain tegmentum associated with enlargement of the third ventricle. Thinning of the tectum particularly in its superior part may also be observed in sagittal sections (25) (Fig. 4A). Atrophy of the midbrain is found in 75-89% of the clinically diagnosed PSP patients (19,23,46).
Simple linear measurements of the midbrain have been suggested in axial and sagittal sections (47). No overlap was found between the measures obtained in PSP patients (range of midbrain diameter: 11-15 mm) vs PD patients and normal subjects (range 17-19 and 17-20 mm, respectively), whereas some overlapping occurred between PSP and MSA-P patients (range 14-19 mm, mean 16.7).
As mentioned above, disproportionate dilatation of the third ventricle, compared to the lateral ones, may be caused by diencephalic extension of atrophy. This finding was evident in about one-fourth of our cases.
One absolutely marginal feature that is sometimes seen in PSP patients in sagittal MR images is hyperextension of the head. This feature, however, may be a clue for the diagnosis. The hyperex-tended position may be assumed during the examination by patients who were correctly positioned by the MRI technician at the beginning of the examination. The neuroradiologist who recognizes the unusual hyperextension of the head should immediately observe the midbrain, looking for atrophy and signal changes that may confirm the suspicion of PSP (17,25).
Signal changes in the midbrain are the second aspect that one should look for in suspected PSP patients. When present, they consist of mild hyperintensity in proton density and T2-weighted images, mainly in the tegmentum and periaqueductal region where gliosis and tau-positive lesions are found in pathological sections (48,49) (Fig. 4B). These abnormalities may extend caudally to the pontine tegmentum (49). Signal changes in the midbrain are recognizable in about 60% of the PSP patients (46).
Mild to moderate cerebral atrophy may be observed in most of the patients, sometimes more marked in the frontal and temporal lobes (19), but without the focal and asymmetric distribution found in CBD (46) (see the following section).
Occasional abnormalities in the basal ganglia (atrophy of the lenticular nucleus with pallidal hyperintensity) have sometimes been reported (19), but the typical combination of T2 putaminal hypointensity with lateral hyperintense rim seen in 1.5 T studies of MSA-P patients are absent (25,28).
DWI has only rarely been used to study PSP. In one study, regional ADCs (rADCs) were increased in the basal ganglia of PSP patients; the values allowed discrimination from patients with PD but did not discriminate PSP and MSA-P (35). No results of the MRI observations were given so that it remains undefined whether DWI performs better, equally, or worse than MRI in discriminating PSP, MSA-P, and PD. In another study on only five PSP patients, ADCs in the prefrontal and precentral white matter were higher than in controls (50).
MRS reports on PSP are also rare. NAA deficit in the lentiform nucleus, consistent with neuronal loss, was found in PSP and in MSA patients, whereas patients with PD did not differ from controls (33). Tedeschi et al. (51) examined patients with PD, PSP, and CBD. Compared with the control subjects, PSP patients presented reduced NAA/Cr or NAA/Cho in the brainstem, centrum semiovale, frontal and precentral cortex, and lentiform nucleus. These data demonstrate that, in addition to brainstem and basal ganglia involvement, the frontal association and motor cortex are also affected; abnormalities in the centrum semiovale suggest involvement of the axons connecting the basal gan-
glia to the cortex (51). In another PSP series, in which a decrease of NAA in the lentiform nucleus was found, the authors reported that three of their nine PSP patients presented the "eye of the tiger" sign in the pallida (52). A similar observation was made by other authors in a case of CBD (53), suggesting that the "eye of the tiger" sign described by Sethi et al. (54) is not specific of HallervordenSpatz disease (HSD). In 1.5 T T2-weighted images, the pallidum is normally hypointense in adults or old patients because of normal accumulation of iron. Any hyperintense area within the pallidum, because of lacunar infarct or dilatation of Virchow-Robin spaces or rarefaction of fibers and gliosis that probably accompany calcium deposits, may justify the appearance of the "eye of the tiger." In fact, contrary to what we expected, we have observed some hyperintensity in T2-weighted images in a few cases in which CT demonstrated calcifications in the pallidum, a nearly physiologic finding in old age. What is striking in children or adolescents with HSD, is that the hypointensity owing to iron deposits is extraordinarily marked for their age. Even if HSD may present in adulthood, the age of the patient should be considered when one sees an anteromedial hyperintensity in a hypointense pallidum. In other words, the hypointensity should be evaluated considering the patient's age.
HSD was recently found to be related to mutations in the gene encoding pantothenate kinase 2 (PANK2), and the specific MRI pattern described (54,55) has been found to distinguish patients with PANK2 mutations (56). The same MRI findings may be seen in HARP syndrome (hypoprebeta-lipoproteinemia, acanthocytosis, retinitis pigmentosa, and pallidal degeneration) which is, however, part of the pantothenate kinase-associated neurodegeneration (57). It will, therefore, be very interesting to carefully rule out any possible misinterpretation (i.e., lacunar infarcts or dilated perivascular spaces) and eventually obtain the appropriate genetic tests in the patients with PSP or other disorders presenting the "eye of the tiger" sign or similar signal abnormalities.
Was this article helpful?