Hashimoto encephalopathy (HE) is associated with various mental disorders, the presence of antithyroid antibody, and has a clinical response to steroid administration.1 It is consistent with reversible cerebral inflammation that is probably mediated by an autoimmune mechanism.2 HE brain lesions can be visualized in different locations by conventional MRI as hyperintensities on fluid-attenuated inversion recovery (FLAIR) sequence and nonenhancing abnormalities on T1WI. A change during corticosteroid therapy can also be noted both in clinical manifestations and MRI appearance.3 Previously, lesions have not been characterized in terms of relevant biochemical change using magnetic resonance spectroscopy (MRS). We describe MRS abnormalities in a female HE patient which correlate with histopathologic findings, and propose that MRS is useful in identifying cerebral inflammation changes of HE.
A 52-year-old woman presented at the hospital with low-grade fever, headache, disorientation, amnesia, bad response to communication, numbness in the right hand, blurred vision in the right eye and tonic-clonic seizures in the previous two weeks. Her previous neurological history was unremarkable. Physical examination revealed horizontal nystagmus, bilateral Babinski signs, 4/5 of limb power, and poor cooperation in mental status examinations. EEG demonstrated a diffuse slow wave with intermediate to high amplitude.
Lumbar puncture results showed normal cerebrospinal fluid (CSF) pressure, normal cell counts, and normal protein levels. Tests for microbial infection were negative. The levels of C-reactive protein, serum immunoglobulin G, and complement C4 were slightly beyond the upper limit. Serum tumor markers, anti-encephalic and antinuclear antibodies were negative.
Thyroid function tests revealed a decreased free T3 of 2.04 pg/ml (reference 2.5-3.9 pg/ml), thyroglobulin antibody >500 U/ml (reference 0-60 U/ml), and thyroid peroxidase antibody >1300 U/ml (reference 0-60 U/ml). The levels of T3, T4, free T4, thyroid stimulating hormone (TSH) and TSH receptor antibody were normal.
Brain MRI on admission showed multiple patchy abnormal hyperintensities on T2 weighted imaging (T2WI) (Figure 1A), homogeneous hypointensities on T1WI in the bilateral cerebral hemispheres, and slight hyperintensities were visualized on diffusion weighted imaging (DWI) (Figure 1B). No unusual enhancement was seen on gadolinium-enhanced T1 sequence. Mass effect and midline shift were absent. Because short echotime (TE) acquisition could include more metabolites with shorter T2 decays than a long TE, a TE of 35 ms was selected here for a single voxel proton MRS examination. The MRS in the left diseased temporal lobe showed that three major resonance peaks (N-acetylaspartate, creatine, and myo-inositol) identified by chemical shift had lost the steep angle up from left to right, which should exist in a normal spectrum. Also, a lipid (Lip) peak and lactate (Lac) doublet peak appeared in the normally flat region of the spectrum on the right. The ratio of N-acetylaspartate (NAA) to creatine (Cr) was 1.19 (normal value was expected to be about 1.4), whereas the ratio of choline (Cho) to creatine (Cr) was 1.21 (normal value was expected to be about 0.8). The reduced myo-inositol (mI) peak made the elevated glutamine/glutamate multiplet (Glx) peaks more visible (Figure 2).
An incisional biopsy in the left temporal lobe was performed. A piece of diseased brain parenchyma with a diameter of 2 cm was obtained. Microscopically, a few sparse foci of cerebromalacia (liquefaction and necrosis) were identified. Furthermore, lymphocytic infiltration of venules was seen with degeneration of neurons, accompanied by moderate gliosis in the background (Figure 3), and there was no evidence of tumor. Electron microscopy demonstrated swelling and degenerative mitochondria in the neuron, microglia hyperplasia in the grey matter, as well as demyelination, axonal vacuolization, and some cystic spaces in the white matter. No virus inclusion bodies were found.
In this case, infections and cancers were ruled out by CSF analysis, immunological screening, MRI, and the pathological results from biopsy. Mitochondrial encephalomyopathy was eliminated by relevant genetic testing and trial treatment. The clinical evidences did not support demyelinating diseases, such as multiple sclerosis or acute disseminated encephalomyelitis. A clinical diagnosis of HE was established based on positive antithyroid antibody. She was treated with intravenous methylprednisolone for 9 days (maximum dose 500 mg/d), followed by an oral taper. The patient’s status improved at first, and then relapsed intermittently. Two years later, a follow-up MRI indicated that previous abnormal signals resolved and there was a remarkable residual cerebral atrophy with degeneration of the white matter in the left occipitotemporal lobes (Figure 1C).
Since the first description of Hashimoto encephalopathy in 1966, there have been approximately 100 reported cases involving a series of neurological clinical manifestations.1 Although the controversy about the diagnosis of HE still exists,4 some studies suggest that HE is consistent with a reversible cerebral inflammation probably mediated by autoimmune mechanisms.2 However, relevant imaging findings in the past were inadequate. While the nonenhancing lesion on T1WI suggests neither angiogenesis nor blood-brain barrier damage, relevant inflammatory changes on MRS and the correlation between MRS and pathologic findings in the diseased area have not been documented.
In addition to conventional MRI, MRS is useful in analyzing pathological changes. The decrease of NAA indicates neuronal and axonal damages, supported here by biopsy. Usually, the elevation of the choline peak is associated with an aggressive neoplastic process. In this case, slight elevation was most likely caused by a benign gliosis and demyelination.5
The elevated lactate and Glx peaks, especially the former, result from excess glycolysis, hinting at hypoxia. Elevation of Glx is not typically seen in aggressive intra-axial neoplasm, but is more suggestive of an inflammatory process.6 The rise of the lipid peak, generally a nonspecific metabolite that can be noted in many destructive processes in the brain, is secondary to the release of free lipids from cell membranes.6 Since mI is a substrate for some cellular messengers that participates in antioxidation and stabilizing osmotic pressure,7 the almost depleted mI in this patient reveals a deprivation of these functions to some extent. Generally, these changes of metabolites in MRS are consistent with an inflammatory and hypoxia process, including neuronal degeneration, gliosis, and demyelination.8
Because there is as yet no clear pathogenic description of HE, the diagnosis of HE is based on excluding other more common diagnoses, along with elevated antithyroid antibodies and a good response to steroid administration. We thought that MRS results could strengthen the diagnosis by providing metabolic insight.
Our conventional MRI findings shared some characteristics with a previous report in multifocal high signal intensity lesions on T2-weighted and FLAIR images, as well as atrophy of cerebral parenchyma,3 but the lesions in this patient progressed faster. Moreover, the findings of slight hyperintensities on DWI and no unusual enhancement on gadolinium-enhanced T1-weighted images are consistent with a prior report.9 Pathological findings are nonspecific for HE, so we consider that MRS could be used as a means of making a diagnosis. Because MRS can be used to monitor therapeutic efficacy in some other diseases,10 it may be that MRS will prove useful as an early indicator of metabolic response to steroid treatment in HE.
1. Chong JY, Rowland LP, Utiger RD. Hashimoto encephalopathy
: syndrome or myth? Arch Neurol 2003; 60: 164-171.
2. Seo SW, Lee BI, Lee JD, Park SA, Kim KS, Kim SH, et al. Thyrotoxic autoimmune encephalopathy: a repeat positron emission tomography study. J Neurol Neurosurg Psychiatry 2003; 74: 504-506.
3. Song YM, Seo DW, Chang GY. MR findings in Hashimoto encephalopathy
. Am J Neuroradiol 2004; 25: 807-808.
4. Schiess N, Pardo CA. Hashimoto’s encephalopathy. Ann N Y Acad Sci 2008; 1142: 254-265.
5. Sheng B, Lau KK, Li HL, Cheng LF. A case of Hashimoto’s encephalopathy with demyelinating peripheral neuropathy. Eur Neurol 2005; 53: 84-85.
6. Panchal NJ, Niku S, Imbesi SG. Lymphocytic vasculitis mimicking aggressive multifocal cerebral neoplasm: MR imaging and MR spectroscopic appearance. Am J Neuroradiol 2005; 26: 642-645.
7. Ross B, Bluml S. Magnetic resonance spectroscopy
of the human brain. Anat Rec 2001; 265: 54-84.
8. Zhao WQ, Li JM, Wang JW, Tuo HJ, Kang ZM, Jiang B, et al. Clinical, imaging and pathological features of Hashimoto’s encephalopathy (report of 1 case). J Clin Neurol 2010; 23: 107-109.
9. Grommes C, Griffin C, Downes KA, Lerner AJ. Steroid-responsive encephalopathy associated with autoimmune thyroiditis presenting with diffusion MR imaging changes. Am J Neuroradiol 2008; 29: 1550-1551.
10. Kaminaga T, Shirai K. Radiation-induced brain metabolic changes in the acute and early delayed phase detected with quantitative proton magnetic resonance spectroscopy
. J Comput Assist Tomogr 2005; 29: 293-297.