Case report

A 39-year-old, previously well, left-handed man presented in 1989 with generalised tonic-clonic epileptic seizures. These were followed by intermittent left-sided facial simple partial motor seizures, complex partial seizures and a progressive left hemiparesis. His response to anti-epileptic medication was poor. In 1990, brain MRI revealed increased signal around the right central sulcus on T2-weighted images. He had right frontal corticectomy and biopsy in 1990. Histology of the specimen showed persistent localised inflammation, perivascular lymphocytic infiltrates, astrogliosis, microglial nodules in grey and white matter, immune complexes in the affected brain tissue and neuronal loss with gliosis. These are changes typical of RE.

Although there was an initial mild improvement after surgery, his epilepsy and hemiparesis again worsened. Over the next few years he also developed progressive dysphasia, cognitive decline and a left homonymous hemianopia. He had right frontoparietal transection in 1995.

By June 1996 his seizures and hemiparesis were incapacitating. On MRI, there was increased T2 signal from most of the right frontal and parietal lobes associated with atrophy. Electroencephalography showed frequent right frontoparietal sharp wave discharges. Cerebrospinal fluid (CSF) contained 5 lymphocytes/mm³. On semiquantitative interictal brain perfusion single photon emission computed tomography (SPECT) scan using ⁹⁹Tc^mHMPAO (hexamethylene propylene amine oxime), there was markedly reduced perfusion of most of the right cerebral hemisphere (figure). In addition there were several contralateral focal perfusion defects.

He started treatment with prednisolone and hIVIg (0.4 g/kg/day for 5 days). In total, he had six high-dose cycles of hIVIg at intervals of 4–6 weeks. He then switched to monthly maintenance hIVIg infusions (0.4 g/kg/day).5

Around 4 months into treatment, he began to improve. His tonic-clonic seizures stopped and simple partial seizures reduced in severity and frequency. His hemiparesis recovered progressively, and he became independently mobile.

A second interictal brain perfusion SPECT scan was performed 8 months into treatment. We used spatial registration of both sets of SPECT images (MultiModality software, Nuclear Diagnostics, UK) to compare the scans. Visually, the registered images show improvement in perfusion to right frontal, temporal and parietal lobes (figure).

We drew regions of interest over the four lobes of the cerebral hemispheres: frontal, temporal, parietal and occipital lobes. Uptake in each of these regions was compared with ipsilateral cerebellar uptake, which served as an internal reference. Occipital ratios are used when either or both cerebellar lobes have reduced perfusion. The lobar to cerebellar uptake ratios were compared with similar ratios calculated from normal data (summed multiple slices) obtained from SPECT scans of 22 healthy volunteers. Variation from this normal database is denoted by the number of standard deviations (SD). Since 95% of values lie within 2 SD either side of the mean for the normal distribution, SD values of 2 or less are associated with a p-value of <0.05.

In our patient, pre-treatment semiquantitative SPECT regional cerebral blood flow (rCBF) data revealed a significant (>2 SD, p < 0.05) reduction in perfusion to right frontal, temporal and parietal lobes and the left frontal lobe. Occipital ratios were used because the patient's right cerebellum showed reduced perfusion compared to the left cerebellum. Comparison of post-therapy values after standardisation revealed significantly improved perfusion to the right frontal (SD 6.65, p< 0.005) and right temporal (SD 3.53, p< 0.005), and to a lesser extent to the right parietal lobe (SD 1.73, p>0.05) (table). Eight months into treatment, CSF was normal. Serial brain MRI showed no change.

Discussion

In our patient, pre-treatment interictal brain SPECT scan showed perfusion defects predominantly in the clinically affected and surgically resected areas of the right cerebral hemisphere. The visual interpretation of hypoperfusion was confirmed by semiquantitative analysis of the rCBF data. Several methods have been developed to increase the sensitivity and reduce the variability of subjective visual reporting of SPECT.6 These techniques increase the reproducibility of reporting by comparing patient rCBF data with those from healthy control subjects. Moreover, serial independent data sets from a patient, here pre- and post-treatment, can be compared indirectly by comparing each data set independently with the same control data.

Perfusion brain SPECT is known to be a useful method to show functional changes associated with RE and many other types of encephalitis and encephalopathy.7-10 The usual findings are focal or multifocal perfusion defects. Typically, in RE, there is decreased perfusion covering confluent and extensive areas of the frontal, temporal and parietal lobes of one cerebral hemisphere.10 While this pattern was seen in our patient, an area of hypoperfusion was also noted in the clinically and MRI normal contralateral cerebral hemisphere. We do not have an explanation for this contralateral perfusion defect. Previous reports of bilateral perfusion SPECT defects in RE patients were ascribed to crossed diasthesis,10 although patients with RE affecting both cerebral hemispheres have been described.4 In other more widespread forms of encephalitis, however, multiple focal perfusion defects are commonly seen bilaterally.7 Thus, the pattern of decrease in rCBF is largely determined by the underlying pathology. By analogy, perhaps the contralateral changes in our patient represented subclinical disease.

This is the first demonstration of improvement in rCBF to the clinically affected brain area in RE in response to therapy. Reperfusion also occurred in regions of previous surgery. Moreover, increases in blood flow were accompanied by the suppression of seizures and striking recovery of functions associated with the right frontal, temporal and parietal lobes. These findings suggest that the potential for recovery persists even in patients with advanced RE and established brain destruction. Brain perfusion defects on SPECT can be produced by several pathogenic mechanisms including direct cellular damage caused by viruses, arteritis causing vessel occlusion,8 and toxic effects of enzymes and cytokines released from damaged brain cells and inflammatory cells.9 RE is an antibody-mediated autoimmune disease characterised by inflammatory infiltrates containing activated lymphocytes, macrophages and glial cells. In our patient, improvement in regional cerebral perfusion was accompanied by resolution of CSF lymphocytosis suggesting that increased perfusion resulted from successful immunomodulation.

We conclude that serial interictal perfusion brain SPECT scanning and semiquantification is a useful and objective method to monitor and validate the response to immunomodulatory treatment in patients with RE. It is a more sensitive method than structural brain imaging, which often fails to demonstrate any improvement in established destructive disease and is particularly difficult to interpret in a patient who has had extensive surgery.