Brain Abnormalities and Epilepsy in Patients with Parry-Romberg Syndrome

DISCUSSION

PRS is clinically suspected with facial asymmetry or hemiatrophy, often in association with neurologic, ophthalmologic, and orodental manifestations. The working diagnosis is made by clinically excluding mimickers such as localized scleroderma, Barraquer-Simons lipodystrophy, fat necrosis, contralateral hemihypertrophy, and congenital hemiatrophy. Rarely, laboratory testing or histopathology is performed to establish the diagnosis. Brain abnormalities are often seen in association with PRS, yet the frequency and severity have received limited attention. In this study, 57% of patients with a working or established clinical diagnosis of PRS and available MR imaging had at least 1 primary brain abnormality associated with PRS. The most common finding was WMD, which was observed asymmetrically in 29% of the cohort. Notably, unilateral-predominant brain abnormalities localized to the side with hemifacial atrophy were found in one-third of patients; due to the atrophic nature of the disorder, these are likely related to the pathophysiology of PRS.

Similar to our results, frequent imaging findings reported in other studies include ipsilateral white matter hyperintense signal on T2-weighted MR imaging corresponding to white matter hypoattenuation on CT, hemispheric brain atrophy, cavernous malformations/microhemorrhages, and subcortical calcifications predominantly noted in the frontal lobe.^{2,13,16⇓⇓⇓⇓-21} We found that the primary brain findings associated with PRS were more common in patients with epilepsy compared with those without epilepsy (94% versus 47%), as well as more commonly present on the side of craniofacial involvement in patients with epilepsy compared with those without epilepsy (81% versus 20%). When present, the degree of hemispheric atrophy, WMD, and microhemorrhage was more severe in those with epilepsy. In addition, 5 patients, including 2 without epilepsy, had remarkable asymmetric sulcal effacement located primarily in the frontoparietal region, a finding that has been rarely reported by other investigators.^10,22⇓-24 Due to the low frequency of other findings, including bilateral frontal horn heterotopia, supraclinoid aneurysms, and small cortical and subcortical infarcts, it remains unclear whether these could be related to PRS.

It is often challenging to characterize the disease course, severity, and prognosis of PRS due to these overlapping features with ECDS or localized scleroderma.³ Currently, the most acceptable etiology associates PRS with an underlying progressive autoimmune neurovasculitis, though other theories have been proposed.^11,21,25,26 Cerebral perivascular inflammation, leptomeningeal and cortical blood vessel endothelial degeneration, hyalinizing changes, and partial obliteration have been reported in PRS, possibly as a consequence of immune-mediated vascular injury and incomplete endothelial regeneration.²⁷ These changes probably account for ischemia, hemorrhages, and breakdown of the blood-brain barrier. Robust response to steroids and immune-modulating treatment in slowing the disease process also supports this etiology, in addition to histologic and laboratory evidence of inflammation.^1,2 Brain abnormalities in PRS have overlap with other conditions, such as Rasmussen encephalitis, an autoimmune disorder of childhood characterized by asymmetric neuroimaging findings, epilepsy, and hemiplegia.⁴ Characteristic craniofacial abnormalities in PRS may help differentiate the disorder from other entities.

Epilepsy can be a debilitating comorbidity in patients with PRS and further compromise quality of life, and it was found in 20% of patients in our series. A global, Internet-based survey of patients with PRS from 29 countries revealed that 11% had coexistent epilepsy,²⁶ with the highest reported epilepsy incidence of 60.5% in a pooled literature review of patients with PRS.²⁸ In a pediatric study of PRS with epilepsy, the median seizure onset was in the first decade.²⁹ In our study, the age of seizure onset was often estimated; however, most patients had onset in childhood. In our cohort, the most common type of seizure was focal-to-bilateral tonic-clonic seizures, followed by focal with impaired awareness. Most cases (81%) of epilepsy had an abnormal EEG finding, with 50% demonstrating interictal epileptiform discharges. On the basis of our observations, the severity of brain abnormalities appears to be a risk factor for development of epilepsy. However, the widespread extent of the abnormalities limits the localizing ability of MR imaging in identifying the seizure-onset zone.³⁰ Nevertheless, it appears that certain MR imaging features predispose to a higher likelihood of developing epilepsy in PRS. Similar to a recent study by Knights et al,³¹ our cohort showed a higher frequency and severity of ipsilateral WMD in epilepsy. Their study, however, failed to show a significantly greater frequency of hemorrhagic foci in epilepsy in contradiction to our findings. Last, abnormal MR imaging findings were more frequent in patients with epilepsy in their cohort, but this was not statistically significant.³¹ Both of these differences between studies may be related to the small number of subjects in their cohort.³¹

Currently, limited evidence exists regarding the role and findings of advanced brain imaging in PRS and epilepsy. Few patients in our cohort had such imaging performed, but findings were largely in agreement with prior reports.^22,32 Hypoperfusion of the affected brain region has been shown in several prior studies using dynamic susceptibility contrast MR imaging or SPECT,^19,22,33 which is in agreement with arterial spin-labeling results from 1 patient in our cohort. In 1 patient, a large region of hypometabolism was observed in the affected region on FDG-PET, which has also been previously reported.³⁴ Last, few studies have shown the findings on ¹H-MR spectroscopy in PRS. While findings have been mixed, most have shown a relative decrease in the Cho:Cr ratio and a decrease in NAA, similar to findings in 1 patient in our cohort.^4,19,22,32

The descriptive findings in this study provide further insight into the need for brain imaging in patients with PRS and neurologic symptoms. Recognition of the frequency and spectrum of brain abnormalities in PRS is critical because the diagnosis was delayed for years and the opportunity for timely management was missed in many of our patients. Further longitudinal studies are needed to better understand the natural history and radiographic progression to diagnose and treat patients with PRS and epilepsy more optimally. Similarly, larger prospective studies are recommended to validate the prevalence of epilepsy and the response to therapy in patients with PRS.

There are several limitations to this study. The retrospective nature leads to selection bias, with heterogeneity and gaps in clinical documentation based on chart review. Nevertheless, the number of patients is relatively large and provides additional details of the association of epilepsy with PRS. The lack of established diagnostic criteria for PRS and frequent overlap with other entities such as ECDS could have led to challenges in establishing a definitive classification of progressive hemifacial atrophy. Referral to a tertiary care center likely overestimates the true incidence of a disease state and severity of associated abnormalities and, therefore, may not be generalizable to other populations.