Introduction

Parry-Romberg syndrome (PRS) is a rare disorder characterized by progressive hemifacial atrophy. The atrophy can involve the skin, underlying soft tissues, cartilage, muscle, and bone, typically in dermatomes of one or more branches of the fifth cranial nerve [2, 34]. In most patients, the disease is slowly progressive or self-limiting, with a female predilection and an onset in the first and second decades of life [35]. However, rare late onset cases have also been reported [4, 7, 34]. The diagnosis is largely based upon clinical findings, with no existing diagnostic criteria [34]. The proposed underlying disease mechanisms are diverse; the most relevant theories include autoimmunity, bacterial (i.e., Borrelia burgdorferi) or viral infections, trauma, and sympathetic nervous system dysfunction [6, 7, 23, 27, 30, 33, 42]. Of these, autoimmune origin is the most supported in the literature, as the disease frequently coexists with other autoimmune disorders such as vitiligo, systemic lupus erythematosus, generalized myopathy, and rheumatoid arthritis, and there frequently is systemic inflammation with or without the presence of a specific autoimmune antibodies [14, 15, 34]. In addition, immunosuppressive therapy has been successfully used in many cases of PRS during its acute phase [2, 4, 22, 33, 35, 43, 45]. Interestingly, a recent case report, where whole exome sequencing was performed on a PRS patient, found mutations in the MTOR and DHX37 genes, possibly related to the disease process [44].

About 10–22% [18, 32, 42] of PRS patients have associated neurological symptoms, including epilepsy, headache, trigeminal neuralgia, facial paresthesia/pain, hemiparesis, cognitive impairment, and cerebral infarction [4, 14, 15, 18, 27, 32, 35, 36, 42, 45]. For those patients with neurologic symptoms, abnormal findings on brain imaging are generally seen ipsilateral to the side of the hemifacial atrophy [42]. The mechanisms of these neurologic manifestations are unclear, with only sparse pathologic reports describing the presence of vascular abnormalities, inflammation, cerebral hemorrhage, and ischemia [1,2,3,4,5, 12, 16, 25, 29, 45].

Clinical summary

At the time of the surgical intervention that generated the pathologic findings that will be the focus of this case report, our patient was a 19-year-old woman with a clinical diagnosis of PRS, evidenced by right hemifacial atrophy most prominent in the forehead and ipsilateral right hemispheric abnormalities on brain imaging. She had a history of seizures starting at the age of 8 years, with progressively increasing seizure frequency over the subsequent years. The right hemifacial atrophy was described first when she was 9 years old. After menarche, the patient’s seizures assumed a catamenial pattern. Subsequently, she established care at our institution age 16. Her peripheral blood indices were normal without anemia or lymphocytosis, and her erythrocyte sedimentation rate (ESR) ranged from normal to slightly elevated (up to 24 mm/h; ref.: 0–20 mm/h) since under our care. Serological studies for autoantibodies were positive only for anti-SSA antibody.

Regarding her nonsurgical management, she received methotrexate and steroids in Mexico and was subsequently treated with intravenous immunoglobulins (IVIG) and rituximab by the rheumatology service. Additionally, a number of anti-epileptic medications have been trialed since her initial seizure manifestation, including zonisamide, clobazam, lamotrigine, levetiracetam, oxcarbazepine, phenytoin, and valproic acid — with the latter three not beneficial or were not tolerated by the patient. In spite of these, however, the patient’s seizures remained medically intractable.

Neuroimaging studies performed initially at our institution showed scattered foci of susceptibility hypointensity and contrast enhancement with surrounding white matter T2 hyperintensity, with subsequent imaging showing progression of the right hemispheric brain abnormalities on MRI with greater involvement of the right parietal lobe (Fig. 1). Additionally, MRS demonstrated a decrease of normal regional metabolites in the affected area (Fig. 2), and decreased uptake was noted in the right frontoparietal regions on FGD PET/CT (Fig. 3).

Fig. 1
figure 1

MRI. Axial T1 (a), axial T1 with contrast (b), axial FLAIR (c), and axial SWI (d) demonstrate thinning of the right frontal scalp and calvarium (arrows). There are underlying right frontal lobe susceptibility hypointense foci (d) with associated enhancement (b) and surrounding white matter FLAIR hyperintensity (c). Perfusion imaging shows corresponding changes in the right frontal lobe with decreased CBV (e) and CBF (f) as well as mildly delayed MTT (g) and TTP (h)

Full size image
Fig. 2
figure 2

MRS. MRS (short echo) with a voxel from the right frontal white matter (a) demonstrates decreased metabolites. A voxel through the normal contralateral left frontal white matter (b) is shown for comparison

Full size image
Fig. 3
figure 3

CT and PET/CT. Axial CT (a) demonstrates a calcific focus in the right frontal lobe with surrounding edema. PET CT (b) shows decreased activity in the right frontal lobe

Full size image

At the age of 17 years, she had a right frontal craniotomy for open biopsy, showing no evidence of an active inflammatory process (see pathology results in the “Pathologic findings”). At this time, she was taking clobazam (10 mg twice a day), lamotrigine (150 mg twice a day), levetiracetam (2000 mg twice a day), zonisamide (400 mg once a day at bedtime), clonazepam (0.5 mg twice a day, when needed), and diazepam (10 mg intranasal, when needed) for her epilepsy.

Two years later, at the age of 19, video electroencephalography (vEEG) studies demonstrated focal onset epilepsy with seizures originating from the right frontal/frontocentral regions prompting the recommendation to pursue a right frontal lobe seizure focus resection, which was guided by intraoperative electrocorticography. Intraoperative electrocorticography revealed interictal discharges in the right frontocentral region just anterior to the supplemental motor area, and this cortex demonstrating abnormal electrical activity was resected as an en bloc specimen. At the depth of this neocortical resection, the white matter was noted to be markedly more rubbery than usual. The dura overlying the resected cortex was also sent for pathologic analysis (see details in the “Pathological findings”).

The patient had a reduced seizure burden for about 3 months after surgery that was only transient, and her seizure burden returned to the preoperative level afterwards.

In light of recurrent seizures, a repeat vEEG admission was completed 6 months after surgery which again demonstrated focal seizures arising from the right frontocentral region. Given the patient’s desire to pursue additional surgical strategies to reduce her seizure burden, a phase 2 epilepsy monitoring surgery was performed in the form of stereoelectroencephalography (sEEG) lead placement to precisely localize active seizure foci 14 months after the resective surgery. Eleven right hemispheric sEEG electrodes were placed, and, over a 4-day monitoring period, the data from this study revealed that the patient’s typical seizures were arising from the right frontal SMA region with early involvement of the right superior parietal lobule region prior to generalization. Given that the right hemispheric lesional region included the right primary motor cortex, placement of a responsive neural stimulation (RNS) device (NeuroPace) was recommended for seizure palliation over consideration of further resective surgery. It is to note that neuromodulation devices, like RNS, are not curative; therefore, anti-epileptic medications are still indicated for seizure control along with them.

Sixteen months after resective surgery, the RNS device was placed — with RNS subdural strip electrodes carefully situated over the right frontal SMA and right superior parietal lobule regions most implicated by the sEEG study. Since implantation, the RNS device has successfully detected seizures and prevented many seizures from generalizing, although the patient still suffers from occasional breakthrough seizure events. Her current anti-epileptic medications are clobazam (25 mg once a day), levetiracetam (2000 mg twice a day), zonisamide (400 mg once a day at bedtime), and cenobamate (100 mg once a day). She continues to have frontal lobe focal aware auras of being anxious with increased heart rate and gaze deviation toward the contralateral side, improving from weekly to 0–2 seizures every 1–2 months. However, there has been an improvement of her debilitating focal to bilateral tonic-clonic seizures, from monthly to a 0–2 every 1–2 months, and the last seizure was 10 months before her 18-month follow-up from the RNS implantation. See Fig. 4 for representative RNS-captured seizures.

Fig. 4
figure 4

Representative RNS-captured seizures. Left: RNS, recording of a type 1 focal motor to bilateral tonic-clonic seizure (FBTCS). Description: She starts feeling anxious, she lose the stability of her eyes, she feels them jumping, versive eyes and head deviation to the left side, and she is described to have stiff left arm, she can still move the right arm, clonic movements of the left side of her face. Her awareness is intact until she developed bilateral tonic-clonic activity. Post-ictally, she has numbness in the left arm and left face with slurred speech. Right: RNS, recording of a type 2 focal aware seizure (FAS). Description: Feeling anxious, eyes may move around

Full size image

Pathological findings

Neuropathologic examination of the brain tissue from the frontal lobe open biopsy at age 17 showed two diminutive cavernous hemangiomas with microvascular and parenchymal calcifications and piloid gliosis in the background brain tissue, with unremarkable overlying dural tissues. On the other hand, the right frontal lobe tissue from the second procedure at age 19 revealed multifocal segmental active lymphocytic arteritis in the parenchymal arterioles (Fig. 5a, b). By immunohistochemistry, the majority of the lymphocytes in the foci of arteritis were immunoreactive for CD3 (Fig. 5c) and negative for PAX-5, confirming a primarily T-cell inflammatory infiltrate. The adjacent brain parenchyma showed variable neuronal loss, reactive gliosis, and meningeal and neocortical dystrophic calcifications (Fig. 5f). Moreover, the overlying leptomeninges and superficial microvasculature showed significant fibrosis, all consistent with chronic tissue injury (Fig. 5d, e). There was no evidence of amyloid deposition seen on Congo red special staining.

Fig. 5
figure 5

Histological findings of the resected brain tissue and dura. a, b H&E staining showing segmental middle-sized vessel vasculitis in the brain parenchyma. c The lymphocytes in the vessel wall are predominantly CD3-positive T lymphocytes. d H&E staining showing hyalinized, thickened vessel walls in the dura. e Trichrome stain shows significant meningeal vessel fibrosis. f H&E staining showing dystrophic calcifications in the brain tissue

Full size image

Discussion

The clinical presentation of PRS is characterized by skin and soft tissue atrophy that typically starts in the periorbital area and may also involve the forehead, lower face, and/or neck areas [42]. The disease typically has an active phase with progressive atrophy that usually abruptly arrests and stabilizes without any obvious reason, referred to as the “burned-out” phase [8, 42]. Seizures are the most common neurologic manifestation of PRS, occurring in ~60% of individuals with neurologic manifestations, and ~30% of patients that seize will go on to develop medically refractory epilepsy [39]. Symptom control is generally the focus of PRS management, involving pain management, anticonvulsants, and reconstructive surgery [4, 45]. In addition, there are reports of successful immunosuppressant therapies in the acute phase of PRS [22, 33, 35, 43], which fits well with the predominant theory of autoimmune cause for the disease. An interesting study showed that facial skin tissue/microsurgical soft tissue from the affected areas from PRS patients has a unique pro-inflammatory gene expression profile, further supporting the inflammatory cause of the disease [9].

Despite autoimmunity being the currently favored etiology of PRS, there are only a few published neuropathologic reports confirming the presence of inflammation, and prior to this current report, there were no reports of vasculitis in patients with PRS. Of note, given the rarity of the disease and the fact that most patients do not undergo neurosurgical resection, there are very few published reports of histopathologic tissue examination in patients with PRS. The previously reported cases show a wide variety of pathologic changes of the brain tissues, including diffuse neuronal loss and microglial nodules [5, 13]; cavernous hemangioma [2]; leptomeningeal fibrosis, degenerative cortical changes, and microvascular malformations [15, 38, 41]; capillary hypertrophy [27]; torturous meningeal midsized vessels with increased mural fibrosis and hyalinization [41]; cortical parenchymal lymphocytic inflammation with microglial nodules and gliosis compatible with Rasmussen encephalitis [29]; and nonspecific vascular changes [8]. Two prior reports described evidence of perivascular lymphocytic inflammation without vasculitis [25, 40], similar to the perivascular infiltrates occasionally seen in skin biopsy specimens from PRS patients [33, 43].

Our case of a PRS patient with medically intractable epilepsy is the first to demonstrate segmental active lymphocytic arteritis in a penetrating parenchymal vessel by histopathologic examination of resected brain tissue. In addition, there was evidence of chronic vascular injury, resulting in fibrosis of the meningeal vessels and leptomeningeal tissues, as well as neuronal loss, gliosis, and regional calcifications, similar to those described in prior pathologic reports. Moreover, our patient’s limited improvement in seizure burden following epilepsy surgery interventions is consistent with the pathological findings of an active inflammatory process. It is to note that the previously described histologic changes can be well-explained by localized vasculitis, leading to vascular wall fragility and subsequent vascular or brain tissue injury.

In patients with PRS, reports on brain imaging are more common than histological studies. The described changes are typically ipsilateral to the skin and soft tissue atrophy or are bilateral. These changes include cavernomas, aneurysms, hemorrhages/microhemorrhages, ischemic lesions, calcifications, cortical and white matter hyperintense lesions on T2-weighted sequences, meningeal enhancement, cortical thickening, and dysgenesis [1,2,3,4, 12, 15, 16, 25, 45]. Additionally, contrast-enhancing lesion, cerebral hypoperfusion, and hypometabolism have also been reported [1, 37]. Autoimmune tissue injury is in the neuroradiologic differential for these reported neuroradiologic abnormalities [25, 39]. In this case, there were scattered enhancing foci and calcifications in the brain parenchyma seen by MRI, with evidence of decreased FDG activity on PET scan in the affected area.

In the spectrum of autoimmune diseases, there is considerable overlap in the clinical manifestations of specific disease entities. Relevant to this discussion, there is a close relationship between PRS and linear scleroderma (en coup de sabre — if it is in the frontotemporal region [21]). Both diseases fall under the umbrella term localized craniofacial scleroderma [24], and the current literature treats these two entities either as a two spectrums of the same disease, overlapping diseases, or two completely different diseases ([31], Tollefson et al 2007, [17, 27, 43]). En coup de sabre is a form of linear scleroderma, where the fibrous plaque has a band-like distribution on the frontotemporal area of the skin, typically surrounded by scarring alopecia. The two conditions have many clinical features in common, such as young age of onset, female predilection, worsening during pregnancy, variably elevated antibody titers, and the presence of a skin lesion [43]. These are, however, non-specific and can be seen in many autoimmune disorders. More specific supporting evidence includes that the PRS skin lesion histologically is consistent with morphea; and that about half of the PRS patients have a skin lesion consistent with en coup de sabre [35]. In addition, approximately 40% of patients with linear scleroderma have hemifacial atrophy [35], and approximately 10% of linear scleroderma patients have associated neurological abnormalities, with epilepsy being the most common manifestation [10, 27]. The pathogenesis of localized scleroderma is unknown, although an inflammatory process, particularly vasculitis, is suspected [10, 34], supported by evidence of linear scleroderma cases improving with immunosuppressant therapy [31].

Central nervous system histologic findings in patients with ECDS are similar to the ones in PRS and include the following: perivascular and/or parenchymal lymphocytic infiltrate [5, 19, 26, 31], gliosis [11, 26], leptomeningeal sclerosis [11, 26], intraparenchymal and intravascular calcifications [11], and sclerotic and/or ectatic blood vessels [11, 19]. Abnormal ectatic cerebral vessels have also been seen on imaging in patients with linear scleroderma/ECDS [10, 27]. In addition, intraparenchymal calcifications [19, 28], microhemorrhages [24], cavernomas [2, 16], mild cerebral atrophy [43], and white matter lesions [28], enhancing T2-hyperintense lesions [24], along with further radiologic signs of cerebral vasculitis [20] have also been described, which are very similar to those findings in PRS.

In conclusion, our report is the first to demonstrate active cerebral vasculitis in a patient with PRS, supporting the previous suspicion of inflammatory etiology in this disease. In addition, the widespread vascular fibrosis in the meningeal vessels and focal cortical ischemia support the presence of previous inflammatory vascular processes in the area. These findings add to the sparse pathologic descriptions of central nervous system tissue in patients with PRS and lend support to the hypothesized mechanism of inflammatory/autoimmune tissue injury and treatment with immune modulation.