Background
Juvenile xanthogranuloma (JXG) is a rare ‘histiocytic’ disorder and belongs to the broad group of non-Langerhans cell histiocytosis [
1]. As noted in a report of this condition by Helwig and Hackney in 1954, Rudolf Virchow was the first to describe a child with cutaneous xanthomas in 1871 [
2]. Other early reports of JXG were published in 1905 by Adamson [
3] and in 1912 by McDonagh [
4]. The real incidence is unknown, but it may be higher than is generally appreciated, because JXG is often underdiagnosed, in particular in people with dark skin. In the Kiel Paediatric Tumor Registry spanning 35 years JXG accounted for 129 (0.5%) out of 24.600 paediatric lesions. It is predominantly a disease of infancy or early childhood with a median age of onset between 5 months and 1 year [
5], but congenital-type juvenile xanthogranuloma is also reported [
6]. More males are affected than females, with a ratio of 1.4:1. JXG may affect all ethnicities, but few black patients with JXG have been reported [
7]. Pathogenesis of JXG has not been uncovered, however it is most likely a reactive and not a neoplastic process. Kitchen et al. assumed a disordered macrophage response resulting from a nonspecific injury [
8].
A triple association of juvenile xanthogranuloma, neurofibromatosis Type I (NF1) and juvenile myelomonocytic leukaemia (JMML) is often reported, but is the subject of frequent debate. In 2004 Burgdorf and Zelger analysed the literature and all available information pertaining to the association and found that patients with NF1 are, indeed, at an increased risk for developing JMML and JXG, but that the triple association of these findings (assuming the worst odds) is < 1% per year. However, regardless of the presence of JXG, children with NF1 are at a 200 to 500-fold greater risk for this hematologic malignancy. With regard to these rare events, lesions of JXG and NF1 may sometimes be clinically very similar and difficult to differentiate without histology. Moreover, lesions of JXGs and skin infiltrates of JMML may sometimes also be difficult to differentiate, clinically as well as histologically, all of which has significant influence on these statistical considerations [
9]. There are also limited reports of the coexistence of JXG and cytomegalovirus infection [
10].
Histopathology, clinical presentation and prognoses show great diversity. The presumed cell of origin of cutaneous JXG is a macrophage, derived in skin from the dermal dendrocyte, which represents a mixed dermal infiltrate of mononuclear cells, multinucleated giant cells and spindle cells [
11]. Immunostaining is important in establishing the diagnosis: JXG stains positive for factor XIIIa, CD68, CD163, CD14 and fascin and is mostly negative for S100 protein and regularly negative for CD1a and anti-langerin (CD207), which are specific for Langerhans cells [
12]. The typical clinical feature is a solitary, reddish or yellowish-tanned papule, plaque or nodule with a size of 0.5–2 cm, which generally appears on the head, neck, or trunk. Nevertheless, lesions can occur at any location in the body including lung, liver, spleen, lymph nodes, gastrointestinal tract, heart, kidney, bone marrow and central nervous system [
13]. Also eye involvement is described [
14]. For skin lesions, spontaneous regression within 1 to 5 years is the rule and treatment is rarely required [
15]. JXG with systemic (extracutaneous) involvement is an uncommon disorder in which significant morbidity and occasional death may occur. Implications for the patient’s condition depend on the degree of visceral dysfunction from the benign mass. Therefore, therapy must be initiated when JXG interferes with vital organ functions. Various treatment strategies including chemotherapy (LCH-III protocol, a Langerhans cell disease-based regimen including corticosteroids and vinca alkaloids) [
16], surgical resection [
17] and radiation are reported.
To illustrate the various spectra of JXG we present two completely different cases, the way to reach a diagnosis, the clinical course, treatment and differential diagnoses of both cases.
Discussion and conclusions
General
Cutaneous juvenile xanthogranuloma is a common ‘histiocytic’ disorder, but a detailed review of the literature reveals only a small number of cases of systemic juvenile xanthogranulomatosis in the neonatal period [
18] and less than 15 cases of spinal JXG [
19]. Although cutaneous JXG is generally regarded as a self-limited condition, systemic JXG may be associated with significant morbidity and occasional deaths so that aggressive medical care is necessary [
20]. To illustrate this point, we report on two affected children, both born within 1 year in Austria, who were confirmed to have JXG.
The originality of our observation is the clinically atypical and completely different presentation of this rare disease by the multi-lesional and multisystemic nature of its pathology. Moreover, it illustrates the difficulty of classifying this disorder, because the clinical and radiological presentation is nonspecific. Therefore, correlation with histopathology is mandatory and the gold standard for diagnosis of JXG.
Clinical spectrum
In the first patient we describe cutaneous JXG, which follows a benign course and gradual regression of the lesion without treatment. The diagnosis was established quickly, although the skin lesions were not typical of JXG. The typical presentation is a solitary erythematous or yellowish, well-circumscribed skin papule on the head, neck or trunk. Our patient presented with blueberry-muffin spots. Excision biopsy of the lesions was performed and established the JXG diagnosis. The absence of the typical yellowish colour was due to the lack of xanthomatization because of lesion immaturity. Thus, this case together with four more case reports in the literature [
21‐
24] indicates that the diagnosis of JXG should be included in the differential diagnosis of clinical presentation of a blueberry muffin baby.
With the second patient we report on one of the few documented cases (fewer than 45) of congenital systemic JXG [
25], presenting with a reduced general condition, a mass on the temple, fever, weight loss, and discrete skin involvement. Because of typical lesions in MRI (lesion in the skull and vertebra plana) and difficulties obtaining a usable biopsy for adequate histopathological analysis, the diagnosis of systemic JXG was delayed for several weeks. Despite the fact that fatal cases of systemic JXG - particularly central nervous system and hepatic involvement - have been reported only rarely [
26‐
29], prompt diagnosis and treatment are essential.
Imaging
In accordance with other reports [
30,
31] diagnostic work-up with ultrasound showed a well-defined, homogeneous, hypoechoic lesion without demonstrable blood flow in the dermis (Patient 1) or viscera (Patient 2) in both patients [
30,
31].
Magnetic resonance imaging (1.5 T) demonstrated the broad extension of the disease. In the literature, enhancement is described as a reliable feature of JXG lesions [
32]. The typical imaging ranges from iso- to hyperintense on T1 and iso- to hypointense on T2 [
31,
33]. MRI findings in our Patient 2 showed the big thoracic tumor conglomerate on T1 and on T2 slightly hyperintense to muscle, furthermore multiple nodular lesions in the liver, hyperintense in TIRM and T2 and hypointense in T1-weighted sequences. MRI imaging is nonspecific and variable. However, it is the first option for localizing the lesion.
Cytogenetics
The molecular cytogenetic findings in Patient 2 with systemic JXG showed 9p-(ptercen), 9p-(p21.3p21.1) and 9q rearrangements (9q33.3qter) positive, which could be a possible chromothripsis region involved in cancer and congenital diseases. The MYCN oncogene presented no indication for an amplification (2p/MYCN-negative). To date little is known about the genetic profile of juvenile xanthogranuloma. However, previous studies have reported that systemic JXG showed multiple genomic alterations, while solitary JXG usually has normal genomic profiles [
34].
Histopathologic features
Because of its typical clinical appearance, diagnosis of JXG is established clinically in most cases. However, its heterogeneous appearance may cause misdiagnosis. To confirm the clinical findings, skin biopsy for histology and immunostaining is essential. However, even this does not always provide a clear result, because more than 100 different subtypes of histiocytosis with a wide range of histological and immunohistochemical presentation have been described.
Classic histology of JXG shows a dense, sheet-like, noncapsulated, well demarcated cell infiltration in the dermis and the upper portion of the subcutaneous fat, while the epidermis and adnexal skin structures are spared. Cellular infiltrate includes five main cell types (vacuolated, xanthomatized, spindle-shaped, scalloped and oncocytic) in variable proportions (from monomorphous to mixed variants) with different types of giant cells (nonspecific, foreign body, Touton and “ground-glass”). Appearance mostly depends on the age of the lesion: while early lesions show a monomorphic infiltrate of lipid-free macrophages that can occupy most of the dermis, mature lesions contain abundant vacuolated, foamy macrophages and Touton-type multinucleated giant cells, particularly in the superficial dermis. Immunohistochemically, JXG lesions typically stain positive with macrophage markers including CD68, CD163, KiM1P, anti-FXIIIa, vimentin and anti-CD4 and usually are negative for S-100 protein and regularly negative for CD1a and CD207 (anti-langerin), which is specific for Langerhans cells [
35].
In Patient 1 the lesion showed a diffuse infiltration of epithelioid cells, sparing the papillary dermis and periadnexal connective tissue. There were monomorphic vacuolated cells without cellular atypia or increased or atypical mitoses. The immunohistochemical findings (Fig.
2c) were negative for mast cell and Langerhans cell markers: S-100 protein, CD1a, CD207 (anti-langerin), toluidine blue histochemistry, c-kit (CD117). The markers for macrophages CD68 and CD163 exhibited significant reactivity.
In Patient 2 the diagnosis was much more difficult and required three biopsies for histological and immunohistochemical work-up - including a referral report - to get the correct diagnosis. The first biopsy, a skin punch, showed eosinophils with strong mitotic activity. Immunohistochemistry showed S-100 protein and CD99 positivity, while CD1a stained negative, typical for a neoplasia of the Ewing/PNET group. The second skin biopsy from the soft tissue lesion on the infant’s left temple was sent to a reference centre and showed sheets of foamy macrophages admixed with mononuclear cells and numerous multinucleated giant cells. There were admixed lymphocytes and neutrophils, and a very prominent stromal haemosiderin deposition. So-called xanthosiderohistiocytosis was regarded as a morphologic variant of xanthoma disseminatum. Small areas consisted of the more monomorphic mononuclear cells similar to those seen in the initial skin biopsy. There was no atypia or pleomorphism and mitoses were scarce. Immunostaining showed strong and diffuse positivity for CD163, while S-100 protein was negative. It was labelled as an unclassified benign xanthogranulomatous lesion. However, the appearances did not match well with that of a conventional juvenile xanthogranulomatous lesion, so we performed another – computed tomography – assisted - biopsy of the mass in the posterior mediastinum showing cellular infiltrates of foamy macrophages with prominent nucleoli and eosinophilic granulocytes. Immunohistochemical work-up demonstrated a homogeneous and intensive CD68 and CD163 positivity, while NSE and CD99 showed nonspecific reaction patterns. CD207 (anti-langerin) and CD1a as well as HMB-45 remained negative. S-100 protein showed isolated dendritic background cells; otherwise it remained mostly negative, except for a non-specific reaction in the macrophages. Thus, definitive diagnosis was xanthogranuloma or xanthogranulomatous reaction.
ALK immunoreactivity was observed in a novel type of systemic histiocytic proliferative disorder that may suggest a storage disorder and should be a possible marker for systemic involvement with xanthogranulomas [
36]. We performed ALK immunostaining in our cases, which, however, was negative in both patients, so that we could not confirm the previous study [
36] suggesting that ALK might be a marker for systemic involvement.
Differential diagnoses
In Patient 1 the main symptom was the blueberry muffin-type rash, which is a potentially life-threatening condition with severe sequelae requiring extensive and prompt diagnostic work-up. Differential diagnoses can be divided into several broad categories: the first category includes haematological and non-haematological malignancies. Especially the differential diagnosis between JXG, in particular the Shapiro variant which is seen in this case, and cutaneous manifestations of JMML can be tricky and difficult to differentiate. The isolated myelosarcoma of skin in childhood is a rare manifestation of acute myeloid leukaemia preceding bone marrow involvement by weeks to months. Case reports in the literature describing the clinical presentation as blueberry muffin spots or symptoms of infection and anaemia are rare [
37]. Histologically, most cases are classified as monoblastic or myelomonocytic leukaemia with atypical mitoses. Immunohistochemically, CD43 and lysozyme stain a large proportion of neoplastic cells, with MPO and CD117 being the most sensitive of markers for myeloid differentiation, while monocytic precursors consistently strongly express CD68 and CD163 [
38]. Due to the small number of cases available for isolated myelosarcoma in children, prognostic statements are difficult. Spontaneous remission of congenital myelosarcoma is reported; however, the majority of cases progressed to AML within months. Taking into account the course of the disease in older patients, one could speculate that the prognosis is rather unfavourable. In synopsis of all findings, the benign clinical course of Patient 1 (at the age of 10 months the patient was in complete remission and after 3 years there is still no evidence of disease), the unremarkable laboratory findings (normal blood counts), the imaging (well-defined, homogeneous, hypoechoic lesion without vascularity), the histological (sparing of papillary dermis and periadnexal connective tissue as seen in our case, missing presence and number of (atypical) mitoses, low proliferation index with Ki-67) and immunohistochemical findings (positive for macrophage markers CD68 and CD163) the JXG diagnosis seems confirmed and valid. The second category includes congenital infections. However, TORCH work-up was negative in our patient. Finally, the third group includes extramedullary haematopoiesis in severe fetal and neonatal anaemia of any cause, but there was no evidence of a haemolytic disease like AB0 or Rh incompatibility or hereditary spherocytosis.
In Patient 2, histological and immunohistochemical findings were a little deceptive. JXG is mostly immunohistochemically negative for S-100 protein. However, case reports of S-100 protein-positive JXG were already reported in 1998 [
39], complemented by a longitudinal observation study in 2009 [
40], which demonstrates that S-100 protein reactivity cannot be reliably used as definitive marker for differentiating JXG from other histiocytoses, such as Rosai-Dorfman disease (RDD) or indeterminate cell histiocytosis. The latter also shows reactivity, with additional markers of Langerhans cells, namely CD1a and CD207 (anti-langerin), being absent in our cases. Both these entities frequently show the presence of eosinophils, which in our case were initially very prominent, in due course only very subtly present. Emperipolesis is a condition that can be observed in many physiological and pathological conditions, where hematopoietic cells in living and intact state are seen in the cytoplasm of the host cell without damage. Usually, JXG shows no emperipolesis. Yet, a high degree of emperipolesis in JXG, simulating Rosai-Dorfman disease, has been reported in individual series [
41]. Macrophages in RDD are frequently foamy and can be multinucleated, so that they are difficult to differentiate from JXG. RDD derives from sinus histiocytic macrophages that are positive for S-100 protein, fascin, CD68, CD14, CD163 and HLA-DR and negative for CD1a and CD207. In our case another peculiarity of JXG may be helpful for delineation from RDD, namely iron deposition in siderophages. This phenomenon is well known for the reaction pattern of xanthogranuloma, then entitled xanthosiderohistiocytosis, but has to the best of our knowledge (so far) not been described in RDD.
Conclusions
Juvenile xanthogranuloma belongs to the heterogeneous group of non-Langerhans cell histiocytoses and generally tends to have a good prognosis. However, the development of systemic disease can be detrimental if not diagnosed in a timely manner.
This report highlights the wide variety of clinical presentations: the first patient with an unusual skin manifestation, the second with visceral (lung, liver, pancreas, kidneys), skeletal (spine) and skin involvement and extension into soft tissue.
To make an early diagnosis and prompt adequate therapy it is pivotal, that all pediatricians be aware of this rare disease, because they are often the first to see these patients.