Erdheim-Chester Disease with BRAF V600E Mutation and a Concomitant Myeloid Malignancy Sharing NRAS and IDH2 Mutations

Erdheim-Chester disease (ECD) is a rare non-Langerhans cell histiocytosis that is currently classified under histiocytic neoplasms [1]. Only approximately 1,500 cases have been described worldwide, highlighting the rarity of this particular entity [2]. ECD shows a slight male predominance and affects mainly adults with an age range of 55–60 years with only rare confirmed pediatric cases described. The disease is chronic with signs and symptoms that vary by the mix of organs involved and the aggressiveness of the process. Most cases show multiorgan involvement with preferential distribution involving bone or the central nervous system, but periaortic area, retroperitoneum, heart, pericardium, lungs, periorbital tissue, and skin can also be affected [3]. Radiologically, ECD has many manifestations; however, long bone symmetrical osteosclerosis, retroperitoneal symmetrical and bilateral dense infiltration of the perinephric space (“hairy kidney”), and circumferential sheathing of the aorta (“coated aorta”) are typical [4].

The histopathological manifestations of ECD feature sheets of bland histiocytes that are epithelioid to xanthomatous, often with Touton giant cells. A background of fibrosis and mononuclear reactive lymphocytes and plasma cells is typical, with only rare neutrophils or eosinophils. Immunohistochemical studies show the histiocytes to be positive for CD14, CD68, CD163, factor XIIIa, fascin and negative for CD1a and CD207 (langerin). S100 is either negative or dim and variable [1, 5, 6].

The molecular landscape for ECD is characterized by recurrent alterations affecting the MAPK signaling pathway [7, 8]. The most clinically relevant activating mutation is BRAFV600E which is present in approximately 60% of cases, albeit other mutations such as ARAF, N/KRAS, and MEK, as well as fusions in ALK and NTRK1, have been described. The association of concomitant clonal hematopoiesis [9] or overt myeloid malignancy [10] (myeloproliferative neoplasms, myelodysplastic syndromes, and mixed myeloproliferative neoplasms/myelodysplastic syndromes) is increasingly being described in the setting of ECD. Here, we demonstrate a unique case of BRAFV600E-positive ECD presenting with clinical findings consistent with a myeloid malignancy with co-occurrence of NRAS and IDH2 mutations in both peripheral blood and ECD-related lesions.

An 86-year-old male presented with a past medical history significant for recent eruption of multiple cutaneous fibrohistiocytomas, long-standing end-stage renal disease, and treated high-grade bladder urothelial carcinoma in remission for multiple years. A pelvic MRI urogram for bladder cancer surveillance revealed bilateral perinephric fat stranding with diffuse perinephric and urothelial soft tissue thickening (Fig. 1). A PET scan revealed mild fluorodeoxyglucose uptake (standardized uptake value: 2.2) associated with bilateral soft tissue thickening surrounding both kidneys and in the posterior lower pelvis peritoneum anterior to the sacrum but no evidence of malignant lymphadenopathy or further distant metastatic disease. The PET scan did not reveal any bone involvement. The patient underwent a CT-guided core biopsy of this perinephric lesion which revealed mature fibroadipose soft tissue featuring a population of relatively bland-appearing histiocytes characterized by abundant foamy cytoplasm (Fig. 2). A polymorphous background was seen including a predominance of mononuclear cells (small mature lymphocytes and plasma cells) with only patchy/modest distribution of eosinophils and mature neutrophils. Immunohistochemical staining (Fig. 3) revealed the histiocytes exhibiting a CD68 (+), CD1a (−), langerin (−), and S100 (−) immunophenotype, along with positivity for a BRAF VE1 (V600E) stain.

Additional molecular testing using a targeted customized panel (Oncomine Focus Assay panel – 52 genes) including BRAF, KRAS, NRAS, PIK3CA, MAP2K1, and ALK, but not including DNMT3A or ASXL1, was conducted. This demonstrated a BRAF pV600E mutation in addition to NRAS p.G12R and IDH2 p.R140Q mutations (Table 1). At this time, significant peripheral blood abnormalities were noted, including a marked leukocytosis (105.7 10*3/μL; normal: 4.2–9.1 10*3/μL) with left-shifted maturation, anemia (8.8 g/dL; normal: 13.0–18.0 g/dL) with only modest thrombocytopenia (121 10*3/μL normal: 150–400 10*3/μL). Leukocytosis consisted of 15% lymphocytes, 0% eosinophils, 0% basophils, 36% mature neutrophils, 2% monocytes, 46% immature myeloid elements, and only 1% blasts. Peripheral blood flow cytometry demonstrated largely corroborated these findings, reporting left-shifted granulopoiesis with only modest increase in blasts (0.40% of total events) but without evidence for a lymphoproliferative disorder. Overall, given these abnormal blood counts, peripheral blood molecular next-generation sequencing (NGS) testing was performed (NeoGenomics; 128 gene legacy lymphoid molecular profile) and was reported negative for a BRAF aberration but identified identical NRAS p.G12R and IDH2 p.R140Q mutations seen in the peri-renal biopsy, with the addition of an ASXL1 frameshift (pG646Wfs*12) and DNMT3A (slice site c1015-1 G >C) mutations (Table 1). Importantly, aberrations implicated in myeloproliferative neoplasms were negative including JAK-2, CALR, MPL, and BCR-ABL1. Karyotype of the peripheral blood reported no metaphases for analysis, but a fluorescence in situ hybridization panel for myeloid/myelodysplastic disorders reported a monosomy 7 (15%; normal <3.1%) and deletion of 12p (74.0%; normal <6.3%). A bone marrow biopsy was strongly recommended at this time, but the patient declined.

The patient months prior presented with multiple skin lesions involving the head neck area that were described as small (8–15 mm) monomorphic dome-shaped brown/yellow papules with fine telangiectasia, and a skin biopsy was performed. Histology at that time revealed a mixed-cell infiltrate of both spindle cells and histiocytoid elements with ovoid nuclei and varying amounts of pale-staining cytoplasm intermixed with rare Touton giant cells in a background with thickened collagen bundles. Shave skin biopsies taken at the time (prior to presentation with the peri-renal CT findings) were interpreted as consistent with eruptive fibrohistiocytoma. These biopsies were subsequently re-examined in light of the diagnosis of ECD and showed morphological overlap with the peri-renal fat biopsies. Given the histological similarities, additional molecular testing (Ion Torrent sequencer using Oncomine Focus Assay) on the skin biopsies demonstrated a BRAF pV600E clone with the addition of NRAS p.G12R and IDH2 p.R140Q mutations (Table 1). Thus, the overall findings were consistent with cutaneous manifestations of ECD.

The patient was treated based on the molecular findings with the targeted BRAF inhibitor vemurafenib. However, shortly after initiation of therapy, the patient was admitted for a syncopal event, and peripheral blood findings at this time showed progressive leukocytosis (270.2 10*3/μL; normal: 4.2–9.1 10*3/μL) with persistent anemia (7.5 g/dL; normal: 13.0–18.0 g/dL) and thrombocytopenia (130 10*3/μL normal: 150–400 10*3/μL). Leukocytosis at this time consisted of 5% lymphocytes, 0% eosinophils, 0% basophils, 49% mature neutrophils, 22% monocytes, 17% immature myeloid elements, and 5% blasts. The patient was subsequently started on hydroxyurea, but soon after it was decided given the progressive peripheral blood findings superimposed on a patient with end-stage renal disease, palliative care was pursued and the patient expired soon after.

ECD was first described in 1930 by Jakob Erdheim and William Chester as a “lipoid granulomatosis” [11, 12]. For much of the time subsequent, it was unclear if the disease represented a benign or malignant disorder as evidence for clonality or distinct driver mutations was lacking. However, since 2012, it has been elucidated that recurrent activating kinase mutations and fusions involving the canonical MAPK (RAS-RAF-MEK-ERK) have been discovered in the majority of ECD patients [7, 13, 14]. The cell of origin in ECD, as well as other histiocytic disorders, has been a contentious question, but recently, with the understanding of the molecular landscape of these disorders, as well as the high frequency of myeloid neoplasms and clonal hematopoiesis, the cell of origin of at least a proportion of patients with ECD is consistent with that of hematopoietic progenitor cells [9].

Our case reported here shows a shared founding clonal profile for ECD and a myeloid malignancy. This association of myeloid malignancy [10] (acute myeloid leukemia, myeloproliferative neoplasms, myelodysplastic syndromes, and mixed myeloproliferative neoplasms/myelodysplastic syndromes) or clonal hematopoiesis is increasingly being described in the setting of ECD. In fact, in a recent study involving 120 patients with confirmed ECD, surprisingly 42.5% were reported to have clonal hematopoiesis and 15.8% went on to develop an overt hematological malignancy. The most frequently mutated genes in the bone marrow they described involved TET2, ASXL1, DNMT3A, and NRAS [9].

Of note, our patient was originally thought to have multiple eruptive dermatofibromas which have been associated with myeloid disorders [15]. Histologically, dermatofibroma shows interlacing fascicles of spindled cells, within a loose collagenous stroma and includes admixture of fibroblasts, macrophages, and blood vessels. Some of the lesions may show Grenz zone. A lymphocytic infiltrate can sometimes be observed. Another helpful diagnostic feature is the presence of individual collagen bundles surrounded by lesional cells, imparting a somewhat hyalinized (sclerotic) appearance. The histopathological findings of ECD skin biopsies show the pattern of a xanthogranulomatous infiltrate which usually includes dermal, and subcutaneous infiltrate of lipidized, and occasional spindle-shaped histiocytes and Touton and multinucleated giant cells, admixed with fibroses or sclerosis. An accompanying lymphocytic infiltrate can be seen with scattered deposits of hemosiderin. Since there is a considerable morphologic overlap between these two entities, they are not infrequently mistaken for one another [5, 16, 17]. Cutaneous involvement affects approximately one-third of patients with ECD and most commonly presents as xanthelasma-like tumors on the eyelids, followed by the lower extremity, but other sites including the neck, axilla, trunk, and groin, as well as mucosal infiltration of the genital area, have been described. Our patient had a rather unusual presentation in that he presented with primary skin lesions even before the manifestations of ECD, which is not a common occurrence [17]. Therefore, in a patient with multiple “eruptive fibriohistiocytoma,” an underlying clonal hematological process should be considered, especially in the context of chronically elevated white cell counts. Additional testing including bone marrow biopsy must be performed which could possibly help in early detection of these processes.

Unfortunately, a precise diagnosis regarding an underlying myeloid neoplasm is lacking as no bone marrow biopsy could be obtained; however, the cytogenetic FISH findings of a monosomy 7 and deletion of 12p, taken along with the NGS results, are consistent. In regards to the identification of founding IDH2 and NRAS mutations, the IDH2 mutation variant allele frequency was 47.1% in the peripheral blood, 29.5% in the peri-renal mass, and 29.9% in the skin, whereas the variant allele frequency measured for the NRAS variant was much lower, 9.9% in the peri-renal mass and 2.2% in the skin lesion, than that measured in the peripheral blood 41.9%. Peripheral blood contamination for NRAS would remain a possibility.

Although various treatment modalities have been tried in ECD including interferon, steroids, imatinib, rapamycin, and methotrexate, the results have been disappointing. However, with the discovery of the MAPK pathway and BRAFV600E mutations, targetable drugs are now being increasingly used in the management of these histiocytic disorders [6]. As per proposed consensus recommendations, in patients with multisystem BRAFV600-mutant ECD with life-threatening cardiac or neurologic involvement, first-line recommendation is to consider BRAF inhibitors such as vemurafenib or dabrafenib. However, an association of a paradoxical increase in white blood cells, thought to be related to activation of cytokine signaling in cells harboring non-BRAFV600E kinase mutations upon exposure to RAF inhibitors, may in some cases limit their utility [10]. Paradoxical treatment response is seen in patients treated with a BRAF inhibitor while there is an underlying myeloid neoplasm or any other RAS-driven neoplasm. For patients without BRAF-V600–ECD, NGS is recommended to evaluate other MAPK-ERK pathway alterations that could potentially be targeted with MEK inhibitor therapy. If no targeted therapy is available, IFN-α/PEG–IFN-α and cladribine therapy have been shown to be efficacious with a variable clinical response [18]. In conclusion, future studies are needed to examine the precise underlying molecular mechanisms of co-occurring ECD and myeloid malignancy; regardless, it is increasingly becoming evident that patients should be carefully screened for a co-existing myeloid neoplasm, particularly in any patient with unexplained peripheral blood counts.

This case report is submitted with no patient-specific information. All personal identifying information has been removed. This report adheres to the guidelines of human studies as per the World Medical Association Declaration of Helsinki. As per the institutional policies of New York University Grossman School of Medicine, anecdotal reports on a single patient or series of patients seen in one’s own practice and a comparison of these patients to existing reports in the literature is not research and does not require institutional review board (IRB) approval. Written informed consent was obtained from the next of kin for publication of this case report and any accompanying images.

The authors have no conflict of interest to declare.

The authors have no funding sources to declare.

Dr. Nitya Prabhakaran and Dr. Nicholas Ward have compiled the manuscript, including obtaining photomicrographs, and gathered the molecular data. Dr. George Jour has provided valuable input toward the dermatopathology aspect of this article. Dr. Arjun Balar has provided the radiographic images.

All data generated or analyzed during this study are included in this article. Further inquiries can be directed to the corresponding author.