Systemic Xanthogranuloma or Erdheim–Chester disease: A systematic review re‐examining the classification challenge

Erdheim–Chester disease (ECD) is a rare histiocytic neoplasm first described in 1930, initially characterized by bilateral osteosclerosis of the long bones (distal femur and proximal tibia/fibula). Over time, ECD has become increasingly recognized as a systemic, clonal disorder of the monocyte–macrophage lineage, often involving the retroperitoneum, cardiovascular system, lungs, central nervous system (CNS) and skin.1 In 2016, ECD was classified as a distinct histiocytic neoplasm by the World Health Organization (WHO) following the discovery of recurrent somatic mutations in the MAPK pathway, particularly BRAFV600E mutations in up to 50% of cases.2 Subsequently, the Histiocyte Society introduced a classification system distinguishing ‘classical’ ECD, defined by radiographic evidence of long-bone involvement, from ‘ECD without long-bone involvement’ (which we will refer to as ‘non-classical’ ECD), which lacks the classic skeletal findings but may still involve other organs.3 In a recent cohort of 101 patients, 86% had classical ECD and 14% were classified as non-classical.4 Interestingly, BRAFV600E mutations were significantly more frequent in classical ECD (74%) compared to non-classical ECD (14%), and classical cases also demonstrated more extensive organ involvement.4 Histologically, ECD is characterized by xanthogranulomatous inflammation composed of lipid-laden histiocytes with abundant eosinophilic cytoplasm and Touton-type multinucleated giant cells. Immunohistochemically, these histiocytes express CD68, CD163 and Factor XIIIa, but lack expression of CD1a and CD207.1 These features are shared by other non-Langerhans histiocytic disorders, particularly the xanthogranuloma (XG) family, including XG lesions historically labelled ‘juvenile’ XG (JXG) or ‘adult’ XG (AXG)- designations that are clinically convenient but biologically arbitrary. While JXG is recognized by the WHO as a benign histiocytic neoplasm typically seen in children, AXG lacks similar nosological clarity and is not formally recognized as a neoplastic process in adults.5 This overlap raises diagnostic and conceptual challenges, particularly when adults present with extracutaneous XG lesions. In the absence of long-bone involvement or other radiographic hallmarks of ECD (e.g. coated aorta, hairy kidney), systemic AXG may be used as a label, but it is not formally classified by either the WHO or Histiocyte Society.2, 3 As such, the distinction between non-classical ECD and systemic AXG remains ambiguous. It is unclear whether these represent distinct diseases, points along a shared biological continuum or nosological artefacts shaped by classification limitations and historical terminology. To better understand this question, we conducted a systematic review of cases describing systemic xanthogranuloma in adults. We queried PubMed and Embase for English-language publications from 1946 to June 2025 using combinations of terms including “xanthogranuloma”, “adult”, “systemic”, “multiple”, “multifocal”, “multicentric”, “extracutaneous” and “visceral”. After removal of duplicates and paediatric cases, 1058 studies were reviewed by two independent reviewers (V.R. and S.K.), and discrepancies were adjudicated by a senior author (G.G.). Inclusion criteria required histopathological confirmation of xanthogranulomatous lesions involving at least one extracutaneous site in adults (≥18 years), with exclusion of cases attributed to other histiocytic disorders or cases reported solely as abstracts. A flow chart depicting the methodology is included in Figure S1. This systematic review is reported in accordance with PRISMA 2020 guidance (checklist in supplement). Given that included studies were case reports/series, we did not apply a formal risk-of-bias instrument; instead, we prespecified key reporting domains (diagnostic confirmation, staging/whole-body imaging, molecular testing and follow-up) and summarized limitations descriptively. From this review, we identified 50 cases of systemic AXG. We also include a recent illustrative case from our institution in the subsequent section. A summary of all cases is included in Table 1 with details of each individual case in Table S1. A 54-year-old man with a 17-year history of progressive dysarthria and ataxia was initially diagnosed and treated for multiple sclerosis with corticosteroids and ocrelizumab. Brain biopsy demonstrated lymphohistiocytic infiltrates but was non-diagnostic. Due to a lack of clinical or radiographic response, he was additionally treated with mycophenolate mofetil, cyclophosphamide, and intrathecal methotrexate. Brain magnetic resonance imaging (MRI) showed post-gadolinium T1-weighted enhancing lesions in the cerebellum and brainstem (Figure 1B) and both enhancing and non-enhancing lesions in the spinal cord (not shown). Whole-body Fluorodeoxyglucose-Positron Emission Tomography/Computed Tomography (FDG-PET/CT) revealed no extracranial disease. Subsequent dermatological examination identified a yellowish papule on the right thigh (Figure 1A). Skin biopsy revealed CD68+, CD163+, S100+, CD1a– and cyclin D1+ histiocytes with Touton giant cells (Figure 1A). Molecular testing revealed a NOTCH1 p.C1025* nonsense mutation but did not identify MAPK pathway mutations. This case meets our inclusion criteria for systemic AXG; however, given CNS-predominant disease and the absence of a recognized ‘systemic AXG’ category in major classifications, the patient was managed as ECD-spectrum disease and treated with trametinib with partial radiographic improvement (Figure 1C). The literature review revealed that, among 51 cases (including ours), the median age at diagnosis was 38 years (range 18–75), with an equal sex distribution. The most commonly affected extracutaneous sites were the spinal vertebrae/spinal cord (29%), brain (22%), oral cavity (18%) and bones (14%). Despite systemic involvement, only 16% reported whole-body imaging, and molecular testing was reported in just 12% of cases. Most patients (78%) underwent surgical resection as first-line treatment, with systemic therapies rarely used (Table S2). Median follow-up duration was 22 months (range 1–142 months). Our review highlights how current nomenclature and classification may influence staging and molecular evaluation in adults with extracutaneous xanthogranulomatous disease, particularly given the low reported rates of whole-body imaging and mutational testing. Terminology may contribute to this gap: the label ‘xanthogranuloma’ can be perceived as descriptive or benign and may not signal a histiocytic neoplasm with systemic potential; using ‘XG histiocytosis’ or 'XG family of histiocytic neoplasms' as a descriptor in clinical communication may better prompt appropriate evaluation.6 CNS and orofacial lesions were common in our review and are often difficult or high-risk anatomic sites to biopsy, potentially delaying diagnosis or leading clinicians to rely on imaging and clinical correlation rather than tissue confirmation. Such cases may also be more likely to be reported due to diagnostic complexity or prolonged clinical courses contributing to further reporting and referral bias. Importantly, CNS involvement as seen in our case is well described in ECD and may further reflect overlapping disease biology rather than a distinct predilection towards systemic AXG. This is also the clinical scenario in which ‘benign’ labelling may be most consequential, as CNS-predominant disease can progress while comprehensive work-up is deferred. As our institutional case highlights, diagnosis often hinges not on histopathology alone—since XG, AXG and ECD are indistinguishable under the microscope—but on a combination of clinical judgement, imaging and context-specific molecular profiling. In our case, MAPK alterations were not reported, yet radiographic response to trametinib illustrates how classification and clinical context can influence access to targeted therapy even when molecular characterization is incomplete. Supporting biological overlap, in a recent study of adult patients with multiple cutaneous adult xanthogranulomas, approximately 28% had systemic involvement.7 Notably, 57% harboured alterations in the MAPK pathway, including BRAFV600E mutations in 11%, underscoring the overlap between XG and ECD.7 These findings challenge the notion that all XG lesions are benign and reinforce the need for thorough evaluation of systemic involvement in adults. The distinction between ‘non-classical ECD’ and systemic AXG is clinically significant but conceptually tenuous. While classical ECD with long bone involvement is well defined, non-classical ECD lacks clear boundaries and risks being conflated with systemic AXG. Moreover, the label ‘non-classical’ implies an incomplete or variant form of ECD, when in fact it may represent a biological spectrum rather than a distinct entity. This is not purely academic: recognition of ECD as a neoplastic disorder affects access to oncological care, clinical trials and targeted therapies, whereas AXG may be perceived as benign, contributing to under-staging and under-treatment. Rather than proposing new diagnostic criteria, our data support a pragmatic approach: adults with extracutaneous xanthogranulomatous lesions should undergo comprehensive staging (e.g. whole-body FDG-PET/CT for assessment of skeletal and other organ involvement) and molecular profiling when feasible. Identification of canonical MAPK-pathway alterations, characteristic ECD-pattern radiographic findings or involvement of critical organs should prompt consideration of ECD-spectrum disease to facilitate standardized monitoring and access to targeted therapy. Prospective cohorts with uniform phenotyping/genotyping are needed to determine whether outcomes differ between systemic AXG and non-classical ECD. We recognize that extracutaneous xanthogranulomatous lesions have a broad differential diagnosis, including inflammatory, infectious and neoplastic mimics; however, our focus here is the nosological overlap with clonal histiocytoses, specifically systemic AXG versus ECD-spectrum disease, and the downstream impact on staging and molecular evaluation. Limitations of our study include biases inherent to case reports/series such as selection/publication bias, incomplete/selective reporting and diagnostic heterogeneity. Thus, these results should be interpreted descriptively and as hypothesis generating. In summary, our results suggest that extracutaneous xanthogranulomatous disease in adults is frequently managed as a benign ‘XG’ label rather than a histiocytic neoplasm with potential ECD-spectrum overlap, contributing to low rates of staging and molecular profiling. Given the substantial clinicopathological and molecular overlap between systemic AXG and non-classical ECD, and the potential consequences of misclassification, a more comprehensive approach to diagnosis and staging is needed. We advocate for heightened clinical awareness, broader use of molecular diagnostics and an open dialogue about whether current histiocytic classifications adequately capture the realities of adult-onset xanthogranulomatous disease. Further consensus efforts may benefit from harmonizing terminology across histiocytic neoplasms using behaviour-based descriptors (e.g. localized vs. multisystem) rather than age-based labels. Vyshnavi Rallapalle performed the research, analysed the data and wrote the manuscript. Sailesh I.S. Kumar and Brinda Shukla performed the research and analysed the data. Karen L. Rech provided pathological images and edited the manuscript. Ronald S. Go and Jithma P. Abeykoon helped write and edit the manuscript. Pradeep Bhambhvani provided radiographic images and edited the manuscript. Shuko Harada helped write and edit the manuscript. Julio C. Sartori-Valinotti, W. Oliver Tobin, Nabeela Nathoo, Aditi Saha, Lucy Bailey and Daniel Sokolowski helped write and edit the manuscript. Aishwarya Ravindran provided pathological images, created the figures and edited the manuscript. Gaurav Goyal designed the research study, wrote the manuscript, edited the manuscript and supervised the data collection. The authors have nothing to report. This paper was supported by the American Cancer Society award RSG-24-1317006-01-CTPS (G.G.). G.G. is a Scholar in Clinical Research for Blood Cancer United. G.G. has served on the advisory board for Seagen and Sobi and received consulting fees from Recordati and Pharmassentia. All data are included in the manuscript and in the Supporting Information. Data S1. Figure S1. Table S1. Table S2. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.