Rosai–Dorfman disease (RDD), as a synonym for sinus histiocytosis with massive lymphadenopathy, was first described in 1965 (1), although it was not recognized as a distinct clinical entity until 1969 (2). RDD is a histiocytic proliferation disorder characterized by large S100-positive histiocytes exhibiting emperipolesis (3). RDD usually involves the lymph nodes, most frequently in the neck. As a result, patients typically present with painless bilateral cervical lymphadenopathy, in conjunction with fever, leukocytosis, and elevated erythrocyte sedimentation rate (4). Extranodal disease may occur as a primary process or in association with bone involvement.

Primary intraosseous RDD is an extremely rare disease. To date, only about 100 cases of intraosseous RDD have been reported, mainly as case reports rather than as study series. Because of its rarity, the clinicopathological characteristics and prognostic features of primary intraosseous RDD have not been well described. In an attempt to expand the known clinicopathologic and molecular genetic characteristics, as well as prognostic features, we retrospectively analyzed four cases of primary intraosseous RDD in our hospital with a review of an additional 62 cases with complete follow-up data from the literature.

RDD is a rare histiocytic proliferative disorder of uncertain pathogenesis, which was first described in 1965 (1) and recognized as a distinct clinical entity in 1969 by Rosai and Dorfman (2). In the fifth edition of the WHO classification of soft tissue and bone tumors, RDD was classified as a hematopoietic neoplasm of the bone (3). Most of the patients present with lymph node involvement manifesting as adenopathy. In the present retrospective study, a total of 18 cases of primary lymph node, skin, or other non-osseous site-based RDD were diagnosed in our hospital. Extranodal disease may occur as a primary process or in association with nodal involvement (5). However, primary RDD of the bone is extremely rare and the occurrence of this entity remains unknown. Primary intraosseous RDD was identified in 0.14% (4/2,800) of total bone biopsies at our institution over the study period.

Due to its rarity, the clinicopathologic features and prognosis of primary intraosseous RDD have not been well characterized. To date, no known risk factors have been identified. After an extensive search of the English literature, only 62 cases of primary intraosseous RDD with complete survival data were identified. The median age of the patients was 24 years old, and there was no significant difference in age between genders. Pain of the affected area was the most common presenting symptom. Interestingly, we found that primary intraosseous RDD most occurred in the bones of the extremities, with the proximal tibia being the most common location. In addition, most of the tumors were single osteolytic lesions.

The etiology of RDD remains uncertain. The proposed mechanisms include immune dysfunction and possible viral infection. A few studies have reported that RDD may be related to the Epstein–Barr virus, human herpesvirus infections, and so on (43, 44). However, no Epstein–Barr virus or human herpesvirus infections were observed in primary intraosseous RDD from our present study and those reported in the literature, so any potential correlation between viral infection and the pathogenesis of primary intraosseous RDD cannot be clarified.

RDD has always been considered as a disease of histiocytic polyclonal hyperplasia, but in recent years, molecular genetics suggests that it may have a potential monoclonal pathogenesis. Fatobene and Haroche reported one case of nodal RDD with confirmed BRAF-V600E mutation, representing a promising therapeutic target, especially for patients with refractory or extensive disease (45). In addition, mutually exclusive KRAS and MAP2K1 mutations were described in one-third of the cases of RDD, suggesting that this subgroup is clonal and involves activation of the MAPK/ERK pathway (46). Wu et al. reported an NRAS mutation in cutaneous RDD (47), indicating that NRAS mutations in the MAPK/ERK pathway may be involved in the pathogenesis of cutaneous RDD. Nevertheless, there was only one report on the molecular genetics of primary intraosseous RDD. Dong et al. detected BRAF-V600E and KRAS mutations in seven undecalcified primary intraosseous RDD cases, which showed that only one case with the concurrence of RDD and Langerhans cell histiocytosis (LCH) had BRAF-V600E mutation, suggesting that the BRAF-V600E mutation may be caused by LCH lesions rather than RDD lesions. Similarly, no BRAF-V600E, KRAS, and NRAS mutations were detected in our cases and all samples were not decalcified. Unfortunately, a larger NGS panel was not performed on our cases and the reported cases to assess for the presence of mutations in other genes within the MAPK pathway. Like our results, some studies (25, 48) have demonstrated immunohistochemical cyclin D1 expression in RDD cases including bone lesions, reflecting constitutive MAPK pathway activation in the pathogenesis of RDD. However, since few studies have been reported and no larger NGS panel has been performed, it is still unclear whether the primary intraosseous RDD is related to the activation of the MAPK pathway. Therefore, not only more cases but also more robust mutational analysis are needed to further confirm the findings.

Tracht et al. demonstrated a possible histologic overlap between RDD and the more common IgG4-related disease (IRD), which could cause problems in pathologic diagnosis (49). Another study demonstrated that RDD of the breast can show a significant increase in IgG4+ plasma cells as well as fibrosis, which may further complicate matters (50). In our present study, one case showed increased IgG4-positive plasma cells and IgG4/IgG ratio, but serum IgG4 and IgG levels were normal. Therefore, we speculate that RDD histomorphology may be associated with that of IRD, but a relationship between RDD and IRD has not been definitively established. A diagnosis of concurrence of RDD and IRD should integrate the pathological features, the number of IgG4-positive plasma cells, clinical manifestations, serological examinations. and so on.

The imaging characteristics of primary intraosseous RDD are not specific and often misleading. Radiographically, bone lesions are often misdiagnosed as osteomyelitis or LCH. Other entities in differential diagnosis include Erdheim–Chester disease, lymphoma, plasma cell myeloma, and metastatic disease (5). Osteomyelitis is a necrotizing and sclerosing bone disease dominated by inflammation, often including numerous neutrophils and with frequent periosteal reactive bone formation. Histologically, the mixed inflammatory infiltrate with focal neutrophilic micro-abscesses, occasional multinucleated giant cells, and granuloma-like histiocyte collections may suggest infection or granulomatous disease. However, the characteristic S100-positive histiocytes with emperipolesis are not seen in either condition. As the name suggests, LCH is dominated by the proliferation of Langerhans histiocytes. Langerhans histiocytes are usually found in granuloma-like clusters and have characteristic elongated, indented, grooved, or convoluted nuclei with inconspicuous nucleoli. In addition to the S100 protein, they are consistently positive for CD1a and langerin, which are not expressed by RDD histiocytes. Erdheim–Chester disease is a multisystemic proliferative histiocytic disorder, characterized by long bone involvement with bilateral and symmetrical sclerotic lesions. Frequently, there is extraskeletal involvement including the cardiovascular system, central nervous system, kidneys, and lungs. Histologically, there is a proliferation of foamy histiocytes within the marrow spaces with associated fibrosis and thickening of bone trabeculae. The proliferating histiocytes are positive for CD163 and CD68 and are usually negative for S100. They are also negative for CD1a and langerin. Emperipolesis is not seen in Erdheim–Chester disease. Metastatic carcinoma and melanoma can be ruled out by histomorphology and the lack of expression of epithelial and melanocytic markers.

Surgical resection or curettage is the most common treatment of primary intraosseous RDD. At present, there is great controversy about the relative benefits of postoperative adjuvant radiotherapy and steroids therapy (51). The prognosis is good, but local recurrences or secondary lesions in other locations may occur after surgery in some cases. Our analysis showed that the 5-year PFS of patients with primary intraosseous RDD was 57.5%. Interestingly, our analysis of all prior reported cases showed that female patients had a trend toward superior PFS compared with male patients. No statistically significant difference was found in PFS between patients with different age groups, tumor locations, or number of lesions.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

All patients provided written informed consent for the collection and publication of their medical information during the first visit to the hospital.

XW designed the study and wrote the manuscript. YJY and MZ contributed to the clinical collection, as well as data analysis. CC and YHS provided the cases. XKW performed molecular detection. RRZ and HHG performed immunohistochemical staining and in-situ hybridization (ISH). WL collected the follow-up information. QZX offered imaging data. XL provided supervision, conceived the project, and provided leadership. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.