https://orcid.org/0000-0001-9793-7032
TO THE EDITOR:
Indeterminate dendritic cell histiocytosis (IDCH) is a rare histiocytic neoplasm usually presenting in the skin.1-5 Since its first description,1 ∼100 cases have been described.5 IDCH mainly affects adults but can occur in children.5,6 Histologically, the disease is characterized by CD1a+ histiocytic cells, and the absence of CD207/Langerin expression distinguishes it from Langerhans cell histiocytosis (LCH).2,3 In 2015, Brown et al reported 3 cases with an ETV3::NCOA2 fusion, a translocation almost specific to IDCH.7,8 However, this translocation was not identified in several other cases.7,9 Thus, the complete genomic landscape of IDCH remains to be elucidated.
We searched for patients with IDCH in the Dutch Nationwide Pathology Databank (Palga) and in University Hospitals Leuven.10,11 Archival tissue slides and blocks and pseudonymized clinical data and images were obtained.10,11 Some patients (cases 3, 8, P1, and P3) were identified through prior studies on LCH,12,13 after examining CD207/Langerin expression. Three patients were previously described (cases 16, 514, and 615,16). Central pathology review was conducted (P.C.W.H.) and additional immunohistochemical stains were performed when indicated. DNA was isolated from microdissected tissue and subjected to targeted locus capture–based next-generation sequencing according to a published protocol,17 except for 2 cases analyzed by whole-transcriptome sequencing through routine diagnostics. A custom capture probe panel was used, targeting exons of 126 genes, as well as introns of 55 of 126 genes to also detect structural variants. A list of genes and detailed methods are presented in another publication.18 Targeted RNA sequencing to investigate gene fusions at the transcriptomic level was performed using Archer technology, as described previously.19 The study was approved by the institutional review board of Leiden University Medical Center (B19.074), the ethics committee of University Hospitals Leuven (S64542/S66281), the Palga scientific council and privacy committee (LZV-2016-183), and the biobank and data access committee of the Princess Máxima Center for Pediatric Oncology (PMCLAB2024.0512).
Twelve patients with IDCH confirmed by central pathology review were included. In addition, we analyzed 6 cases with a potential diagnosis of IDCH, exhibiting some atypical histologic features (eg, partial expression of CD1a). Detailed clinicopathologic characteristics are presented in supplemental Tables 1 and 2. Furthermore, we analyzed 1 malignant histiocytosis with an indeterminate dendritic cell phenotype (supplemental Figure 1) and 2 histologic mimics (supplemental Figure 2), including a reticulohistiocytoma with rare CD1a+ cells and a case of primary cutaneous T-cell lymphoma (mycosis fungoides) accompanied by a prominent CD1a+ infiltrate.20
The 12 patients with IDCH comprised 8 adults and 4 children (supplemental Table 1), with a median age of 57.5 years (range, 0-87). All children were infants and 3 had skin lesions at birth. Nine patients were male. Eight had isolated skin lesions, whereas 4 patients had multisystemic disease, including skin lesions in 3 of 4 patients. Skin lesions were multifocal in 10 of 11 patients and generally consisted of papules or nodules on the trunk, extremities, and/or face (Figure 1A). Patients with multisystemic disease comprised 1 infant (case 9) and 3 adults (cases 10, 11, and 12), with disease involvement of the skin, bone marrow, and spleen (case 9); the skin and lungs (case 10); the intestines, lymph nodes, and spleen (case 11); or the skin and breasts (case 12). Notably, case 12 developed the breast lesions 3.4 years after diagnosis. Isolated skin disease was treated with (topical) steroids and/or phototherapy in 5 patients and spontaneously regressed in 3 others. Among patients with multisystemic disease, 1 adult died soon after diagnosis, 2 adults received systemic treatment but died from their disease after 4 and 7 years, respectively, and the infant achieved complete remission with vinblastine/prednisone chemotherapy. Four adults were diagnosed with an additional hematologic malignancy, including 1 with myeloid sarcoma (case 5), 2 with chronic myelomonocytic leukemia (cases 6 and 7), and 1 with myelodysplastic syndrome (case 10). One malignancy was diagnosed concurrently with IDCH (case 6), whereas the other 3 were diagnosed between 2 and 6 years after the diagnosis of IDCH (supplemental Table 1).
Clinical and pathologic features of IDCH. (A) Photographs of lesions in cases 1, 3, 4, 5, 9, 10, and 12. Shown are congenital purpuric papules in case 1, which spontaneously regressed over several weeks.6 Case 3 was a 9-month-old child with multiple yellow-brown papules spread over the body and an erythematous, crusted nodule on the right nasal ala. The lesion on the nose was successfully treated with topical steroids, whereas the other lesions regressed spontaneously over several months. Cases 4 and 5 were adults with multiple brown-red or skin-colored, asymptomatic papules on the trunk and extremities, which responded to ultraviolet B phototherapy. Case 9 was an infant with congenital purpuric skin lesions who developed anemia, thrombocytopenia, and splenomegaly in the first month after birth, reminiscent of infants developing high-risk LCH. The child achieved complete remission with vinblastine/prednisone chemotherapy. In contrast, cases 10 and 12, both adults, died from progressive disease despite systemic treatment. Case 10 presented with >100 papules spread over the body and bilateral lung nodules (indicated by white arrows). Case 12 presented with skin lesions on the right cheek but later developed locally invasive disease requiring facial surgery and bilateral breast lesions (not shown). (B) Photomicrographs of tissue slides stained with hematoxylin and eosin (H&E) or stained for specific proteins; the latter are indicated at the bottom of the images. Case numbers are provided in the upper left corners; black bars in the lower left corners indicate 20 μm. Shown are infiltrates of histiocytoid cells with frequent nuclear grooves or indentations, as highlighted in several inlets. Clusters of cells were often located around vessels, as illustrated by H&E images of cases 1 and 10. In case 4, a small intraepithelial focus of abnormal histiocytes was observed in the lesion of the back (right H&E image); this lesion was also notable for more cellular atypia and CD163+ multinucleated cells (inlet). Dense infiltrations of similar, histiocytoid cells were observed in extracutaneous lesions, often accompanied by prominent lymphocytic infiltrates (not shown). As illustrated by multiple photomicrographs of case 4 (with an ETV3::NCOA2 fusion), lesional cells stained positive for CD1a and CD68, and were negative for CD207/Langerin and S100. Yet, CD68 was negative and S100 was positive in other cases (supplemental Table 2). The pattern of CD68 expression was diffuse granular cytoplasmic expression, with occasional Golgi dot-like accentuation (most notable in cases 1 and 7). The pattern of S100 expression was diffuse cytoplasmic and nuclear expression, although nuclear expression was less apparent in case 11 (shown). As illustrated by the photomicrograph of the back lesion of case 4, CD1a staining sometimes revealed a remarkable branching pattern in the deeper dermis, in which histiocytes were often located around vessels and, from there, spread through the dermal collagen. CD163 was generally negative, with rare positive lesional cells, whereas cyclin D1 frequently stained the lesional cells, and PU.1 was diffusely positive. Mutational status of all patients is depicted in Figure 2A. Detailed clinical and pathologic information is provided in supplemental Tables 1 and 2.
Clinical and pathologic features of IDCH. (A) Photographs of lesions in cases 1, 3, 4, 5, 9, 10, and 12. Shown are congenital purpuric papules in case 1, which spontaneously regressed over several weeks.6 Case 3 was a 9-month-old child with multiple yellow-brown papules spread over the body and an erythematous, crusted nodule on the right nasal ala. The lesion on the nose was successfully treated with topical steroids, whereas the other lesions regressed spontaneously over several months. Cases 4 and 5 were adults with multiple brown-red or skin-colored, asymptomatic papules on the trunk and extremities, which responded to ultraviolet B phototherapy. Case 9 was an infant with congenital purpuric skin lesions who developed anemia, thrombocytopenia, and splenomegaly in the first month after birth, reminiscent of infants developing high-risk LCH. The child achieved complete remission with vinblastine/prednisone chemotherapy. In contrast, cases 10 and 12, both adults, died from progressive disease despite systemic treatment. Case 10 presented with >100 papules spread over the body and bilateral lung nodules (indicated by white arrows). Case 12 presented with skin lesions on the right cheek but later developed locally invasive disease requiring facial surgery and bilateral breast lesions (not shown). (B) Photomicrographs of tissue slides stained with hematoxylin and eosin (H&E) or stained for specific proteins; the latter are indicated at the bottom of the images. Case numbers are provided in the upper left corners; black bars in the lower left corners indicate 20 μm. Shown are infiltrates of histiocytoid cells with frequent nuclear grooves or indentations, as highlighted in several inlets. Clusters of cells were often located around vessels, as illustrated by H&E images of cases 1 and 10. In case 4, a small intraepithelial focus of abnormal histiocytes was observed in the lesion of the back (right H&E image); this lesion was also notable for more cellular atypia and CD163+ multinucleated cells (inlet). Dense infiltrations of similar, histiocytoid cells were observed in extracutaneous lesions, often accompanied by prominent lymphocytic infiltrates (not shown). As illustrated by multiple photomicrographs of case 4 (with an ETV3::NCOA2 fusion), lesional cells stained positive for CD1a and CD68, and were negative for CD207/Langerin and S100. Yet, CD68 was negative and S100 was positive in other cases (supplemental Table 2). The pattern of CD68 expression was diffuse granular cytoplasmic expression, with occasional Golgi dot-like accentuation (most notable in cases 1 and 7). The pattern of S100 expression was diffuse cytoplasmic and nuclear expression, although nuclear expression was less apparent in case 11 (shown). As illustrated by the photomicrograph of the back lesion of case 4, CD1a staining sometimes revealed a remarkable branching pattern in the deeper dermis, in which histiocytes were often located around vessels and, from there, spread through the dermal collagen. CD163 was generally negative, with rare positive lesional cells, whereas cyclin D1 frequently stained the lesional cells, and PU.1 was diffusely positive. Mutational status of all patients is depicted in Figure 2A. Detailed clinical and pathologic information is provided in supplemental Tables 1 and 2.
Histologic features of IDCH skin lesions included diffuse or clustered dermal infiltrations of histiocytoid cells with frequent nuclear grooves/indentations (Figure 1B; supplemental Table 2). Epidermotropism was generally not observed, although 3 cases had small intraepithelial foci of abnormal histiocytes (cases 2, 4, and 7). Mitoses were rare; eosinophils were infrequent. Lesional cells stained positive for CD1a and negative for CD207/Langerin. Yet, CD207/Langerin expression was observed in the pulmonary lesion of case 10 (supplemental Figure 3), representing a rare case with both IDCH (in the skin) and LCH (in the lungs). S100 and CD68 were positive in a subset of patients, whereas PU.1 and CD4 were consistently positive (supplemental Table 2). CD163 was generally negative, with occasional positive lesional cells. Cyclin D1 expression was absent (1/10), rare (2/10), occasional (2/10), or frequent (5/10).
Putative driver alterations were identified in all 12 patients with IDCH (Figure 2A; supplemental Tables 3 and 4). Alterations included ETV3::NCOA2 fusions in 4 patients and mitogen-activated protein kinase (MAPK) pathway gene alterations in 8 others, including KRAS, NRAS, BRAF, and MAP2K1 alterations. Notably, 3 of 4 patients with BRAF alterations had mutations other than BRAFV600E. ETV3::NCOA2 fusions were also detected at the transcriptomic level in all 3 analyzed cases. In addition, 7 of 8 adults harbored mutations in epigenetic regulators, including TET2, ASXL1, EZH2, PHF6, and KMT2D, or in the splicing factor ZRSR2. In 5 cases, 2 separate histiocytosis lesions were sequenced, and identical genetic alterations were identified in all pairs (Figure 2B-C). For example, the ETV3::NCOA2 fusion was detected in IDCH skin lesions from the abdomen and the leg in case 4 and in the IDCH (skin) and LCH (lung) lesions in case 10. In 3 of 5 patients, mutations unique to 1 lesion were also detected. Patients with ETV3::NCOA2 fusions included 1 child and 3 adults, including an adult with multisystemic disease (case 10). The other 3 patients with multisystemic disease had BRAF and/or KRAS mutations.
Molecular characterization of IDCH and additional myeloid malignancies. (A) Oncoprint depicting clinicopathologic and molecular features of 12 patients with IDCH, 6 patients with potential IDCH, and 1 patient with MH-IDC (depicted on the far right). Every column represents 1 patient; case numbers are provided at the bottom of the plot. Blue squares indicate the presence of a clinicopathologic characteristic, orange squares indicate the presence of a gene fusion, red squares indicate the presence of a SNV or insertion and/or deletion (indel), and purple squares indicate the presence of CNV. (B) Genetic findings in 4 patients of whom 2 separate IDCH lesions were molecularly analyzed, demonstrating identical mutations in paired lesions in all cases. In addition, mutations unique to 1 lesion were sometimes identified. (C) Genetic findings in 4 patients of whom the additional myeloid malignancy was molecularly analyzed. Shared genetic alterations between the IDCH and additional hematologic malignancy were identified in 2 of 4 cases, strongly suggesting a clonal relationship of these hematopoietic neoplasms. Details on detected genetic alterations, including variant allele frequencies, are provided in supplemental Tables 3 and 4. CNV, copy number variation; MH-IDC, malignant histiocytosis with an indeterminate dendritic cell phenotype; SNV, single-nucleotide variant.
Molecular characterization of IDCH and additional myeloid malignancies. (A) Oncoprint depicting clinicopathologic and molecular features of 12 patients with IDCH, 6 patients with potential IDCH, and 1 patient with MH-IDC (depicted on the far right). Every column represents 1 patient; case numbers are provided at the bottom of the plot. Blue squares indicate the presence of a clinicopathologic characteristic, orange squares indicate the presence of a gene fusion, red squares indicate the presence of a SNV or insertion and/or deletion (indel), and purple squares indicate the presence of CNV. (B) Genetic findings in 4 patients of whom 2 separate IDCH lesions were molecularly analyzed, demonstrating identical mutations in paired lesions in all cases. In addition, mutations unique to 1 lesion were sometimes identified. (C) Genetic findings in 4 patients of whom the additional myeloid malignancy was molecularly analyzed. Shared genetic alterations between the IDCH and additional hematologic malignancy were identified in 2 of 4 cases, strongly suggesting a clonal relationship of these hematopoietic neoplasms. Details on detected genetic alterations, including variant allele frequencies, are provided in supplemental Tables 3 and 4. CNV, copy number variation; MH-IDC, malignant histiocytosis with an indeterminate dendritic cell phenotype; SNV, single-nucleotide variant.
Similar mutations were detected in 2 of 6 patients with potential IDCH, including KRAS and ASXL1 mutations in the unifocal lesion of case P4 and an identical TET2 alteration in both skin lesions of case P5 (Figure 2A; supplemental Figure 4). In the malignant histiocytosis with an indeterminate dendritic cell phenotype, BRAF, KRAS, CDKN2A, and CREBBP mutations were detected. In the mimics, an ABCA1::BRAF fusion and TP53 mutation were identified in the reticulohistiocytoma with rare CD1a+ cells, whereas no alterations were detected in the microdissected CD1a+ infiltrate accompanying mycosis fungoides.20 The latter supports a reactive rather than a neoplastic nature of this histiocytic infiltrate, although an accompanying histiocytosis cannot be fully excluded.
In all 4 adults with an additional hematologic malignancy, tissue samples of the malignancy could be retrieved for comparative molecular analysis. In 2 of 4 patients, identical genetic alterations were identified (Figure 2C). Both had chronic myelomonocytic leukemias that shared NRAS/KRAS, TET2, and either ASXL1 (case 6) or ZRSR2 and PHF6 (case 7) mutations with their IDCH. The histiocytosis in case 6 also harbored unique EZH2 and BRAF mutations.
This study describes a large molecularly characterized cohort of patients with IDCH. We confirm the recurrent occurrence of ETV3::NCOA2 fusions7 and expand its identification to pediatric cases and aggressive, multisystemic histiocytosis. In addition, we reveal frequent MAPK pathway gene alterations, particularly KRAS and BRAF mutations, which are targetable.21 Our findings demonstrate that IDCH is molecularly distinct from LCH, which is more often driven by BRAFV600E, BRAF exon 12, and MAP2K1 mutations.12,13,22,23 IDCH and LCH also vary clinically, with IDCH’s predilection for the skin, lack of bone involvement, and predominant presentation in adults as distinguishing factors. Yet, differentiating between the 2 diseases can be challenging on an individual basis, as IDCH can appear very similar to isolated skin or multisystemic LCH (eg, in cases 1 and 9). Histologically, IDCH lesions often exhibit few eosinophils but can otherwise mimic LCH, underscoring the importance of CD207/Langerin immunohistochemistry. By demonstrating identical genetic alterations in separate lesions of 5 patients, we prove the clonal relationship of different sites of the disease. This suggests that IDCH may derive from mutated precursor cells in the blood or bone marrow that can seed different tissues. Together, our findings may inform (molecular) diagnostics and treatment of future patients.
Among adults, we demonstrate frequent mutations in epigenetic regulators like TET2 and ASXL1, which are tumor suppressor genes mutated in many myeloid neoplasms.24 A recent literature review revealed that a third of reported adult patients with IDCH had an additional hematologic malignancy.5 This high frequency is substantiated by the incidence among our adult cohort (4/8) and exceeds the frequencies observed in LCH or Erdheim-Chester Disease.13,25 Previously, shared genetic alterations have been detected in histiocytic neoplasms and associated hematologic malignancies.13,15,16,25 Accordingly, we revealed identical mutations in the IDCH and additional myeloid leukemia in 2 patients, strongly suggesting a common clonal origin. These results further support that IDCH may arise from mutated myeloid precursor cells and underscore that patients should be carefully evaluated for associated hematologic cancers.
Acknowledgments: The authors thank the immunohistochemical laboratory of Leiden University Medical Center for performing additional immunohistochemical stains. The authors also thank A. Beishuizen (Princess Máxima Center for Pediatric Oncology), D.J.P. Willemsen (Anna Hospital), A.M. van Tuyll van Serooskerken (Haga Hospital), and D. Dittmar (University Medical Center Groningen) for providing clinical data and/or images of individual patients.
This study was funded by a grant from Stichting de Merel (P.G.K., T.v.W., A.G.S.v.H., and P.C.W.H.). F.J.S.H.W.-A.-J. personally financed routine molecular diagnostic tests for her patients when these were not reimbursed. P.G.K. received an MD/PhD grant from Leiden University Medical Center. T.T. holds a Mandate for Fundamental and Translational Research from the “Stichting tegen Kanker” (2014-083 and 2019-091) and is supported by the ME TO YOU Foundation. I.V.B. is a recipient of a postdoctoral fellowship sponsored by the Clinical Council for Research and Education from University Hospitals Leuven.
Contribution: F.J.S.H.W.-A.-J., K.D.Q., C.v.d.B., L.N., and J.A.M.v.L. were involved in the clinical care of patients and provided pseudonymized clinical data and images; I.V.B., M.J.K., K.M.H., M.A.S.-V., G.F.H.D., R.M.V., P.M.J., E.H., T.T., A.H.G.C., and T.v.W. were involved in the routine (molecular) pathologic evaluation of included patients and/or provided archived tissue samples; R.H.P.V. subjected the sequencing data to the somatic variant calling pipeline; E.S. and J.F.S. assisted in targeted locus capture–based next-generation sequencing (TLC-NGS) panel design, performed TLC-NGS, and analyzed sequencing data for structural variants; T.v.W. reviewed small genetic variants called by the variant calling pipeline; A.H.B. helped with selecting the Palga search strategy and assisted in the Palga intermediary procedure; P.G.K. collected information and samples, performed microtomy, immunohistochemistry, and tissue microdissection, made the figures and tables, and drafted the manuscript; T.v.W., A.G.S.v.H., and P.C.W.H. revised the manuscript; and all authors reviewed and approved the manuscript before submission.
Conflict-of-interest disclosure: E.S. and J.F.S. are employees of Cergentis BV (a Solvias company). The remaining authors declare no competing financial interests.
Correspondence: Paul G. Kemps, Department of Pathology, Leiden University Medical Center, PO Box 9600, 2300 RC Leiden, The Netherlands; email: p.g.kemps@lumc.nl.
