Blastic Plasmacytoid Dendritic Cell Neoplasm

Epidemiology

It remains challenging to accurately quantify the incidence of BPDCN due to the variety of names the disease had previously and because it was only recently given a code in the International Classification of Diseases (ICD). BPDCN incidence has been estimated by various sources as approximately 0.5% of all hematologic malignancies, 0.45/1,000,000 population, or 200 to 500 cases per year in the United States. The true incidence of BPDCN may be underestimated due to unfamiliarity with the disease among clinicians and pathologists, and because at least 20% of patients with BPDCN have a pre-existing myeloid neoplasm, such as chronic myelomonocytic leukemia (CMML), myelodysplastic syndrome (MDS), or AML. Careful clinical and pathological examination is necessary to distinguish between BPDCN and AML, particularly in patients with a known myeloid disorder (discussed in more detail later). In particular, BPDCN should be considered when patients with a myeloid neoplasm develop skin lesions. As recognition of BPDCN as a distinct entity improves and because there is an approved BPDCN-specific targeted therapy available and others in development, biopsy-confirmed incidence of the disease may continue to rise.

The median age of diagnosis in adult patients is approximately 60 to 65 years old. The disease also occurs in children and there is a bimodal incidence pattern, similar to what is observed in B-cell acute lymphoblastic leukemia, with a separate pediatric peak of BPDCN incidence at age 8 to 10 years. There is a striking male predominance in adults with BPDCN, with men affected 3 to 4 times more frequently than women. The disease is not nearly as gender-biased in children, where the male incidence is only 1.2 times higher than in females. There is no exposure history clearly linked to BPDCN, such as prior chemotherapy, radiation, or environmental toxins. Different from conventional dendritic cells (cDCs), which present processed antigens to lymphocytes, pDCs’ primary role is to sense pathogens and activate the immune system by producing interferon α and β. Thus, there has been interest in determining if BPDCN has an infectious etiology. However, there has been no infection linked to BPDCN, neither in empirical assessment of specific viruses, such as Epstein-Barr virus, nor in unbiased microbial metagenomics using next-generation sequencing.

Pathobiology

pDCs develop from a common myeloid-dendritic progenitor that is found in the bone marrow. After maturation, they circulate as 0.5% to 1% of peripheral blood mononuclear cells and are also found in secondary lymphoid organs such as lymph nodes, tonsil, and spleen. They are not frequently detected in the skin under normal circumstances. It is not clear if a specific stage of hematopoietic stem cells, pDC progenitors, or mature pDCs is the target for transformation to BPDCN. Similarly, we do not know if the initial transformation event takes place in bone marrow, peripheral blood, lymphoid organs, or skin. Investigation of these questions regarding the developmental ontogeny of BPDCN is an active area of ongoing research.

At least 50% of BPDCNs have an abnormal karyotype by cytogenetic analysis. However, there is no consistent translocation, mutation, deletion, or amplification event that defines all cases of BPDCN. Recurrent copy loss of specific chromosomes is a common feature, including minimal deleted regions that include known tumor suppressor genes, many of which are cell cycle regulators: 7p12 ( IKZF1 ), 9p21 ( CDKN2A / CDKN2B ), 12p13 ( CDKN1B ), 13q11 ( LATS2 ), 13q13 ( RB1 ), and 17p ( TP53 ). Rearrangements of 8q24 involving the MYC gene were reported in 40% of cases in a Japanese BPDCN population, with at least half only detectable using fluorescence in situ hybridization (FISH). In that series, MYC rearrangement was associated with poorer overall survival (OS) in univariate analysis. In a subset, the translocation t(6;8)(p21;q24) positions an enhancer region of the dendritic lineage-specific gene RUNX2 on 6p21 upstream of the MYC gene, driving its high expression. Despite these findings, in our center’s BPDCN population, we have not seen these high rates of MYC rearrangement, possibly related to differences in North American and Japanese populations. Therefore, we do not routinely perform FISH for MYC abnormalities. Pediatric BPDCN may also be distinct cytogenetically as it is enriched compared to adults for translocations involving the MYB locus and does not have other adult-type BPDCN-associated gene mutations.

Somatic mutations acquired in BPDCN cells overlap with the genes mutated in myeloid malignancies, supporting a similar cell of origin. The affected genes are particularly overlapping with those observed in patients with MDS and CMML, and include TET2 , ASXL1 , RNA splicing factors (e.g., SF3B1 , SRSF2 , U2AF1 , ZRSR2 ), TP53 , and others mutated in AML, including IDH1 , IDH2 , KRAS , and NRAS . Some series identified mutations in transcription factors that may be involved in pDC lineage-related transformation, including IKZF1 , IKZF2 , IKZF3 , and ZEB2 . Unlike AML, mutations in DNMT3A , FLT3 , and NPM1 are not common.

Gene expression profiling, including by single cell RNA sequencing, support that the closest normal counterpart of BPDCN is the pDC or its immediate precursors. BPDCNs also express high levels of the antiapoptotic gene BCL2 and have upregulated expression of nuclear factor kappa B (NF-κB) gene signatures, which may suggest targets for therapy.

Clinical Manifestations and Recommended Diagnostic Evaluation

The most striking and unique feature of BPDCN is its involvement of the skin. At least 80% of patients have skin involvement at the time of initial diagnosis. The skin lesions can be single, or widespread and numerous; they can be nodular and indurated or can be patchy erythema that mimics inflammatory or allergic conditions. However, the most common skin presentation is one or more violaceous, bruise-like plaques or tumors that can be found in any region of the body ( Fig. 64.1 ). The disease usually begins with a single lesion, which often then spreads to contiguous areas or more distantly in the skin. Interestingly, up to 50% of patients will not have concurrent bone marrow involvement (so-called “skin-only” disease). In those with bone marrow disease, only a minority will have circulating peripheral blood BPDCN cells. Anemia or thrombocytopenia are signs that the bone marrow may be involved. Lymph node and/or splenic involvement is present in approximately 40%. BPDCN has an increased propensity to involve other visceral sites such as the lung, breast, oral cavity, and other mucosa such as the cervix. In our experience, extramedullary involvement by BPDCN is more frequent than in AML, including compared to AMLs with monocytic differentiation (FAB M4 and M5), which can involve mucosal surfaces, gingiva, and other organs including the central nervous system.

We recommend that the initial evaluation for a patient with suspected BPDCN includes a full body skin exam, preferably by a dermatologist or oncologist with experience in skin malignancies, with skin biopsy of suspicious lesions; a bone marrow aspiration and biopsy; and full body imaging with either contrast-enhanced CT scan or positron emission tomography (PET)-CT scan.

Central nervous system (CNS) involvement by BPDCN is more frequent than in AML or other myeloid neoplasms such as MDS and CMML. Occult CNS involvement at diagnosis may be as high as 30%, and symptomatic CNS disease is often a feature of relapsed or refractory disease. Therefore, we recommend lumbar puncture for cytology and flow cytometry, plus prophylactic intrathecal chemotherapy for all patients at diagnosis, and for all patients with neurologic symptoms or signs at any time in their disease course. For patients with no CNS involvement at diagnosis, we recommend a full course of prophylactic intrathecal chemotherapy, concurrent with systemic therapy, following a regimen similar to that used for high-risk acute lymphoblastic leukemia.

Histopathology

Skin biopsy reveals a dense monomorphous infiltrate in the dermis and subcutis of blastoid cells ( Fig. 64.2 ). The epidermis is typically spared and there is usually a clear “grenz zone,” an uninvolved strip of papillary dermis just below the epidermis. The individual tumor cells are medium to large sized with scant cytoplasm and have similar appearance in skin, marrow, lymph nodes, blood, and other involved sites. Some have a plasmacytoid appearance with eccentric nuclei, cytoplasmic vacuolation, and/or extended cytoplasmic projections similar to “hand-mirror” cells that are seen in acute lymphoblastic leukemia. Mitotic figures are often present.

The classic immunophenotype of BPDCN cells is positivity for CD123 (IL-3 receptor α chain), CD4, and CD56, on CD45-positive cells indicating hematopoietic origin. The cells fall in the CD45 dim, side-scatter low “blast gate” on flow cytometry. Staining for the proto-oncogene T-cell leukemia/lymphoma protein 1 A (TCL-1) is often positive, although this marker is not specific for BPDCN and can also be expressed in several lymphoid malignancies. Terminal deoxynucleotidyl transferase (TdT) is positive in approximately one-third of cases. The pDC antigen CD303 (BDCA-2) is lost in some BPDCNs, but when positive, its expression supports the diagnosis. CD4 or CD56 may be negative in a minority (approximately 8%) of cases. One diagnostic schema suggests that BPDCN is likely if the tumor cells express at least 4 of 5 markers among CD123, CD4, CD56, TCL-1, and CD303. CD34 is usually negative. Myeloperoxidase and lysozyme are negative in BPDCN and when either is positive makes AML more likely.

Differential Diagnosis