BPDCN is a rare disease characterized by heterogeneous clinical features, complex diagnostic processes, and numerous changes in nomenclature and classification over the decades. Therefore, there is a lack of knowledge regarding BPDCN. In the United States, BPDCN account for only 0.44% of all hematological malignancies, 0.7% of all cutaneous leukemias/lymphomas, and 0.04 cases per 100,000 [11]. There is an extreme male predominance of 2–4:1 and a median age of 53 years [12]; a bimodal pattern was observed in older (>60 yr) and younger (<20 yr) individuals [12]. Additionally, the male predominance has been associated with ZRSR2 mutations on the X chromosome [13].

A BPDCN is a clinically aggressive myeloid malignancy that involves the skin, bone marrow, lymphatic organs, and central nervous system (CNS)/cerebrospinal fluid (CSF) [14, 15]. Most patients (80%) have skin involvement and 50% have only skin lesions. Skin lesions range from solitary brown or violaceous bruise-like lesions, erythematous patches, plaques, and ulcerative lesions to disseminated cutaneous spread [16, 17]. Most patients present with localized skin lesions that progressively spread throughout the body, whereas some patients initially present with leukemia without any skin lesions. More than 60% of patients have bone marrow involvement and diffuse involvement, and progressive cytopenia is common [16, 18]. More than 20% of patients have a prior or concomitant myeloid malignancy, such as AML, MDS, or CMML [15, 16]. BPDCN rapidly disseminate in the bone marrow, lymph nodes, liver, spleen, and CNS. In one study, flow cytometric analysis identified CSF positivity in 60% of patients with BPDCN at diagnosis and in 100% of patients at relapse [19]. Thus, the clinical manifestations of BPDCN are highly variable and often misdiagnosed or delayed. A BPDCN has a poor prognosis, with a median overall survival (OS) of 2 years [16].

BPDCN has no known environmental or hereditary genetic factors that contribute to its development, although it frequently exhibits abnormal karyotypes on 5q, 12p, 17p, 13q, 6q, 15q, and 9 monosomy [20]. The 9p21.3 deletion is associated with a poor prognosis [21]. Furthermore, MYC, MYB, and MLL rearrangement is common in BPDCN [22-24]. In recent studies, BPDCN exhibited recurrent mutations in ASXL1, IDH1-2, IKZF1-3, NPM1, TET1-2, TP53, U2AF1, and ZEB2, as well as the overexpression of IncRNA-3q or TCF4 [25-27]. Genetic abnormalities are heterogeneous; therefore, their molecular mechanisms and prognostic significance remain controversial.

Its diagnostic workup is summarized in Table 1. The diagnosis of BPDCN is usually initiated with suspicion regarding patients with skin lesions or progressive pancytopenia. Although the shape, size, color, and pattern of BPDCN skin lesions vary, their morphological features and immunohistochemical expression patterns are typical. Immature blastic cells infiltrate deeply into the dermis, sparing the epidermis and preventing angioinvasion, coagulative necrosis, or inflammatory cells [28]. These cells typically express CD4 and CD56, as well as at least one of the pDC-associated markers (CD123, CD303, CD304, TCL1, and TCF4) in the absence of lineage-specific markers for myeloids, T lymphocytes, B lymphocytes, or NK cells [28, 29]. Additionally, there is a pearl-necklace appearance of blast cells in bone marrow puncture smears with tiny vacuoles extending in a pseudopod-like pattern along the cytoplasmic membrane [29]. They are medium-sized cells with scant cytoplasm, fine chromatin, and irregular nucleoli [29]. However, there are a few cases in which is negative for CD4, CD56, or both.

The immunophenotypic diagnostic criteria for BPDCN are listed in Table 2 [4]. Clinical and morphological features, as well as immunophenotyping, should be considered when diagnosing BPDCN.

There is a need for a multidisciplinary approach to the diagnosis of BPDCN. As shown in Table 3, BPDCN can be distinguished from other CD56-positive hematologic malignancies that affect the skin, such as leukemia cutis, myeloid sarcoma, extranodal NK/T cell lymphoma, and cutaneous T cell lymphoma [29].

BPDCN can occur as an isolated disease or in the context of other hematologic neoplasms. However, the relationship between BPDCN and other myeloid malignancies has not yet been elucidated. Therefore, the differential diagnosis of BPDCN is necessary when pDCs are present in patients with myeloid malignancies or ambiguous immunophenotypes. This is based on CD56 positivity and the presence of pDC markers (Fig. 1) [5]. Compared to BPDCN, pDC-AML exhibits RUNX1 mutations, fewer skin lesions, and expresses CD34-positive and CD56-negative pDCs [8]. Furthermore, NPM1-mutated AML with a pDC-like phenotype expresses CD123-positive and other pDC marker-negative cells and has poor outcomes [30].

There is currently no standard treatment for BPDCN. Historically, BPDCN treatments have been based on multi-agent chemotherapy regimens for lymphoma, acute lymphoblastic leukemia, and AML [28]. In addition, acute leukemia regimens achieve high complete response (CR) rates ranging from 40–90% and allogeneic hematopoietic cell transplantation (allo-HCT) can result in durable remission in some patients. However, their rarity and heterogeneity make it difficult to determine the most effective therapeutic strategies [11]. Owing to recent advances in molecular biology and genetics, targeted treatment strategies have been developed. In December 2018, the FDA approved tagraxofusp, a first-in-class CD123-targeting therapy for treatment-naïve or relapsed/refractory BPDCN [31]. However, unfit, relapsed, or refractory patients continue to require effective therapeutic strategies.

The most common induction regimens for BPDCN are CHOP (cyclophosphamide, doxorubicin, vincristine, prednisone) hyper-CVAD (hyperfractionated cyclophosphamide, vincristine, doxorubicin, and dexamethasone alternating with high-dose methotrexate and cytarabine); high doses of methotrexate combined with asparaginase; and the 7+3 regimen (cytarabine combined with idarubicin or daunorubicin) [32]. Several retrospective studies have revealed better outcomes with lymphoid-type regimens than with myeloid regimens; however, only allo-HCT was associated with prolonged OS [15, 16]. Recently, the French BPDCN network and the German Society of Hematology and Oncology recommended acute leukemia regimens rather than CHOP or CHOP-like regimens [14, 32]. A prospective phase II trial to evaluate the effectiveness of three cycles of methotrexate, L-asparaginase, idarubicin, and dexamethasone, followed by allo-HCT, for eligible patients is ongoing (NCT03599960).

Allo-HCT is the standard consolidation treatment for fit patients with BPDCN who achieve CR1 [33, 34]. Based on a meta-analysis of 128 patients who underwent allo-HCT, the pooled OS was 70% in CR1 and 7% beyond CR1 [35]. However, no randomized controlled trials have compared allo-HCT with other therapies, especially in the era of novel therapies. A North American collaborative study revealed similar outcomes between myeloablastive conditioning (MAC) and reduced-intensity conditioning regimens [36]. Data from the EBMT registry indicate MAC based on TBI results in long-term remission after allo-HCT for BPDCN [33]. Due to limited data, autologous HCT may lack efficacy against BPDCN [36]. However, patients who are ineligible for allo-HCT may benefit from autologous HCT [33, 37].

Tagraxofusp, an interleukin (IL) 3-diphtheria toxin fusion protein, targets the IL3 receptor alpha subunit or CD123, which is highly expressed on BPDCN cells [38]. In patients with treatment-naïve BPDCN, tagraxofusp resulted in a 75% overall response (CR: 57%) with a median OS of 15.8 months (range, 9.7 to 25.8) [39]. In the previously treated group, 67% of the patients achieved an overall response with a median OS of 8.5 months [31, 39]. The median response durations for the treatment-naïve and previously treated BPDCN groups were 24.9 months (range, 3.8 to not reached) and 2.8 months (range, 0.7 to 14.0), respectively [31]. The most common adverse events were the elevation of liver enzymes, hypoalbuminemia, thrombocytopenia, peripheral edema, and capillary leak syndrome (CLS) [31]. CLS was the most serious, resulting in three deaths [31]. CLS is treated by holding tagraxofusp and administering intravenous albumin or steroids and volume control [39]. In addition to tagraxofusp (a recombinant fusion protein targeting CD123), CD28/4-1BB CD123 CAR T cells, UCART123 (anti-CD123 CAR-T cells), XmAb14045 (a CD123/CD3 bispecific antibody), and IMGN632 (CD123 antibody-drug conjugates) have been investigated [24, 40-42]. IMGN632 is a humanized monoclonal antibody that targets CD123 fused to a potent DNA-alkylating agent. A phase I clinical trial of IMGN632 in CD123-positive hematologic malignancies is currently ongoing (NCT03386513).

BPDCN highly express anti-apoptotic protein B cell leukemia/lymphoma-2 (BCL-2) [43]. Several cases have revealed the efficacy of venetoclax, a BCL-2 inhibitor [43, 44]. In addition, epigenetic alterations are common in BPDCN, and the efficacy of 5-azacytidine (5-Aza) in patients with BPDCN has been reported [45]. However, monotherapy does not demonstrate long-term efficacy. Venetoclax in combination with hypomethylating agents or tagraxofusp for BPDCN has been reported to be synergistic [46, 47]. Therefore, several clinical trials have targeted these mechanisms, including those using 5-Aza, venetoclax, and tagraxofusp.

Several case reports have suggested a potential role of myeloma-based therapy in the treatment of BPDCN. Bortezomib for the constitutive activation of the NF-κB pathway, lenalidomide for IKZF1 depletion, and daratumumab, which targets CD38 overexpression in BPDCN, were used [48-50]. Marmouset et al. [49] described a promising combination of bortezomib/lenalidomide and dexamethasone in 2019; Mirgh et al. [51] demonstrated the efficacy of daratumumab in combination with bortezomib in CD38-positive BPDCN.

The genetic landscape of BPDCN is characterized by recurrent MYB/MYC rearrangements, epigenetic alterations, the downregulation of the liver X receptor (LXR), and the dysregulation of the TCF4-/BRD4-dependent transcriptional network [24, 52]. NF-κB inhibitors, BRD4 inhibitors, and LXR agonists are potential therapeutic targets for BPDCN. Furthermore, the use of bromodomain and extra-terminal motif inhibitors (BETi), which antagonize the TCF4 amplification loop, can induce blastic pDC apoptosis [26]. Further research is required to determine the most effective combination or sequential therapies for BPDCN, including allo-HCT and novel agents.