2  Diagnosis of Blastic Plasmacytoid Dendritic Cell Neoplasm

A Visual Guide to Clinical Presentation, Immunophenotype, Genomic Landscape & Differential Diagnosis

Author
Affiliation

Based on Kharfan-Dabaja MA, Lane AA, Pemmaraju N, et al.

Blood 2025; 145:567-574 | Luskin & Lane, Haematologica 2024 | Pemmaraju N, et al. J Clin Oncol 2026

Published

August 19, 2026

NoteKey Diagnostic Principles at a Glance

BPDCN is an ultra-rare (<0.5% of hematologic malignancies), aggressive neoplasm derived from plasmacytoid dendritic cell precursors. Diagnosis requires a high index of suspicion and is confirmed by identifying blasts expressing CD123 + CD4 and/or CD56 with at least one additional pDC marker (TCF4, TCL1, CD303, CD304), while excluding T-cell, B-cell, and myeloid lineage markers. Remember the mnemonic: “123-4-56” for CD123, CD4, CD56.

3 Background & Historical Context

  • Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is an aggressive hematologic malignancy derived from precursors of plasmacytoid dendritic cells (pDCs), listed as a separate entity in the 5th edition of the WHO classification of hematolymphoid tumors.
  • The disease has undergone a series of name and classification changes over decades — previously termed CD4⁺CD56⁺ hematodermic neoplasm, blastic NK-cell lymphoma, and agranular CD4⁺ natural killer cell leukemia — reflecting historical confusion before its true pDC lineage was established.
  • BPDCN constitutes approximately 0.44% of all hematologic malignancies (~0.05 cases per 100,000), but this likely represents an underestimation due to heterogeneous presentation and diagnostic challenges.
  • The approval of tagraxofusp (2018) and ongoing development of pivekimab sunirine have made accurate diagnosis imperative, as CD123-directed therapies require precise identification of this entity.

3.1 Historical Nomenclature

Show code 
library(ggplot2)

nomenclature <- data.frame(
  era = factor(c(
    "1990s",
    "2001",
    "2005",
    "2008\n(WHO 4th Ed.)",
    "2016\n(WHO Revised)",
    "2022\n(WHO 5th Ed.)"
  ), levels = c(
    "1990s",
    "2001",
    "2005",
    "2008\n(WHO 4th Ed.)",
    "2016\n(WHO Revised)",
    "2022\n(WHO 5th Ed.)"
  )),
  name = c(
    "Agranular CD4+\nNK-cell leukemia",
    "Blastic NK-cell\nlymphoma",
    "CD4+CD56+\nhematodermic\nneoplasm",
    "Blastic plasmacytoid\ndendritic cell\nneoplasm",
    "BPDCN\n(AML-related\nprecursor neoplasm)",
    "BPDCN\n(Separate entity)"
  ),
  y = rep(1, 6),
  insight = c(
    "NK lineage\nassumed",
    "Reclassified\nas lymphoma",
    "Skin tropism\nrecognized",
    "pDC origin\nestablished",
    "Grouped with\nmyeloid neoplasms",
    "Standalone\nclassification"
  )
)

color_gradient <- c("#d6604d", "#d6604d", "#f4a582", "#4393c3", "#2166ac", "#2166ac")

ggplot(nomenclature, aes(x = era, y = y)) +
  geom_segment(
    data = data.frame(
      x = levels(nomenclature$era)[1:5],
      xend = levels(nomenclature$era)[2:6],
      y = 1, yend = 1
    ),
    aes(x = x, xend = xend, y = y, yend = yend),
    inherit.aes = FALSE,
    arrow = arrow(length = unit(0.2, "cm"), type = "closed"),
    color = "#888888", linewidth = 1
  ) +
  geom_point(size = 16, color = color_gradient) +
  geom_text(aes(label = gsub("\\n.*", "", as.character(era))),
            size = 2.8, fontface = "bold", color = "white") +
  geom_text(aes(label = name, y = y + 0.18), size = 2.5, color = "#333333",
            lineheight = 0.85) +
  geom_text(aes(label = insight, y = y - 0.15), size = 2.3, color = "#777777",
            fontface = "italic", lineheight = 0.85) +
  scale_y_continuous(limits = c(0.65, 1.35)) +
  labs(
    title = "Evolution of BPDCN Nomenclature",
    subtitle = "From NK-cell leukemia to a distinct WHO entity — three decades of reclassification"
  ) +
  theme_void(base_size = 13) +
  theme(
    plot.title = element_text(hjust = 0.5, face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(hjust = 0.5, size = 11, color = "#555555",
                                 margin = margin(b = 15)),
    plot.margin = margin(20, 20, 20, 20)
  )
Figure 3.1: Evolution of BPDCN nomenclature and classification over three decades, reflecting progressive understanding of the disease’s plasmacytoid dendritic cell origin.

4 Clinical Presentation

4.1 Demographics

Show code 
library(ggplot2)
library(dplyr)
library(tidyr)

demo_data <- tibble(
  Characteristic = c(
    "Median age ≥65 years",
    "Male sex",
    "White race",
    "ECOG PS 0",
    "ECOG PS 1",
    "Skin involvement",
    "Bone marrow",
    "Lymph nodes",
    "Peripheral blood",
    "CNS (occult or overt)"
  ),
  `Tagraxofusp\nPivotal (n=47)` = c(83, 81, 91, 47, 53, 94, 51, 45, 17, 12),
  `PVEK CADENZA\n(N=84)` = c(71, 82, 82, 36, 61, 77, 48, 36, 16, 8),
  `French BPDCN\nNetwork` = c(73, 77, 88, NA, NA, 84, 51, 45, 45, 15)
)

demo_long <- demo_data |>
  pivot_longer(-Characteristic, names_to = "Cohort", values_to = "Percentage") |>
  filter(!is.na(Percentage)) |>
  mutate(
    Characteristic = factor(Characteristic, levels = rev(demo_data$Characteristic)),
    Cohort = factor(Cohort, levels = c(
      "Tagraxofusp\nPivotal (n=47)",
      "PVEK CADENZA\n(N=84)",
      "French BPDCN\nNetwork"
    ))
  )

cohort_colors <- c(
  "Tagraxofusp\nPivotal (n=47)" = "#2166ac",
  "PVEK CADENZA\n(N=84)" = "#d6604d",
  "French BPDCN\nNetwork" = "#4daf4a"
)

ggplot(demo_long, aes(x = Percentage, y = Characteristic, fill = Cohort)) +
  geom_col(position = position_dodge(width = 0.8), width = 0.7, alpha = 0.9) +
  geom_text(
    aes(label = paste0(Percentage, "%")),
    position = position_dodge(width = 0.8),
    hjust = -0.15, size = 2.8, fontface = "bold"
  ) +
  scale_fill_manual(values = cohort_colors) +
  scale_x_continuous(
    limits = c(0, 110),
    breaks = seq(0, 100, 25),
    labels = function(x) paste0(x, "%")
  ) +
  labs(
    title = "Baseline Patient Characteristics at BPDCN Diagnosis",
    subtitle = "Cross-study comparison of demographic and disease involvement patterns",
    x = "Patients (%)",
    y = NULL,
    fill = "Study Cohort"
  ) +
  theme_minimal(base_size = 13) +
  theme(
    plot.title = element_text(face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(size = 12, color = "#555555", margin = margin(b = 15)),
    legend.position = "bottom",
    legend.title = element_text(face = "bold"),
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank(),
    axis.text.y = element_text(size = 11),
    plot.margin = margin(15, 30, 15, 15)
  )
Figure 4.1: Key demographic and clinical features of BPDCN at diagnosis, synthesized from published cohorts (Kharfan-Dabaja et al., Blood 2025; Pemmaraju et al., NEJM 2019; Pemmaraju et al., JCO 2026).
ImportantSkin Is the Sentinel Organ

Skin involvement occurs in 65–94% of patients and is the most common presenting feature. Characteristic lesions include dark purple or violaceous nodules, plaques, and tumors — often initially in sun-exposed areas. BPDCN has been misdiagnosed as melanoma, leprosy, and even dermatomyositis. Importantly, tumor phylogenomics shows that premalignant clones in the bone marrow give rise to skin tumors, highlighting that even seemingly “localized” skin disease is a systemic malignancy.

4.2 Disease Compartments at Diagnosis

Show code 
import matplotlib.pyplot as plt
import numpy as np

compartments = ['Skin', 'Bone Marrow', 'Lymph Nodes', 'Peripheral Blood', 'CNS', 'Viscera']
percentages = [84, 51, 45, 45, 15, 14]

fig, ax = plt.subplots(figsize=(10, 5.5))

colors = ['#2166ac', '#4393c3', '#67a9cf', '#92c5de', '#d1e5f0', '#e0e0e0']

bars = ax.barh(
    range(len(compartments)), percentages,
    color=colors, edgecolor='white', linewidth=1.5, height=0.6
)

for bar, pct in zip(bars, percentages):
    ax.text(
        bar.get_width() + 1.5, bar.get_y() + bar.get_height()/2,
        f'{pct}%', va='center', ha='left', fontsize=13, fontweight='bold',
        color='#333333'
    )

ax.set_yticks(range(len(compartments)))
ax.set_yticklabels(compartments, fontsize=12, fontweight='bold')
ax.set_xlim(0, 100)
(0.0, 100.0)
Show code 
ax.set_xlabel('Patients (%)', fontsize=12)
ax.set_title(
    'BPDCN Disease Compartments at Diagnosis',
    fontsize=16, fontweight='bold', color='#1a5276', pad=15
)

ax.invert_yaxis()
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
ax.spines['left'].set_visible(False)
ax.tick_params(left=False)
ax.grid(axis='x', alpha=0.3, linestyle='--')

fig.text(0.5, -0.02,
         'Data synthesized from French BPDCN Network (Garnache-Ottou et al., Blood Adv 2019) and Kharfan-Dabaja et al., Blood 2025.',
         ha='center', fontsize=8, fontstyle='italic', color='#888888')

plt.tight_layout()
plt.show()
Figure 4.2: Multi-compartment involvement in BPDCN at diagnosis. BPDCN commonly presents with simultaneous involvement of skin, bone marrow, lymph nodes, peripheral blood, and CNS.

4.3 Hematologic Findings at Diagnosis

Show code 
library(ggplot2)

heme_data <- data.frame(
  Finding = factor(c(
    "Blast infiltration\n(bone marrow)",
    "Thrombocytopenia",
    "Anemia",
    "Cytogenetic\nabnormality",
    "Neutropenia",
    "Dysplastic features"
  ), levels = c(
    "Blast infiltration\n(bone marrow)",
    "Thrombocytopenia",
    "Anemia",
    "Cytogenetic\nabnormality",
    "Neutropenia",
    "Dysplastic features"
  )),
  Percentage = c(96, 75, 64, 60, 38, 31),
  Category = c("Marrow", "Cytopenias", "Cytopenias", "Genetics", "Cytopenias", "Morphology")
)

cat_colors <- c(
  "Marrow" = "#2166ac",
  "Cytopenias" = "#d6604d",
  "Genetics" = "#4daf4a",
  "Morphology" = "#984ea3"
)

ggplot(heme_data, aes(x = Finding, y = Percentage, fill = Category)) +
  geom_col(width = 0.6, alpha = 0.9) +
  geom_text(aes(label = paste0(Percentage, "%")),
            vjust = -0.5, fontface = "bold", size = 5, color = "#333333") +
  scale_fill_manual(values = cat_colors, name = "Category") +
  scale_y_continuous(limits = c(0, 110), breaks = seq(0, 100, 25),
                     labels = function(x) paste0(x, "%")) +
  labs(
    title = "Hematologic Findings at BPDCN Diagnosis",
    subtitle = "French BPDCN Network — Peripheral blood and bone marrow abnormalities",
    x = NULL,
    y = "Patients (%)"
  ) +
  theme_minimal(base_size = 14) +
  theme(
    plot.title = element_text(face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(size = 11, color = "#555555", margin = margin(b = 15)),
    legend.position = "bottom",
    legend.title = element_text(face = "bold"),
    panel.grid.major.x = element_blank(),
    panel.grid.minor = element_blank(),
    axis.text.x = element_text(size = 10, face = "bold"),
    plot.margin = margin(15, 20, 15, 15)
  )
Figure 4.3: Hematologic abnormalities reported at BPDCN diagnosis from the French BPDCN Network (Garnache-Ottou et al., Blood Adv 2019).

5 Immunophenotypic Diagnosis

The cornerstone of BPDCN diagnosis is its distinctive immunophenotype. Two complementary diagnostic frameworks are currently in use: the WHO 5th Edition criteria and the NCCN guidelines (v3.2024).

Criterion Markers Requirement
Pathway A CD123 + CD4 and/or CD56 + ≥1 pDC marker Expression of CD123, CD4/CD56, and ≥1 of: TCF4, TCL1, CD303, CD304
Pathway B ≥3 pDC markers without lineage markers Any 3 of: CD123, TCF4, TCL1, CD303, CD304 — AND absence of CD3, CD14, CD19, CD34
Expected negative Lineage exclusion markers CD3 (T-cell), CD19 (B-cell), CD14 (monocytic), lysozyme, myeloperoxidase
Criterion Markers Requirement
Positive markers ≥4 of 6 BPDCN antigens CD4, CD56, CD123, TCL1, CD2AP, CD303/BDCA-2
Negative markers Myeloid lineage exclusion MPO, lysozyme, CD14, CD34, CD116, CD163
Negative markers Lymphoid lineage exclusion T-cell and B-cell markers
Marker Expression Level
CD4 Always expressed High
CD123 Always expressed High
HLA-DR Always expressed High
TCL1 Always expressed High
CD56 Very often expressed Variable
CD303/BDCA-2 Very often expressed Variable
CD304 Very often expressed Variable
CD36, CD38, CD45RA Very often expressed Variable

5.1 Immunophenotypic Diagnostic Algorithm

Show code 
library(ggplot2)

steps <- data.frame(
  step = factor(c(
    "1. Suspect\nBPDCN",
    "2. Confirm\nCD123+",
    "3. Check\nCD4/CD56",
    "4. Add pDC\nMarkers",
    "5. Exclude\nLineage",
    "6. Diagnosis\nConfirmed"
  ), levels = c(
    "1. Suspect\nBPDCN",
    "2. Confirm\nCD123+",
    "3. Check\nCD4/CD56",
    "4. Add pDC\nMarkers",
    "5. Exclude\nLineage",
    "6. Diagnosis\nConfirmed"
  )),
  y = rep(1, 6),
  phase = c("Clinical", "Core", "Core", "Extended", "Exclusion", "Final"),
  detail = c(
    "Skin lesions\nCytopenias\nBlasts",
    "Universal in\nBPDCN\n(IL3Rα)",
    "CD4+ and/or\nCD56+",
    "TCF4, TCL1\nCD303, CD304",
    "CD3−, CD19−\nCD14−, MPO−",
    "BPDCN\nconfirmed"
  )
)

phase_colors <- c(
  "Clinical" = "#f4a582",
  "Core" = "#2166ac",
  "Extended" = "#4393c3",
  "Exclusion" = "#d6604d",
  "Final" = "#1a5276"
)

ggplot(steps, aes(x = step, y = y)) +
    geom_segment(
    data = data.frame(
      x = levels(steps$step)[1:5],
      xend = levels(steps$step)[2:6],
      y = 1, yend = 1
    ),
    aes(x = x, xend = xend, y = y, yend = yend),
    inherit.aes = FALSE,
    arrow = arrow(length = unit(0.3, "cm"), type = "closed"),
    color = "#888888", linewidth = 1.2
  ) +
  geom_point(aes(color = phase), size = 20) +
  geom_text(aes(label = gsub("\\d+\\. ", "", gsub("\n.*", "", as.character(step)))),
            size = 2.8, fontface = "bold", color = "white") +
  geom_text(aes(label = detail, y = y - 0.18),
            size = 2.5, color = "#444444", lineheight = 0.85) +
  scale_color_manual(values = phase_colors, name = "Phase") +
  scale_y_continuous(limits = c(0.65, 1.25)) +
  labs(
    title = "BPDCN Diagnostic Algorithm — Stepwise Immunophenotyping",
    subtitle = "Remember \"123-4-56\": CD123 (universal) → CD4 + CD56 → pDC markers → Lineage exclusion"
  ) +
  theme_void(base_size = 13) +
  theme(
    plot.title = element_text(hjust = 0.5, face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(hjust = 0.5, size = 11, color = "#555555",
                                 margin = margin(b = 15)),
    legend.position = "bottom",
    plot.margin = margin(20, 25, 20, 25)
  )
Figure 5.1: Stepwise immunophenotypic approach to BPDCN diagnosis. The mnemonic ‘123-4-56’ captures the core markers: CD123, CD4, CD56. Dual expression of TCF4 and CD123 is reported to be highly sensitive and specific for BPDCN (Sukswai et al., Am J Surg Pathol 2019).

5.2 Marker Expression Landscape

Show code 
library(ggplot2)
library(dplyr)
library(tidyr)

markers <- tibble(
  Marker = c(
    "CD123", "CD4", "TCL1", "HLA-DR",
    "CD56", "CD303", "CD304", "TCF4",
    "CD36", "CD38", "CD45RA", "CD7",
    "CD3", "CD19", "CD14", "CD34",
    "MPO", "Lysozyme", "TdT"
  ),
  Category = c(
    rep("Core pDC", 4),
    rep("Extended pDC", 4),
    rep("Often Positive", 3),
    "Variable",
    rep("Expected Negative", 6),
    "Variable"
  ),
  Expression = c(
    99, 98, 95, 97,
    85, 80, 75, 90,
    70, 75, 65, 30,
    2, 2, 3, 5,
    1, 2, 15
  ),
  Status = c(
    rep("Always+", 4),
    rep("Usually+", 4),
    rep("Often+", 3),
    "Variable",
    rep("Negative", 6),
    "Variable"
  )
)

markers <- markers |>
  mutate(Marker = factor(Marker, levels = rev(Marker)))

fill_colors <- c(
  "Always+" = "#2166ac",
  "Usually+" = "#67a9cf",
  "Often+" = "#d1e5f0",
  "Variable" = "#f7f7f7",
  "Negative" = "#f4a582"
)

ggplot(markers, aes(x = Expression, y = Marker, fill = Status)) +
  geom_col(width = 0.7, alpha = 0.9) +
  geom_text(
    aes(label = ifelse(Expression >= 10,
                       paste0(Expression, "%"),
                       paste0("<", Expression + 3, "%"))),
    hjust = -0.15, size = 3.2, fontface = "bold", color = "#333333"
  ) +
  geom_vline(xintercept = 50, linetype = "dashed", color = "grey50", linewidth = 0.4) +
  annotate("text", x = 52, y = 19.5, label = "50%", hjust = 0, size = 3, color = "grey40") +
  scale_fill_manual(values = fill_colors, name = "Expression Pattern") +
  scale_x_continuous(limits = c(0, 115), breaks = seq(0, 100, 25),
                     labels = function(x) paste0(x, "%")) +
  labs(
    title = "BPDCN Immunophenotypic Marker Landscape",
    subtitle = "Hierarchical expression patterns — Core pDC markers are universally expressed",
    x = "Approximate Expression Frequency (%)",
    y = NULL
  ) +
  theme_minimal(base_size = 13) +
  theme(
    plot.title = element_text(face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(size = 11, color = "#555555", margin = margin(b = 15)),
    legend.position = "bottom",
    legend.title = element_text(face = "bold"),
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank(),
    axis.text.y = element_text(size = 11, face = "bold"),
    plot.margin = margin(15, 30, 15, 15)
  )
Figure 5.2: Immunophenotypic marker expression landscape in BPDCN. Green = consistently positive; red = consistently negative (expected). The combination of CD123+/CD4+/CD56+ with negative lineage markers is pathognomonic.
NoteTCF4 + CD123: A Highly Sensitive & Specific Combination

Dual expression of TCF4 and CD123 has been reported to be highly sensitive and specific for BPDCN, providing a practical two-marker screen for cases where the diagnosis is suspected (Sukswai et al., Am J Surg Pathol 2019). This combination can be particularly valuable in differentiating BPDCN from AML with leukemia cutis.

6 Differential Diagnosis

One of the most critical aspects of BPDCN diagnosis is distinguishing it from morphologic mimics — both hematopoietic and non-hematopoietic neoplasms that can present with similar skin involvement or immunophenotypic overlap.

6.1 Differential Diagnosis Table

Show code 
import matplotlib.pyplot as plt
import numpy as np

fig, (ax1, ax2) = plt.subplots(1, 2, figsize=(14, 7))

# Panel A: Hematopoietic differentials
heme_entities = [
    'AML with\nleukemia cutis',
    'Mature pDC\nproliferation',
    'T-ALL',
    'B-ALL',
    'NK/T-cell\nlymphoma',
    'ATLL',
    'ALCL'
]
heme_key = [
    'MPO+, CD34+\nCD123 variable',
    'CD34+, CD56−\nMature morphology',
    'CD3+, TdT+\ncytoplasmic CD3',
    'CD19+, CD10+\nTdT+',
    'CD3ε+, CD56+\nEBV+ (EBER)',
    'CD4+, CD25+\nHTLV-1+',
    'CD30+, ALK±\nCD4−'
]

colors_heme = ['#2166ac'] * len(heme_entities)
y_pos = range(len(heme_entities))

for i, (entity, key) in enumerate(zip(heme_entities, heme_key)):
    ax1.barh(i, 1, color='#2166ac', alpha=0.15, height=0.7)
    ax1.text(0.02, i, entity, va='center', ha='left', fontsize=10,
             fontweight='bold', color='#1a5276')
    ax1.text(0.55, i, key, va='center', ha='left', fontsize=9,
             color='#555555', linespacing=1.3)

ax1.set_xlim(0, 1)
(0.0, 1.0)
Show code 
ax1.set_ylim(-0.5, len(heme_entities) - 0.5)
(-0.5, 6.5)
Show code 
ax1.invert_yaxis()
ax1.set_title('Hematopoietic Neoplasms', fontsize=14, fontweight='bold',
              color='#1a5276', pad=10)
ax1.axis('off')
(np.float64(0.0), np.float64(1.0), np.float64(6.5), np.float64(-0.5))
Show code 
# Panel B: Non-hematopoietic differentials
non_heme_entities = [
    'Melanoma',
    'Merkel cell\ncarcinoma',
    'Angiosarcoma',
    'Kaposi sarcoma',
    'Dermatomyositis\n(mimicry)'
]
non_heme_key = [
    'S100+, HMB-45+\nMelan-A+',
    'CK20+, CD56+\nNeuroendocrine',
    'CD31+, CD34+\nERG+',
    'HHV-8+\nVascular',
    'Skin rash pattern\nAutoantibodies+'
]

for i, (entity, key) in enumerate(zip(non_heme_entities, non_heme_key)):
    ax2.barh(i, 1, color='#d6604d', alpha=0.15, height=0.7)
    ax2.text(0.02, i, entity, va='center', ha='left', fontsize=10,
             fontweight='bold', color='#8b1a1a')
    ax2.text(0.55, i, key, va='center', ha='left', fontsize=9,
             color='#555555', linespacing=1.3)

ax2.set_xlim(0, 1)
(0.0, 1.0)
Show code 
ax2.set_ylim(-0.5, len(non_heme_entities) - 0.5)
(-0.5, 4.5)
Show code 
ax2.invert_yaxis()
ax2.set_title('Non-Hematopoietic Neoplasms', fontsize=14, fontweight='bold',
              color='#8b1a1a', pad=10)
ax2.axis('off')
(np.float64(0.0), np.float64(1.0), np.float64(4.5), np.float64(-0.5))
Show code 
plt.suptitle('Differential Diagnosis of BPDCN — Key Distinguishing Features',
             fontsize=16, fontweight='bold', color='#1a5276', y=1.02)

plt.tight_layout()
plt.show()
Figure 6.1: Differential diagnosis of BPDCN categorized by hematopoietic and non-hematopoietic neoplasms, with key distinguishing features.
WarningMature pDC Proliferation ≠ BPDCN

Mature plasmacytoid dendritic cell proliferation — frequently associated with CMML, CML, or AML — must be distinguished from BPDCN. Key differentiators include: mature pDC proliferations are generally CD34-positive and CD56-negative, and cells appear morphologically mature. These proliferations are associated with an underlying myeloid neoplasm and do not warrant BPDCN-directed therapy.

6.2 BPDCN vs. AML with Leukemia Cutis

Show code 
library(ggplot2)
library(dplyr)
library(tidyr)

comparison <- tibble(
  Marker = factor(c("CD123", "CD4", "CD56", "TCL1", "TCF4",
                     "CD34", "MPO", "CD14", "Lysozyme"),
                  levels = c("CD123", "CD4", "CD56", "TCL1", "TCF4",
                             "CD34", "MPO", "CD14", "Lysozyme")),
  BPDCN = c(99, 98, 85, 95, 90, 5, 1, 3, 2),
  AML_Cutis = c(60, 30, 15, 5, 5, 70, 65, 40, 35)
)

comp_long <- comparison |>
  pivot_longer(-Marker, names_to = "Disease", values_to = "Expression") |>
  mutate(Disease = ifelse(Disease == "AML_Cutis", "AML with\nLeukemia Cutis", "BPDCN"))

disease_colors <- c("BPDCN" = "#2166ac", "AML with\nLeukemia Cutis" = "#d6604d")

ggplot(comp_long, aes(x = Marker, y = Expression, fill = Disease)) +
  geom_col(position = position_dodge(width = 0.7), width = 0.6, alpha = 0.9) +
  geom_text(
    aes(label = paste0(Expression, "%")),
    position = position_dodge(width = 0.7),
    vjust = -0.5, size = 3.2, fontface = "bold"
  ) +
  scale_fill_manual(values = disease_colors, name = "Diagnosis") +
  scale_y_continuous(limits = c(0, 115), breaks = seq(0, 100, 25),
                     labels = function(x) paste0(x, "%")) +
  labs(
    title = "BPDCN vs. AML with Leukemia Cutis — Immunophenotypic Comparison",
    subtitle = "TCL1 and TCF4 are the strongest discriminators between these entities",
    x = NULL,
    y = "Approximate Expression (%)"
  ) +
  theme_minimal(base_size = 13) +
  theme(
    plot.title = element_text(face = "bold", size = 15, color = "#1a5276"),
    plot.subtitle = element_text(size = 11, color = "#555555", margin = margin(b = 15)),
    legend.position = "bottom",
    legend.title = element_text(face = "bold"),
    panel.grid.major.x = element_blank(),
    panel.grid.minor = element_blank(),
    axis.text.x = element_text(size = 11, face = "bold"),
    plot.margin = margin(15, 20, 15, 15)
  )
Figure 6.2: Key immunophenotypic differences between BPDCN and AML with leukemia cutis — the most common diagnostic conundrum.

7 Genomic Landscape

The mutational landscape of BPDCN is complex, with TET2 mutations being the most commonly reported abnormality. Importantly, ~20% of patients have a prior or concurrent hematologic malignancy (PCHM), often sharing a clonal origin with BPDCN.

7.1 Mutational Spectrum

Show code 
library(ggplot2)
library(dplyr)

mutations <- tibble(
  Gene = c("TET2", "ASXL1", "ZRSR2", "NRAS", "TP53",
           "IDH2", "SRSF2", "IKZF1", "ETV6", "ARID1A",
           "JAK2", "KRAS", "ATM", "BCOR", "ZEB2"),
  Pathway = c(
    "DNA Methylation", "Chromatin Remodeling", "Splicing", "RAS Signaling", "Tumor Suppressor",
    "DNA Methylation", "Splicing", "Transcription Factor", "Transcription Factor", "Chromatin Remodeling",
    "Cytokine Signaling", "RAS Signaling", "Tumor Suppressor", "Chromatin Remodeling", "Transcription Factor"
  ),
  Frequency = c(55, 30, 25, 20, 18,
                15, 14, 12, 10, 10,
                8, 7, 7, 6, 5)
) |>
  mutate(Gene = factor(Gene, levels = rev(Gene)))

pathway_colors <- c(
  "DNA Methylation" = "#2166ac",
  "Chromatin Remodeling" = "#4393c3",
  "Splicing" = "#67a9cf",
  "Transcription Factor" = "#984ea3",
  "RAS Signaling" = "#d6604d",
  "Cytokine Signaling" = "#f4a582",
  "Tumor Suppressor" = "#e7298a"
)

ggplot(mutations, aes(x = Frequency, y = Gene, fill = Pathway)) +
  geom_col(width = 0.7, alpha = 0.9) +
  geom_text(aes(label = paste0(Frequency, "%")),
            hjust = -0.2, fontface = "bold", size = 3.5, color = "#333333") +
  scale_fill_manual(values = pathway_colors, name = "Functional Pathway") +
  scale_x_continuous(limits = c(0, 70), breaks = seq(0, 60, 10),
                     labels = function(x) paste0(x, "%")) +
  labs(
    title = "BPDCN Mutational Landscape by Functional Pathway",
    subtitle = "TET2 truncating mutations predict worse outcomes | ZRSR2 mutations show sex-biased distribution",
    x = "Approximate Mutation Frequency (%)",
    y = NULL
  ) +
  theme_minimal(base_size = 13) +
  theme(
    plot.title = element_text(face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(size = 11, color = "#555555", margin = margin(b = 15)),
    legend.position = "bottom",
    legend.title = element_text(face = "bold"),
    panel.grid.major.y = element_blank(),
    panel.grid.minor = element_blank(),
    axis.text.y = element_text(size = 11, face = "bold"),
    plot.margin = margin(15, 30, 15, 15)
  ) +
  guides(fill = guide_legend(nrow = 2))
Figure 7.1: Mutational landscape of BPDCN organized by functional pathway. TET2 is the most frequently mutated gene, with truncating TET2 mutations associated with worse outcomes (Beird et al., Blood Adv 2023).

7.2 Genomic Pathway Overview

Show code 
import matplotlib.pyplot as plt

categories = [
    'DNA Methylation\n(TET2, IDH2)',
    'Chromatin Remodeling\n(ASXL1, ARID1A, BCOR)',
    'Splicing\n(ZRSR2, SRSF2)',
    'RAS Signaling\n(NRAS, KRAS)',
    'Transcription Factors\n(ETV6, IKZF1, ZEB2)',
    'Tumor Suppressors\n(TP53, ATM)',
    'Cytokine Signaling\n(JAK2, STAT3)'
]
sizes = [35, 20, 15, 12, 10, 5, 3]
colors = ['#2166ac', '#4393c3', '#67a9cf', '#d6604d', '#984ea3', '#e7298a', '#f4a582']
explode = (0.05, 0, 0, 0, 0, 0, 0)

fig, ax = plt.subplots(figsize=(9, 6))
wedges, texts, autotexts = ax.pie(
    sizes, explode=explode, labels=categories, colors=colors,
    autopct='%1.0f%%', startangle=90, pctdistance=0.75,
    textprops={'fontsize': 9}
)

for autotext in autotexts:
    autotext.set_fontweight('bold')
    autotext.set_fontsize(10)
    autotext.set_color('white')

ax.set_title(
    'BPDCN Mutations by Functional Pathway',
    fontsize=16, fontweight='bold', color='#1a5276', pad=20
)

plt.tight_layout()
plt.show()
Figure 7.2: Distribution of BPDCN mutations across functional pathways. Epigenetic regulators (DNA methylation + chromatin remodeling) constitute the largest category.
ImportantPrognostic Genomic Markers

TET2 truncating mutations predict worse outcomes in BPDCN (Beird et al., Blood Adv 2023). ZRSR2 mutations show a sex-biased distribution (located on the X chromosome) and impair plasmacytoid dendritic cell activation and apoptosis (Togami et al., Cancer Discov 2022). Co-occurring myeloid neoplasms (CMML, MDS, AML) are clonally related in at least 10–20% of cases, often sharing a TET2-mutated origin.

8 Staging Workup

Given BPDCN’s propensity for multi-compartment involvement, a comprehensive staging evaluation is essential at diagnosis.

Evaluation Modality Purpose
Skin assessment Biopsy + MSWAT scoring Confirm diagnosis; quantify skin burden
Bone marrow Aspirate + biopsy + flow cytometry Assess marrow involvement and blast %
Peripheral blood CBC + differential + flow Detect circulating blasts
Imaging PET-CT Nodal, visceral, and extranodal disease
CSF Lumbar puncture + cytology + flow CNS assessment (occult involvement common)
Molecular NGS panel Identify mutations and co-occurring neoplasms
Cytogenetics Conventional karyotype + FISH Chromosomal abnormalities (60% have ≥1)
Finding Frequency Implication
Occult CNS at diagnosis ~15% All patients need LP at diagnosis
CNS at any point Up to 33% Serial IT prophylaxis recommended
CNS at relapse Common Highlights need for CNS-penetrant therapy
Standard CNS approach High-risk ALL protocols IT methotrexate or cytarabine
Compartment Method Response Criteria
Skin Biopsy + MSWAT Modified from mycosis fungoides/Sézary
Bone marrow IWG AML criteria Standard morphologic + flow assessment
Lymph nodes Lugano CT criteria Bidimensional measurements
CNS Serial LP Cytology + flow clearance
Overall Composite CR Disappearance in ALL initial disease sites

9 Co-occurring Myeloid Neoplasms

Show code 
library(ggplot2)

pchm_data <- data.frame(
  Condition = factor(c(
    "CMML",
    "MDS",
    "AML",
    "MPN",
    "CML"
  ), levels = c("CMML", "MDS", "AML", "MPN", "CML")),
  Frequency = c(8, 5, 4, 2, 1),
  Note = c(
    "Most common\nassociation",
    "Shared TET2\norigin",
    "Clonally\nrelated",
    "Less\ncommon",
    "Rare"
  )
)

ggplot(pchm_data, aes(x = Condition, y = Frequency)) +
  geom_col(fill = c("#2166ac", "#4393c3", "#67a9cf", "#d1e5f0", "#e0e0e0"),
           width = 0.6, alpha = 0.9) +
  geom_text(aes(label = paste0("~", Frequency, "%")),
            vjust = -0.5, fontface = "bold", size = 5, color = "#333333") +
  geom_text(aes(label = Note, y = Frequency / 2),
            size = 3, color = "white", fontface = "italic", lineheight = 0.9) +
  scale_y_continuous(limits = c(0, 12), breaks = seq(0, 10, 2),
                     labels = function(x) paste0(x, "%")) +
  labs(
    title = "Co-occurring Myeloid Neoplasms in BPDCN",
    subtitle = "Prior or concurrent hematologic malignancies (PCHM) occur in 15–20% of patients",
    x = NULL,
    y = "Approximate Frequency (%)"
  ) +
  theme_minimal(base_size = 14) +
  theme(
    plot.title = element_text(face = "bold", size = 16, color = "#1a5276"),
    plot.subtitle = element_text(size = 11, color = "#555555", margin = margin(b = 15)),
    panel.grid.major.x = element_blank(),
    panel.grid.minor = element_blank(),
    axis.text.x = element_text(size = 12, face = "bold"),
    plot.margin = margin(15, 20, 15, 15)
  )
Figure 9.1: Prior or concurrent hematologic malignancies (PCHM) in patients with BPDCN. Co-occurring neoplasms are clonally related and may affect treatment decisions and outcomes.
NoteClinical Impact of Co-occurring Neoplasms

Approximately 20% of patients with BPDCN have a prior or concomitant hematologic malignancy. These co-occurring diagnoses — most often CMML, MDS, or AML — are frequently confirmed to be clonally related, often sharing a TET2-mutated origin. The presence of PCHM may alter disease biology and affect clinical outcomes, including potentially lower overall survival, and must be accounted for in treatment planning.

10 Summary: Diagnostic Checklist

Show code 
library(ggplot2)

summary_data <- data.frame(
  Step = factor(c(
    "Clinical suspicion\n(skin + cytopenias)",
    "Tissue biopsy\n(skin or marrow)",
    "Core IHC panel\n(CD123, CD4, CD56)",
    "Extended pDC markers\n(TCF4, TCL1, CD303, CD304)",
    "Lineage exclusion\n(CD3−, CD19−, MPO−, CD14−)",
    "NGS mutational panel",
    "Cytogenetics\n(karyotype + FISH)",
    "PET-CT staging",
    "CSF evaluation\n(LP + flow)",
    "Expert hematopathology\nreview"
  ), levels = rev(c(
    "Clinical suspicion\n(skin + cytopenias)",
    "Tissue biopsy\n(skin or marrow)",
    "Core IHC panel\n(CD123, CD4, CD56)",
    "Extended pDC markers\n(TCF4, TCL1, CD303, CD304)",
    "Lineage exclusion\n(CD3−, CD19−, MPO−, CD14−)",
    "NGS mutational panel",
    "Cytogenetics\n(karyotype + FISH)",
    "PET-CT staging",
    "CSF evaluation\n(LP + flow)",
    "Expert hematopathology\nreview"
  ))),
  Category = c(
    "Clinical", "Pathology", "Immunophenotype", "Immunophenotype",
    "Immunophenotype", "Genomics", "Genomics", "Staging", "Staging", "Final"
  ),
  Priority = c(100, 100, 100, 95, 95, 85, 80, 90, 88, 100),
  bar_val = c(100, 100, 100, 95, 95, 85, 80, 90, 88, 100)
)

cat_colors <- c(
  "Clinical" = "#f4a582",
  "Pathology" = "#d6604d",
  "Immunophenotype" = "#2166ac",
  "Genomics" = "#4daf4a",
  "Staging" = "#4393c3",
  "Final" = "#1a5276"
)

ggplot(summary_data, aes(x = bar_val, y = Step, fill = Category)) +
  geom_col(width = 0.6, alpha = 0.9) +
  scale_fill_manual(values = cat_colors, name = "Diagnostic Domain") +
  scale_x_continuous(limits = c(0, 115)) +
  labs(
    title = "BPDCN Diagnostic Checklist — Essential Elements",
    subtitle = "Comprehensive workup integrating clinical, immunophenotypic, genomic, and staging evaluations",
    x = NULL, y = NULL
  ) +
  theme_void(base_size = 14) +
  theme(
    plot.title = element_text(face = "bold", size = 18, color = "#1a5276",
                              margin = margin(b = 5)),
    plot.subtitle = element_text(size = 12, color = "#555555",
                                 margin = margin(b = 20)),
    axis.text.y = element_text(size = 11, hjust = 1, margin = margin(r = 10)),
    legend.position = "bottom",
    legend.title = element_text(face = "bold", size = 12),
    legend.text = element_text(size = 10),
    plot.margin = margin(20, 50, 20, 20)
  ) +
  guides(fill = guide_legend(nrow = 2))
Figure 10.1: Comprehensive diagnostic checklist for BPDCN. Each element contributes to accurate diagnosis and appropriate treatment selection.
NoteKey Diagnostic Takeaways
  1. BPDCN is a distinct WHO entity: No longer classified as an AML subtype — requires specific immunophenotyping and targeted therapeutic strategies
  2. Skin is the sentinel organ: Dark purple nodules, plaques, or tumors in 65–94% of patients; frequently misdiagnosed as melanoma, leprosy, or dermatomyositis
  3. “123-4-56” mnemonic: CD123 + CD4 + CD56 form the diagnostic triad; always confirm with additional pDC markers (TCF4, TCL1, CD303, CD304)
  4. Lineage exclusion is mandatory: CD3−, CD19−, MPO−, CD14− must be documented to rule out mimics
  5. Universal CD123 expression enables targeted therapy: Tagraxofusp (approved) and pivekimab sunirine (investigational) both target CD123
  6. Comprehensive staging is essential: Multi-compartment involvement (skin, marrow, nodes, blood, CNS) must be fully evaluated; occult CNS disease is present in ~15%
  7. Genomic profiling informs prognosis: TET2 truncating mutations predict worse outcomes; co-occurring myeloid neoplasms (CMML/MDS/AML) found in 15–20%
  8. Expert hematopathology review is imperative: Referral to a cancer center with clinical trial availability and BPDCN experience is strongly recommended
TipRendering This Document

To render this Quarto document, ensure the following are installed:

R packages:

install.packages(c("ggplot2", "dplyr", "tidyr"))

Python packages:

pip install matplotlib numpy

Render command:

quarto render bpdcn_diagnosis.qmd

Sources: Kharfan-Dabaja MA, Lane AA, Pemmaraju N. How I treat blastic plasmacytoid dendritic cell neoplasm. Blood. 2025;145(6):567-574. Luskin MR, Lane AA. Tagraxofusp for BPDCN. Haematologica. 2024;109(1):44-52. Pemmaraju N, et al. Pivekimab sunirine in BPDCN. J Clin Oncol. 2026. Pemmaraju N, Lane AA, Sweet KL, et al. Tagraxofusp in BPDCN. N Engl J Med. 2019;380(17):1628-37. Garnache-Ottou F, et al. Blood Adv. 2019;3(24):4238-4251. Togami K, et al. Cancer Discov. 2022;12(2):522-541. Sukswai N, et al. Am J Surg Pathol. 2019;43(10):1429-1437.