Development and Validation of the Blastic Plasmacytoid Dendritic Cell Neoplasm Prognostic Score

Background

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is an uncommon and highly aggressive hematologic malignancy primarily characterized by its frequent involvement of the skin and potential to rapidly progress to systemic disease. Initially misclassified under various categories, such as blastic NK-cell lymphoma, and later as a subset of acute myeloid leukemia (AML), BPDCN has now been distinctly recognized as a unique myeloid neoplasm. Despite recent advancements in understanding its pathogenesis and clinical behavior, the prognosis of BPDCN remains poor and highly variable among patients. Accurate prognostic tools are crucial for improving patient management and outcomes. This study focuses on developing and validating a clinical prognostic score for BPDCN, aiming to stratify patients better based on their risk and guide therapeutic decision-making.

Methods

Using a constructed BPDCN database of 273 cases, we analyzed demographics and clinicopathologic factors. Of these, 218 cases had complete survival and outcomes data. Cox proportional-hazards models and Log-rank tests were used to identify factors impacting overall survival (OS). Factors with significant impact on OS were scored based on their hazard ratios. A validation cohort of 106 BPDCN-confirmed cases was used to test the score. Subsequently, both cohorts were combined and split into adult and pediatric cohorts for further testing and validation across age groups.

Results

The median OS for the initial, validation, adult, and pediatric cohorts were 16, 15, 12, and 60 months, respectively. The following dichotomous variables significantly impacted OS: Age > 60 (10 vs. 30 months), lymphadenopathy (14 vs. 42 months), bone marrow involvement (14 vs. NR months), splenomegaly (10 vs. 16 months), disseminated skin disease (13 vs. NR months). Variables were assigned points: 1 point for each variable, except age > 60, splenomegaly, and disseminated skin disease, which were given 2 points due to their high hazard ratios. The risk model categorized patients into low (0-2 points), intermediate (3-4 points), and high-risk (5-8 points) groups. Median OS for these groups were NR, 24, and 10 months, respectively (p < 0.0001). The BPS similarly identified 3 risk groups in the validation cohort (38 vs. 13 vs. 9 months, p = 0.0001), adult cohort (22 vs. 13 vs. 9 months, p < 0.0001), and pediatric cohort (NR vs. 36 vs. 9 months, p = 0.002).

Conclusion

The BPDCN prognostic score (BPS) is a promising tool that effectively stratifies patients into 3 distinct risk categories with significant differences in overall survival. This score was validated in multiple cohorts, including adult and pediatric groups, but requires prospective validation for further confirmation.