Aggressive natural killer (NK) cell leukemia (ANKL) is a rare disease with a grave prognosis. Patients commonly present acutely with fever, constitutional symptoms, hepatosplenomegaly, and often disseminated intravascular coagulation or hemophagocytic syndrome. This acute clinical presentation and the variable pathologic and immunophenotypic features of ANKL overlap with other diagnostic entities, making it challenging to establish a timely and accurate diagnosis of ANKL. Since its original recognition in 1986, substantial progress in understanding this disease using traditional pathologic approaches has improved diagnostic accuracy. This progress, in turn, has facilitated the performance of recent high-throughput studies that have yielded insights into pathogenesis. Molecular abnormalities that occur in ANKL can be divided into three major groups: JAK/STAT pathway activation, epigenetic dysregulation, and impairment of TP53 and DNA repair. These high-throughput data also have provided potential therapeutic targets that promise to improve therapy and outcomes for patients with ANKL. In this review, we provide a historical context of the conception and evolution of ANKL as a disease entity, we highlight advances in diagnostic criteria to recognize this disease, and we review recent understanding of pathogenesis as well as biomarker discoveries that are providing groundwork for innovative therapies.

Acute myeloid leukemia (AML) with ≥2% plasmacytoid dendritic cells (pDC) has been recently described as AML with pDC differentiation (pDC-AML) characterized by pDC expansion with frequent RUNX1 mutations. In this study, we investigated a cohort of 53 pDC-AML cases representing about 3% of all AML cases. We characterized their immunophenotype and genetic profiles and compared these findings with blastic plasmacytoid dendritic cell neoplasm (BPDCN). pDC-differentiation/expansion was preferentially observed in AML with an immature myeloid or myelomonocytic immunophenotype, where myeloblasts were frequently positive for CD34 (98%), CD117 (94%), HLA-DR (100%) and TdT (79%), with increased CD123 (89%) expression. The median number of pDCs in pDC-AML was 6.6% (range, 2% to 26.3%) and their immunophenotype reminiscent of pDCs in early or intermediate stages of differentiation. The immunophenotype of pDCs in pDC-AML was different from BPDCN (n = 39), with major disparities in CD34 (96% vs. 0%), CD56 (8% vs. 97%) and TCL1 (12% vs. 98%) and significant differences in frequency of CD4, CD13, CD22, CD25, CD36, CD38, CD117 and CD303 expression. At the molecular level, the genetic landscapes of pDC-AML and BPDCN also differ, with RUNX1 mutations detected in 64% of pDC-AML versus 2% of BPDCN. Disparities in TET2 (21% vs. 56%), FLT3 (23% vs. 0%), DNMT3A (32% vs. 10%) and ZRSR2 (2% vs. 16%) (all p < 0.05) were also detected. The distinct immunophenotypic and mutation profiles of pDC-AML and BPDCN indicate that the neoplastic pDCs in pDC-AML and BPDCN derived from different subsets of pDC precursors.