APL-like Subset within NPM1-Mutated Acute Myeloid Leukemia: A Distinct Phenotypic Signature Correlating with Early-Onset Vascular Complications

Background: NPM1 mutations represent the most common genetic alteration in adult acute myeloid leukemia (AML), often in co-occurrence with FLT3-ITD and mutations in genes influencing DNA methylation (DNMT3A, IDH1, IDH2 and TET2). Whether these genetic alterations correlate with distinct immunophenotypic and clinical features is still a matter of debate. There is growing evidence of distinct immunophenotypic subsets driven by co-occurring genetic mutations in NPM1-mutated AML, potentially accounting for the heterogeneity of clinical presentations and outcome. Mason et alhave described a subtype of NPM1-mutated AML, displaying an immunophenotypic profile resembling that of acute promyelocytic leukemia, specifically, the negativity for both CD34 and HLA-DR, and as such defined APL-like.

Aims: We studied the characteristics of a series of patients with APL-like NPM1-mutated AML, focusing on the incidence of vascular events at disease onset, and investigated the impact of some markers (blood counts, coagulation parameters and LDH) reported to correlate with coagulopathy in APL.

Methods: The study cohort included patients diagnosed with NPM1-mutated AML at our Centre according to conventional morphological, immunophenotypic, cytogenetic and molecular criteria. Vascular events were defined according to the revised World Health Organization (WHO) bleeding scale and to the CTCAE grading of thromboembolic events.

Results: From April 2007 to May 2023, 139 patients with a diagnosis of NPM1-mutated AML were enrolled, of whom 31 (22.3%) featured by APL-like phenotype. Their characteristics are detailed in Table 1. APL-like patients were older (64 y) compared to non-APL-like (57 y, P=0.002) NPM1-mutated patients; no further difference emerged for baseline blood count parameters. Vascular complications (n=22 bleeding events and n=2 thrombotic events) were significantly more frequent in the APL-like (n=10, 34.5%) than non-APL-like (n=16, 13.6%, P=0.015) group. There was a trend for more severe (G3-G4) vascular events in APL-like (3/31, 9.7%) vs non-APL-like (3/108, 2.8%, P=0.12). Also, abnormal coagulopathy-related parameters, including INR ≥1.5 and/or fibrinogen below normal level) were more frequent in APL-like patients (27.6% versus 15.7%,). D-dimer levels resulted significantly higher in APL-like patients (median 5998 ng/ml versus 2287 ng/ml, P=.005). The D-dimer/fibrinogen ratio (DD/FBG) showed significantly higher level in APL-like (median 16.84) vs non-APL-like (4.4, P=.017) patients. Of note, in multivariate analysis, APL-like phenotype maintained a value (OR=2.67, P=0.007) on DD levels from WBC count (OR=2.37, P=0.017), thus suggesting an effect independent from leukocytosis. Also, APL-like subset maintained an age-independent impact (OR=2.77, P=0.041) on the rate of vascular events. Among the 85 patients with full molecular information, there was a significant enrichment IDH1 (36.8%, P=0.002) and IDH2 (47.6%, P=.001) mutations in APL-like vs non-APL-like (6.1% and 8.6%, respectively) patients. No difference in the incidence of TET2 mutations was observed (18.8% and 25% in APL-like and non-APL-like, respectively, P=0.744). As regards outcome, complete remission rate was 85.7% and 78.9% in APL- and non-APL-like (P=0.43). We did not observe any significant difference between APL- and non-APL-like patients in disease-free (13.2 vs 22.0 months; P=0.83) or overall (15.9 vs 15.1 months, respectively; P=0.38) survival.

Conclusions: Our findings suggest APL-like signature to be a potential predictor of susceptibility to vascular events within NPM1-mutated AML. Our results deserve validation in a larger patient set and might indicate the utility of an intensive monitoring and supportive care to prevent early vascular, especially hemorrhagic, events in this patient category.