AUTHORS
Bruce Liu, Jack H. Peplinski, MSc2, Rhonda E. Ries, MA3, Grace Hawkins4, Katherine Tarlock I, MD5, Soheil Meshinchi, MD, PhD6, Danielle C. Kirkey, MD7
AFFILIATIONS
1 Fred Hutchinson Cancer Center, Seattle, Washington, United States
2 Fred Hutchinson Cancer Center, Seattle, Washington, United States
3 Fred Hutchinson Cancer Center, Seattle, Washington, United States
4 Fred Hutchinson Cancer Center, Seattle, Washington, United States
5 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
6 Fred Hutchinson Cancer Center, Seattle, Washington, United States
7 Fred Hutchinson Cancer Center, Seattle, Washington, United States, Seattle Children's Hospital, Seattle, Washington, United States
ABSTRACT
Extramedullary disease (EMD) in acute myeloid leukemia (AML) has traditionally been viewed as an end-organ manifestation of leukemia rather than a biologically active source of disease progression. In a complementary study, we demonstrate that localized extramedullary leukemia disseminates to the peripheral blood, bone marrow, and distant organs, supporting a new biologic model in which EMD functions as a protected reservoir of viable leukemia capable of reseeding systemic disease. We hypothesize that, if this reservoir model is clinically relevant, patients presenting with non-CNS EMD would exhibit an increased risk of subsequent systemic relapse and inferior clinical outcomes. To test this hypothesis, we evaluated patients with newly diagnosed AML for the presence of non-CNS EMD and examined its association with clinical and genomic characteristics, treatment response, relapse, and survival.
We analyzed 2,228 children, adolescents, and young adults with newly diagnosed AML enrolled on prior COG trials. Five-year event-free survival, disease-free survival, overall survival, and relapse risk (EFS, DFS, OS and RR) were compared between patients with and without EMD. Recurrent fusions and somatic mutations were evaluated for association with EMD, the prognostic impact of EMD was assessed within major molecular subgroups, and the effect of allogeneic hematopoietic stem cell transplantation (HSCT) on outcomes among patients with EMD was examined.
Among the cohort, 297 (13.3%) presented with documented non-CNS EMD at diagnosis, whereas 1,931 (86.7%) had no evidence of EMD. The EFS for those with and without EMD was 39% vs. 47% (P<0.001), OS 58% vs. 64% (P=0.02), and RR of 52% vs. 41% (P<0.0001) indicating inferior durable leukemia control among patients with EMD.
To determine if this adverse outcome reflected enrichment of biologically distinct AML subtypes, we evaluated the genomic landscape of EMD. Non-CNS EMD was significantly enriched among patients with KMT2A-MLLT3 (20.1%, Padj=0.001), KMT2A-MLLT10 (31.9%, Padj<0.0001), KMT2A-ELL (26.1%, Padj=0.005), rare KMT2A-r (29.5%, Padj=0.001), RUNX1-RUNX1T1 (17.4%, Padj=0.001) and ETV6-MNX1 (33.3%, Padj=0.04) when compared with no fusion AML (8.8%) by Fisher’s Exact Test with Benjamini-Hochberg adjustment for multiple testing correction. No mutations were significantly enriched for EMD, however CEBPA mutations were notably depleted in non-CNS EMD (2.7% vs 13.9%, Padj = 0.014).
To assess whether the adverse impact of EMD was confined to specific subtypes, outcomes were analyzed within individual molecular groups. Among patients with KMT2A-r AML, EFS of 28% vs 38% for those with and without EMD (P=0.006) and RR of 69% vs 48% (P=0.038) was observed. Within the KMT2A-r patients, those with high-risk fusion partners had identical EFS (21%) and RR (76%) regardless of EMD presence (P=NS). However, those with standard risk fusion partners with and without EMD had an EFS of 33% vs. 48% (P=0.005) and corresponding RR of 63% vs. 49% (P=0.05). This difference was most pronounced in patients with KMT2A-MLLT3 (EFS 23% vs. 47%, P<0.001). Presence of EMD in those with RUNX1-RUNX1T1 was not associated with adverse outcome. However, those with CBFB-MYH11 with and without EMD had an EFS of 46% vs. 62% (P=NS) with a corresponding RR of 61% vs. 33% (P=0.020). Finally, we evaluated whether HSCT mitigated the adverse impact of EMD and found DFS for patients consolidated with HSCT vs. chemotherapy was 47% and 45% (P=NS) with corresponding RR of 41% vs 54% (P=NS).
These findings provide clinical validation of inferior outcomes in AML with EMD supporting a new biologic model in which non-CNS extramedullary leukemia can function as a protected site that can promote systemic relapse. The particularly poor outcomes in KMT2A-r AML, together with no measurable benefit from HSCT, suggest that conventional therapeutic intensification alone is insufficient to overcome this biology. Because current clinical practice identifies EMD only if clinically apparent, the true burden of EMD is likely underestimated. These findings support prospective evaluation of sensitive whole-body imaging strategies to detect occult EMD, particularly in molecular subgroups enriched for EMD. Collectively, our data suggest that durable leukemia remission may require identification and eradication of both medullary and extramedullary reservoirs, establishing EMD as a critical therapeutic target for future AML treatment strategies.
DISCLOSURES
Bruce Liu: No relevant conflicts of interest to declare.