Outcomes of Blast-Phase MPN and JAK2, MPL, and CALR Mutated de Novo AML Compared to AML with Wild-Type JAK2, MPL, and CALR Genes: A Propensity Score-Adjusted Cohort Study

Background:

The BCR-ABL-negative myeloproliferative neoplasms (MPNs) include polycythemia vera, essential thrombocythemia, and primary myelofibrosis. These MPNs are clonal disorders of hematopoietic stem cells that can transform into secondary acute myeloid leukemia (AML), known as the blast phase of an MPN. Transformation to blast phase occurs in about 10% of patients with MPN. Previous studies have shown that patients with MPN-BP have very poor prognosis, and there is currently no accepted standard of care for these patients. The VIALE-A trial, which showed that combining venetoclax (VEN) with azacitidine improved outcomes in AML patients, excluded patients with MPN-BP. The purpose of this single-center retrospective study is to compare the outcomes of MPN-BP patients to patients with de novo AML with JAK2, MPL, and CALR mutations and AML with wild-type JAK2, MPL, and CALR genes.

Methods:

We conducted a single-center retrospective cohort study to compare composite complete response (CCR), overall survival (OS), and event-free survival (EFS) of patients with MPN-BP, de novo AML with underlying JAK2, MPL, or CALR mutations (AML-D-MPN), and AML without underlying JAK2, MPL, or CALR mutations (AML-WT). Patients with AML who were tested for JAK2, MPL, and CALR mutations at baseline were included in the study. Data collected included age, gender, ethnicity, comorbidities, AML subtypes, cytogenetics, BCR::ABL1 and FLT3 status, mutational profiles (via next-generation sequencing), ECOG performance status, treatment history, and follow-up. CCR was defined as complete response (CR) or CR with incomplete count recovery. Event-free survival (EFS) was defined from the diagnosis date to refractoriness, progression, or death. Overall survival (OS) was from diagnosis date to death. Survival probabilities were estimated using the Kaplan-Meier method, with differences assessed by the log-rank test. Multivariable regression and propensity score (PS) analysis were used to adjust for confounders, including comorbidities.

Results:

From May 2017 to September 2023, 730 AML patients were treated at the Cleveland Clinic Foundation, Ohio. Of the 528 AML patients who had NGS at baseline, 74 had MPN-BP (14%), 22 had AML-D-MPN (4%), and 432 had AML-WT (82%). The median age for the groups above were 62 (IQR: 55-73), 71 (IQR: 66-78), and 66 (IQR: 57-73) years. In patients with MPN-BP, 70% had JAK2 mutations, 28% had CALR mutations, and 2% had MPL mutations. In patients with AML-D-MPN, 100% had JAK2 mutations. The median time for MPN to BP transformation was 49 months (IQR: 21-123). With a median follow-up of 34.9 months (mo) (95CI: 1.8-90.2), the median OS for MPN-BP was 7.1 mo (95CI: 4.1-11), 13 mo (95CI: 6.1-NC) for AML-D-MPN, and 12 mo (95CI: 10-14) for AML-WT (log-rank P<0.01). Similarly, the MPN-BP had worse EFS when compared to AML-D-MPN and AML-WT (P=0.02).

In a combined analysis for MPN-BP and AML-D-MPN, we compared CCR, OS and EFS between first-line treatment arms. Twenty-seven patients received intensive regimens, 17 received VEN + hypomethylating agents (HMA), 11 received HMA, and 22 received other regimens. There was no statistically significant difference in response rate on PS-adjusted logistic regression. The PS-adjusted median OS was 17 mo (95CI: 11-NR) for intensive regimens, 10 mo (95CI: 5.2-NR) for VEN+HMA, 13 mo (95CI: 13-13) for other regimens and 1.2 mo (95CI: 0.85-NR) for HMA regimens. Similarly, the PS-adjusted median EFS was 13 mo (95CI: 407-NR) for intensive regimen, 5.2 mo (95CI: 4.2-NR) for VEN+HMA regimens, 5.1 mo (95CI: 5.1-5.1) for other regimens and 1.2 mo (95CI: 0.85-NR) for HMA regimens.

Conclusions:

While all AML patients in our study had poor survival outcomes, patients with MPN-BP had worse survival when compared to patients with de novo AML with a similar mutational profile and patients with AML negative for these mutations. Intensive regimens in MPN-BP and AML-D-MPN were associated with the largest impact on OS and should be utilized when feasible.