AUTHORS
Jeffrey Zwicker, MD, Michael R. Grunwald, MD2, Stephen Oh, MD, PhD3, John O. Mascarenhas, MD4, Naveen Pemmaraju, MD5, Taehyung Kim, PhD6, Cynthia D Timmers, PhD7, Erin Crowgey, PhD8, Evan Braunstein, MD, PhD9, Abdulraheem Yacoub10
AFFILIATIONS
1 Memorial Sloan Kettering Cancer Center, New York, New York, United States, Weill Cornell Medical College, New York, New York, United States
2 Atrium Health Levine Cancer Institute and Wake Forest University School of Medicine, Charlotte, North Carolina, United States
3 Washington University School of Medicine, St. Louis, Missouri, United States
4 Icahn School of Medicine at Mount Sinai, New York, New York, United States
5 The University of Texas MD Anderson Cancer Center, Houston, Texas, United States
6 Incyte Corporation, Wilmington, Delaware, United States
7 Incyte Corporation, Wilmington, Delaware, United States
8 Incyte Corporation, Wilmington, Delaware, United States
9 Incyte Corporation, Wilmington, Delaware, United States
10 The University of Kansas Cancer Center, Westwood, Kansas, United States
ABSTRACT
Introduction: Disease-initiating mutations in JAK2 or CALR occur in >90% of patients (pts) with myeloproliferative neoplasms (MPNs). JAK2V617F and frameshift mutations in CALR exon 9 (mutCALR) drive constitutive activation of the JAK/STAT pathway, leading to oncogenesis. Disease phenotype and clinical outcomes differ among pts with JAK2V617F or mutCALR. Using the REVEAL (polycythemia vera [PV]) and MOST (essential thrombocythemia [ET]/low-risk myelofibrosis [MF]) observational studies, we investigated the molecular dynamics of JAK2V617F and mutCALR variant allele frequency (VAF) to assess correlations with clinical outcomes, including disease progression.
Methods: Optional biospecimens were collected during REVEAL and MOST. All pts with ≥1 available sample, excluding pts with dual JAK2V617F and mutCALR, were included in this analysis. VAF was determined by whole exome sequencing (WES) or droplet digital PCR (ddPCR) using pipelines for variant calling and copy-neutral loss of heterozygosity (CN-LOH) (Oh et al, EHA 2025).
Results: Of 2052 evaluable pts with JAK2V617F, 1456 (71.0%) were diagnosed with PV, 495 with ET (24.1%), and 101 (4.9%) with MF. Mean (SD) duration from diagnosis to enrollment was 5.7 (5.9) years in PV, 6.2 (6.7) years in ET, and 3.1 (4.8) years in MF. Mean (SD) JAK2V617F VAF at enrollment was 47.4% (29.6) in PV, 20.3% (18.4) in ET, and 44.3% (23.0) in MF. After adjusting for disease duration, JAK2V617F VAF was significantly lower in ET compared with PV or MF (both P<0.001). Disease duration was independently associated with an estimated VAF increase of 0.84% per year across all MPN subtypes (P<0.001). JAK2V617F VAF highly correlated with absolute neutrophil count (ANC) in PV (r>0.5, P<0.001), but less so in ET and MF (r<0.5, P<0.001). Weak correlations between JAK2V617F VAF with platelets, hematocrit, and lactate dehydrogenase (LDH) were observed in PV and ET (r<0.5, P<0.001); JAK2V617F VAF correlated with LDH but not platelets or hematocrit in MF.
The contribution of LOH to JAK2V617F VAF was assessed in a subset of JAK2V617F+ pts at enrollment by WES (PV, n=434; ET, n=495; MF, n=101). LOH was identified in 68.2% of pts with PV, 12.5% with ET, and 37.6% with MF. Most LOH events were CN-LOH. In pts with LOH at JAK2 loci, median (range) JAK2V617F VAF was highest in PV (0.76 [0.17, 0.99]), followed by MF (0.67 [0.29, 0.94]) and ET (0.57 [0.20, 0.92]). Allelic imbalance, a quantitative measure of LOH (Grinfeld et al, New Engl J Med 2018), correlated with JAK2V617F VAF across subtypes.
Among 81 pts with PV that progressed to MF with longitudinal ddPCR data, the greatest rate of JAK2V617F increase was in pts who enrolled earlier in their disease course (<5 years, median 5.9%, n=33; ≥5 years, median 1.5%, n=48). Notably, among 77 pts with longitudinal WES data, 79% had CN-LOH at enrollment, and half of those without CN-LOH at enrollment (8/16) gained LOH during the study.
Of 227 evaluable pts with mutCALR, 187 (82.4%) were diagnosed with ET and 40 (17.6%) with MF. Pts with ET and MF had similar mutCALR VAF at enrollment (mean [SD]: 24.9% [11.9] vs 25.0% [7.1]). Longer disease duration was independently associated with higher mutCALR VAF (increase of 0.35% per year; P≤0.001). CN-LOH was rare and limited to pts with Type 2 mutCALR (2%). At enrollment, there was no correlation between mutCALR VAF and blood counts, including platelets, ANC, and hematocrit.
In pts with Type 1 mutCALR (n=99) with enrollment and post-enrollment ddPCR data, pts with disease duration <5 years had a larger median mutCALR VAF increase from enrollment than pts with disease duration ≥5 years (6.4% vs 0%), similar to the acceleration pattern observed in JAK2V617F; no significant association between disease duration and change in mutCALR VAF was observed among pts with Type 2 mutCALR (n=60) (1.1% vs 6.6%, P=0.487). Changes in mutCALR VAF did not correlate with changes in blood counts (r<0.5).
Conclusions: Longitudinal analyses of driver mutation dynamics in pts with MPNs show that JAK2V617F and mutCALR VAF both increased over time, with the greatest changes occurring within the first 5 years from diagnosis. In pts with LOH at JAK2 loci, JAK2V617F VAF was higher in PV compared with MF or ET, while mutCALR VAF was similar in ET and MF, suggesting that JAK2- and CALR-driven MPN follow distinct trajectories of clonal evolution. JAK2V617F VAF was influenced by CN-LOH in PV, and acquisition of LOH is a potential driver of progression.
DISCLOSURES
Jeffrey Zwicker, MD: Pfizer/Bristol Myers Squibb - Honoraria; CSL Behring - Honoraria; Sanofi - Honoraria; Incyte Corporation - Research Funding; Quercegen - Research Funding; Parexel - Consultancy (Includes expert testimony)