AUTHORS
Hana Qasim, Fellow, Saheli Parekh, MD2, Neha Gutta, MD3, Mohammad Hussaini, MD4, Linsheng Zhang, MD5, Todd Knepper, PhD6, Ghada Araji, MD7, Shehab F. Mohamed, MD8, Daniela Carolina Carolina Bravo Solarte, MD9, Andrew T Kuykendall, MD10, Rami Komrokji, MD11, Eric Padron, MD12, Anthony Michael Hunter, MD13
AFFILIATIONS
1 H. Lee Moffitt Cancer Center & Research Institute, Tampa, United States
2 University of Emory, Atlanta, Georgia, United States
3 University of South Florida/Moffitt Cancer Center, Tampa, United States
4 H. Lee Moffitt Cancer Center & Research Institute, Tampa, United States
5 University of Emory, Atlanta, United States
6 H. Lee Moffitt Cancer Center & Research Institute, Tampa, United States
7 H. Lee Moffitt Cancer Center & Research Institute, Tampa, United States
8 Hamad Medical Corporation, Doha, Qatar
9 H. Lee Moffitt Cancer Center & Research Institute, Tampa, United States
10 H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, United States
11 H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, United States
12 H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, United States
13 Emory University, Decatur, Georgia, United States
ABSTRACT
BACKGROUND:
Activating CSF3R mutations are established drivers of chronic neutrophilic leukemia (CNL), but pathogenic somatic CSF3R alterations are increasingly recognized across a broader spectrum of myeloid neoplasms. Their clinicogenomic characteristics and outcomes outside CNL remain incompletely characterized.
METHODS:
Patients with pathogenic or likely pathogenic somatic CSF3R mutations identified between April 2014 and August 1, 2025 at Moffitt Cancer Center and Emory University were included. Cases were identified using CLIA-certified next-generation sequencing assays. Germline and benign or likely benign variants were excluded. Clinical, laboratory, cytogenetic, molecular, treatment, and outcome data were abstracted from the medical record. Complete blood count parameters and bone marrow blast percentage were recorded at first CSF3R-positive NGS. Overall survival (OS) was calculated from first CSF3R-positive NGS to death; patients without documented death were censored at last known follow-up. Kaplan-Meier methods were used to estimate OS. Ruxolitinib responses were assessed only among treated patients with evaluable follow-up. Hematologic response was defined as cytoreduction and/or reduction in transfusion requirements documented by the treating physician. Splenic response was defined as patient-reported improvement and/or documented reduction in spleen size on physical examination or imaging. Molecular response was defined as clearance of the CSF3R mutation or a reduction in CSF3R variant allele frequency (VAF) on serial next-generation sequencing.
RESULTS:
We identified 133 patients. Median age at first CSF3R-positive NGS was 70.1 years (IQR, 61.2-77.4), and 90/133 (67.7%) were male. Diagnoses included AML (50, 37.6%), CNL (22, 16.5%), MDS (18, 13.5%), CMML (12, 9.0%), MDS/MPN (10, 7.5%), MPN (10, 7.5%), MDS/MPN with neutrophilia (4, 3.0%), and other myeloid neoplasms (7, 5.3%). Among 111 patients with documented date of first NGS, CSF3R was detected at diagnosis in 62 patients (55.9%) and post-diagnosis in 49 (44.1%). Among 47 patients with complete date information, the median interval from initial diagnosis to first CSF3R-positive NGS was 14.5 months (IQR, 7.2–32.1). At first CSF3R-positive NGS, median WBC was 8.7 ×10³/µL (range, 0.2–130.4), hemoglobin 9.1 g/dL (range, 5.6–15.0), platelet count 78.5 ×10³/µL (range, 5–807), bone marrow blasts 5.0% (range, 0–85), and median CSF3R variant allele frequency was 27.0% (IQR, 8.6–43.1; range, 0–82.2). Variants were membrane-proximal in 56 (42.1%), truncating in 53 (39.8%), compound in 9 (6.8%), transmembrane in 8 (6.0%), and other clinically significant variants in 7 (5.3%). Cytogenetics were evaluable in 119 patients; 58 (48.7%) had an abnormal karyotype. The most frequent co-mutations were spliceosome-pathway genes (65/129, 50.4%) and ASXL1 (61/131, 46.6%); among spliceosome mutations, SRSF2 was the most common (39/128, 30.5%). Twenty-seven patients received ruxolitinib, including 14 with membrane-proximal CSF3R mutations. Among evaluable treated patients, hematologic response occurred in 14/19 (73.7%), splenic and/or symptom response in 8/10 (80.0%), molecular response in 4/8 (50.0%), and overall clinical benefit in 16/19 (84.2%). Response rates were similar across mutation classes; however, comparisons were limited by the small number of evaluable non–membrane-proximal cases. Among 91 patients with a non-AML myeloid neoplasm at baseline, 29 (31.9%) progressed to AML. The median time to AML transformation was 14.8 months (IQR, 7.3-29.3 months). Among 133 patients with evaluable OS, 79 deaths occurred; median OS was 16.0 months, with 12- and 24-month OS estimates of 61.8% and 33.8%, respectively. Using initial diagnosis as the baseline disease classification, patients with CNL demonstrated a numerically longer median overall survival than patients with non-CNL/non-AML diagnoses (17.6 vs 12.4 months), although this difference was not statistically significant (log-rank P=0.219).
CONCLUSIONS:
Clinically significant somatic CSF3R mutations occur predominantly across a broad spectrum of myeloid neoplasms and are associated with substantial clinicogenomic heterogeneity and poor overall survival. Ruxolitinib demonstrated frequent clinical benefits among evaluable treated patients, supporting further prospective evaluation of biomarker-directed therapeutic strategies.
DISCLOSURES
Hana Qasim, Fellow: No relevant conflicts of interest to declare.