Evaluating real-world diagnostic and treatment patterns, risk of thromboembolic events, and mortality for polycythemia vera across an integrated academic-community cancer network

Abstract Background: Polycythemia vera (PV) is a chronic myeloproliferative neoplasm often managed in community settings where access to advanced diagnostics and novel therapies can vary. Whether network affiliation of community cancer centers with an academic institution fosters uniform management is unclear. We retrospectively examined diagnoses, treatment patterns, and outcomes of PV patients within an integrated academic-community cancer network. Methods: We identified PV patients treated in our Comprehensive Cancer Center (CCC) affiliated network from 2015-2022 through the institutional cancer registry using International Classification of Diseases for Oncology codes. Our network includes a CCC with a subspecialized leukemia team and 8 community-based affiliates. Baseline demographics, disease characteristics, diagnostic testing, and first and second-line cytoreductive therapy choices were compared between patients treated at the CCC vs network affiliates (NAC). Outcomes of interest were rates of hematocrit (HCT) control (45% for ≥3 months), blastic transformation or fibrotic progression, and thromboembolic events (TEs) post-PV diagnosis (venous, arterial, or intracardiac). Overall survival (OS) was assessed by Kaplan-Meier; Cox regression was used to evaluate whether treatment site was associated with outcomes. Results: Of the 152 patients included, 105 (69.1%) received PV care at a NAC and 47 (30.9%) at the CCC. Most patients (91.4%) were White and had a JAK2 V617F (95.4%) mutation. Median age at diagnosis was older in NAC (69 years, IQR: 57, 74) vs CCC (63 years, IQR: 51, 73); a higher proportion of NAC patients were female (61.9% vs 44.7%), not married (53.3% vs 29.8%), and resided 10 miles from the CCC (98.1% vs 68.1%, p0.05 for all). No significant differences were observed in prevalence of neighborhood disadvantage by area deprivation index or health insurance type, nor were there differences in comorbid autoimmune disease, second malignancy, hereditary or acquired thrombophilia, or tobacco dependence. NAC patients were more likely to present with high-risk (HR) PV (81.9% vs 61.7%, p=0.02) though median white blood cell count, hemoglobin, and HCT were not significantly different. Median time to PV diagnosis from first abnormal lab was 296.5 days (IQR: 26.3, 853.8); NAC patients more often experienced 3-year delay in diagnosis (17.1% vs 4.3% in CCC, p=0.037) and pre-diagnosis TEs (33.3% vs 8.5%, p=0.02). Extended next-generation sequencing was completed in 31.5% of NAC vs 61.8% of CCC patients (p0.0006). In the first-line, 56.2% of the NAC group (6 of 20 low-risk LR, 53 of 85 HR) and 57.4% of CCC group (5 of 18 LR, 22 of 29 HR) were on cytoreductive therapy, most commonly hydroxyurea (95.3%) followed by ruxolitinib (2.3%) and interferons (2.3%). At the time of analysis (07/27/2025), NAC patients with HR PV were less likely to be on cytoreductive therapy beyond aspirin ± phlebotomy compared to those at CCC (72.9% vs 100%, p0.0001), and more often remained on first-line (65.7% vs 25.5%, p0.0001). In second-line, NAC patients were less likely to be on ruxolitinib (7.5% vs 21.3%) or interferons (8.8% vs 40.4%) compared to CCC patients (p0.05 for all). At a median follow-up of 71 months, median OS was 136.4 months (95% CI: 128.0-not reached) for the entire cohort without significant difference between NAC vs CCC. HCT control was achieved by 87.6% of NAC vs 93.6% of CCC patients. Rates of blastic transformation or fibrotic progression were comparable (11.4% in NAC vs 6.4% in CCC, p=0.4). There were 34 TEs among 30 PV patients, and TEs were significantly higher in the NAC vs CCC group (27.6% vs 10.6%, p=0.03). After adjusting for age and sex, NAC patients were at increased risk of post-PV TEs (OR=3.67, 95% CI: 1.15-11.7) relative to CCC patients. Conclusions: Within our integrated academic-community cancer network, a greater proportion of PV patients were treated at community sites, where patients were older and with higher risk disease by both age and history of TEs. CCC patients were more likely to receive cytoreductive therapy and to receive agents other than hydroxyurea for second line therapy. Despite comparable OS, NAC patients had higher risk of post-PV TEs and analyses for association of patient and disease-related factors are ongoing. Our study highlights the need for enhanced collaboration between leukemia specialists and community-based providers to maximize quality of PV care.