Cancers in Polycythemia Vera and Essential Thrombocythemia: A before and after Diagnosis Survey and Analysis of Risk Factors

Background

An increased risk of second cancers (SC), in patients with myeloproliferative neoplasms (MPN), has been suggested by a number of population-based and cancer registry studies (Blood 2011;118:6515). Such studies often included patients belonging to a broad category of MPN entities and did not always account for history of cancer prior to the diagnosis of MPN, associated comorbidities, mutations, or treatments received, in assessing risk association and prediction (Leukemia 2019;33:1996). The current study seeks to fill knowledge gaps, in this regard, in the context of a biologically more homogeneous MPN setting, vis-ẚ-vis polycythemia vera (PV) and essential thrombocythemia (ET).

Methods

The present study was conducted under an institutional review board approved minimum risk protocol that allowed retrospective collection and analysis of data from Mayo Clinic patient records. Diagnostic criteria were according to the International Consensus Classification (Blood 2022; 140:1200). Documented primary cancers were annotated in 17 categories, not including pre-cancerous conditions or MPN progressions. SC was defined as any cancer that developed after 3 months from the initial diagnosis of MPN. Cox proportional hazard regression model was used for univariable and multivariable analysis. Kaplan-Meier plots were used for time-to-event assessment (considering the first in case of multiple events).

Results

The study population included 1,968 consecutive patients (median age 60 years; 56% females) with PV (N=1,001) or ET (N=967), seen at the Mayo Clinic between 1970 and 2024. Median follow-up was 8.6 years. The incidences of arterial/venous thrombosis were14%/12% at diagnosis and 12%/10% after diagnosis. Co-morbid conditions included diabetes (12%), hypertension (52%), tobacco use (16%), and history of autoimmune/inflammatory disease (AID; 6%); hyperlipidemia was annotated in PV cases only (51%). Non-driver mutation information was available in 620 patients: TET2 (13%), ASXL1 (7%), DNMT3A (5%), SRSF2 (3%), SF3B1 (3%) and TP53 (2%). A history of another cancer was documented in 598 (30%) patients and included cancers documented before/at (N=291; 15%) or after (N=404; 20%) the diagnosis of MPN, with only the latter being referred to as “SC”. Most frequent cancer type was non-melanoma skin cancer (NMSC), accounting for 28% of cancers diagnosed before/at MPN, and 47% of SC; other SC types included lymphoid neoplasms (10%), melanoma (10%), prostate cancer (9%), and trachea/lung/bronchial cancer (8%), as the most frequent. On multivariable analysis, cancer history, before/at or after MPN diagnosis, clustered with age >60 years (OR 2.7), male sex (OR 1.5), and history of AID (OR 1.7).

History of cancer prior to MPN diagnosis, NMSC or otherwise, was associated with inferior overall survival (p<0.01); by contrast, neither myelofibrosis-free (p=0.5) nor leukemia-free (p=0.8) survival was affected. Median time from MPN diagnosis to SC was 5.7 years. In multivariable analysis including all cancers, SC-free survival was adversely affected by age>60 years (HR 2.5; p<0.01), male sex (HR 1.7; p<0.01), history of previous cancer (HR 2.5, p<0.01), and AID history (HR 1.6; p<0.01); all these 4 risk factors remained significant after excluding NMSC. Interestingly, among informative cases, SRSF2 mutation also predicted SC (HR 2.6; p=0.02), even after NMSC cases were excluded (HR 3.4; p=0.02). Pre-MPN history of cancer did not affect either arterial (p=0.9) or venous thrombosis-free survival (p=0.7), in age- and sex-adjusted analysis. Conversely, after excluding post-SC thrombotic events, venous (p=0.02) but not arterial (p=0.1) events, occurring after MPN diagnosis, were predictive of SC. This association remained significant in multivariable analysis that included age, sex, and venous thrombosis history (p=0.03). Additional observations from this study, include an independent association between SC and hyperlipidemia in PV (p<0.01) and ruxolitinib monotherapy with NMSC (HR 9; p<0.01).

Conclusions

The risk of SC in patients with MPN is significantly higher in older patients, men, and in the presence of pre-MPN cancer history and AID. Incidental venous, but not arterial, thrombosis was associated with subsequent SC, not NMSC. The observed association of SC with SRSF2 mutations and hyperlipidemia is noted but requires further validation.