Classical JAK2V617F+ Myeloproliferative Neoplasms emergence and development based on real life incidence and mathematical modeling

Mathematical modeling allows us to better understand myeloproliferative neoplasms (MPN), a group of blood cancers, emergence and development. We test different mathematical models on an initial cohort to determine the emergence and evolution times before diagnosis of JAK2V617F+ classical MPN (Polycythemia Vera (PV) and Essential Thrombocythemia (ET)). We consider the time before diagnosis as the sum of two independent periods: the time (from embryonic development) for the JAK2V617F mutation to occur, not disappear and enter proliferation, and a second time corresponding to the expansion of the clonal population until diagnosis. We prove that the rate of active mutation occurrence increases exponentially with age following the Gompertz model rather than being constant. We find that the first tumorous cell takes an average time of $63.1 \pm 13$ years to appear and start proliferation. On the other hand, the expansion time is constant: $8.8$ years once the mutation has emerged. These results are validated in an external cohort. Using this model, we analyze JAK2V617F ET versus PV, and obtain that the time of active mutation occurrence for PV takes approximately $1.5$ years more than for ET to develop, while the expansion time was similar. In conclusion, our age-dependent approach for the emergence and development of MPN demonstrates that the emergence of a JAKV617F mutation should be linked to an aging mechanism, and indicates a $8-9$ years period of time to develop a full MPN.

💡 Research Summary

**
This study applies mathematical modeling to elucidate the timing of emergence and subsequent growth of classical JAK2V617F‑positive myeloproliferative neoplasms (MPNs), specifically polycythemia vera (PV) and essential thrombocythemia (ET). The authors decompose the interval from birth to clinical diagnosis (TM) into two independent components: T1, the time from embryogenesis until a JAK2V617F‑bearing stem cell both appears and acquires proliferative capacity, and T2, the period required for that clone to expand to a diagnostic mass of roughly 10¹² cells (≈10⁶ cell divisions).

A key innovation is the abandonment of the traditional assumption of a constant mutation rate. Instead, the “active mutation rate” δ (the product of the underlying mutation rate τ and the probability p that a mutation leads to a proliferative clone) is modeled as age‑dependent, following a Gompertz function δ(t)=A·exp(k·t). This captures the empirically observed increase in MPN incidence with age. For T2, two alternatives are considered: a fixed expansion time α (reflecting the averaging of many cell‑division intervals) and a log‑normal distribution (allowing for stochastic diagnostic delays).

Parameter estimation is performed via a generalized Expectation‑Maximization (EM) algorithm on two independent cohorts: a regional French cancer registry (264 patients, 1990‑2020) and the national BCB‑FIMBANK dataset (1,111 patients). Age‑specific mortality rates are used to adjust raw incidence counts, ensuring that only individuals who survived to the age of diagnosis are considered.

Model comparison using chi‑square goodness‑of‑fit tests favors the simpler Model A.1 (Gompertz δ, fixed T2=α). The estimated parameters are A≈0.0012, k≈0.09, and α≈8.8 years. Consequently, the average age at which the first tumor‑initiating cell appears is 63.1 ± 13 years, while the subsequent clonal expansion to a detectable disease state consistently takes about 8.8 years across patients.

When PV and ET are analyzed separately, the model reveals that the active mutation occurrence for PV is delayed by roughly 1.5 years relative to ET, whereas the expansion time remains indistinguishable between the two diseases. This aligns with known biological differences: PV often involves homozygous JAK2V617F, requiring an extra mutational step, whereas ET typically harbors a heterozygous mutation.

The study’s strengths include (1) a biologically plausible age‑dependent mutation model grounded in Gompertz dynamics, (2) rigorous EM‑based parameter inference, and (3) validation across independent real‑world datasets. Limitations are acknowledged: the reduction of complex mutational processes to a single δ parameter, the assumption of a fixed diagnostic cell count, and the omission of co‑occurring driver mutations (e.g., TET2, ASXL1) and treatment effects.

Future directions suggested by the authors involve extending the framework to incorporate additional genetic lesions, patient‑specific clonal fitness measures, and therapeutic interventions, thereby enabling personalized risk prediction and earlier detection strategies. By quantifying both the long latency before a JAK2V617F clone becomes active and the relatively short, uniform expansion phase, the work provides a valuable quantitative foundation for understanding MPN natural history and for designing age‑targeted screening programs.