Refined iRR6 Model Improves Risk Stratification in Intermediate-1 Myelofibrosis

An international effort has refined the early prognostic assessment of myelofibrosis (MF) patients treated with ruxolitinib, emerging with the development of the revised iRR6 model. Published in Cancer, this new model, a subanalysis of the large RUX-MF study (NCT06516406), provides a more precise method for identifying patients at intermediate-1 risk who may benefit from early therapy adjustments.1 

The iRR6 may help identify those who are unlikely to derive long-term benefit from ruxolitinib and could be considered for alternative therapies. | Image credit: mdaros - stock.adobe.com

MF, a Philadelphia-negative chronic myeloproliferative neoplasm, often presents challenges in identifying patients at higher risk for early mortality. The research builds upon the previously developed "response to ruxolitinib after six months (RR6)" score, which was designed to identify patients likely to experience poor long-term outcomes despite early therapy.2 While the RR6 model demonstrated prognostic separation in mixed-risk cohorts, investigators note that it was "less applicable to lower-risk patients."1 As the authors explain, "In the 428 intermediate‑1 patients, the RR6 model did not discriminate between intermediate and low‑risk patients (5‑year OS, 74.4% vs. 72.0%; P = .24)." Only the high-risk group within this intermediate-1 cohort still demonstrated "a significantly poorer 5-year OS of 57.4% (P < .01).

To address this limitation, investigators developed the intermediate-1 specific RR6 (iRR6) model by incorporating 3 new prognostic variables, each of which was significantly associated with overall survival (OS). These included underdosing of ruxolitinib relative to platelet count at one or more time points (HR, 3.91; P < .001), absence of ≥50% palpable spleen reduction at 6 months (HR, 1.45; P = .02), and red blood cell transfusion requirement at all time points (HR, 1.85; P = .01). Scoring within the iRR6 placed patients into three survival categories: low-risk (score 0), intermediate-risk (1–2), and high-risk (>2). This refined scoring system was then applied to the 428 intermediate-1 patients, effectively stratifying them into 3 distinct risk categories with significantly different survival outcomes. The low-risk group (20.3% of patients) achieved an 5-year OS of 84.8%, the intermediate-risk group (45.8% of patients) had a 5-year OS of 76.4%, and the high-risk group (33.9% of patients) showed a 5-year OS of 56.6% (P < .0001).

Validation in an external cohort of 95 intermediate-1 risk patients treated at the Moffitt Cancer Center confirmed the findings. In this external cohort, 5-year OS was 83.3% for low-risk, 71.7% for intermediate-risk, and 54.5% for high-risk patients (P = .01). The distribution of predictive factors in this validation cohort was comparable, with 55.8% failing to achieve ≥50% spleen reduction and 37.9% receiving underdosed ruxolitinib at 6 months.

These findings hold significant implications for clinical practice. The study emphasizes that "optimal treatment of intermediate-1 risk patients is a challenge," as they are often not considered for allogeneic stem cell transplantation despite potentially benefiting from it. The iRR6 model now provides a valuable tool to "identify the subset of intermediate-1 risk patients who are unlikely to respond well so that we can optimize their therapy (or otherwise refer them for transplantation)."

Furthermore, the research highlights the importance of optimal ruxolitinib dosing, noting that "Underdosing of ruxolitinib, often due to concerns over anemia, has been reported in previous studies in a substantial fraction of patients. Notably, the underdosing of ruxolitinib was associated not only with reduced response rates but also with poorer survival outcomes."

Another important finding was the prognostic value of a ≥50% reduction in spleen size at 6 months. The new model suggests that a 50% spleen reduction is a more meaningful indicator of disease control for this patient population than the previous 30% indicator.

The study reinforces that intermediate‑1 risk patients, despite their younger age and less severe presentation compared with higher-risk MF, remain highly heterogeneous. "Optimal treatment of intermediate‑1 risk patients is a challenge. They were historically not included in the COMFORT studies, yet they benefit from ruxolitinib therapy, with higher response rates, fewer cytopenias, and lower discontinuation rates. Furthermore, they constitute majority of MF patients receiving ruxolitinib in a real‑life context. Finally, they are frequently not considered for transplantation, despite being younger patients who could potentially benefit more from the procedure," the authors explain.

The investigators acknowledge the study's limitations, including its retrospective design, reliance on spleen palpation for response assessment, variability in adherence data, and incomplete molecular profiling due to the long enrollment period. However, the external validation completed at Moffitt aims to strengthen the model's generalizability. "A high degree of similarity was observed in the distribution of iRR6 variables between the Moffitt cohort and the original RUX‑MF cohort, suggesting that the iRR6 model can be applied broadly to intermediate‑1 patients treated with ruxolitinib, regardless of geographic or institutional disparities," the authors report.

By integrating treatment-related factors such as dose intensity, transfusion needs, and depth of spleen response, the iRR6 may help identify those who are unlikely to derive long-term benefit from ruxolitinib and could be considered for alternative therapies, including clinical trials or early transplantation.

References

1. Palandri F, Branzanti F, Bonifacio M, et al. Revised "iRR6" model in intermediate-1 risk myelofibrosis patients treated with ruxolitinib. Cancer. 2025;131(17):e70062. doi:10.1002/cncr.70062

2. Maffioli M, Mora B, Ball S, et al. A prognostic model to predict survival after 6 months of ruxolitinib in patients with myelofibrosis. Blood Adv. 2022;6(6):1855-1864. doi: 10.1182/bloodadvances.2021006889