Allogeneic hematopoietic stem cell transplantation is the sole potentially curative treatment option for selected patients with myelofibrosis and is associated with substantial risks and potential benefits. Shared decision making with regard to transplant is informed by disease risk models, but additional considerations extend beyond disease risk alone.

Allogeneic hematopoietic stem cell transplantation remains the only potentially curative treatment for individuals with myelofibrosis. It is an intensive procedure associated with substantial risks—but also considerable potential benefit. The decision to proceed with a transplant and the determination of optimal timing depend on a number of factors.

The first consideration is when a patient should be referred for transplant evaluation. I recommend early referral, which is typically once a patient reaches intermediate-1 risk myelofibrosis and, in some cases, even while the disease is classified as low risk. Transplants require significant resources, and patients are better served by having adequate time to prepare rather than having to arrange logistics under time pressure. In addition, the volume of information involved in understanding transplantation is considerable, and earlier referral gives patients the time that is needed to fully understand the process.

I approach the discussion of allogeneic hematopoietic stem cell transplantation with patients in a systematic manner, beginning with a review of the methods used to assess risk in myelofibrosis, including the available prognostic scoring systems. Although numerous studies have identified specific risk groups for whom a transplant should or should not be considered, the decision is more complex than risk category alone. In general, I think that these scoring systems are good guide rails, but I still do not think that we fully understand how to use the data they provide.

Several prognostic scoring systems have been developed over time. The International Prognostic Scoring System (IPSS) was introduced first, followed by the Dynamic International Prognostic Scoring System (DIPSS) and the DIPSS Plus. These systems were derived from clinical variables that were readily available at the time of their development, including hemoglobin level, white blood cell count, circulating blast percentage, constitutional symptoms, and age; DIPSS Plus additionally incorporates karyotype, transfusion dependence, and thrombocytopenia. As our understanding of the molecular heterogeneity of these diseases has advanced, the Mutation-Enhanced International Prognostic Score Systems (ie, MIPSS70 and MIPSS70+) were developed to incorporate both driver mutation status (JAK2, CALR, or MPL) and somatic nondriver mutation status.

As with most prognostic models, these systems were designed to identify disease-associated risk factors, generate a composite score, and estimate survival. Survival estimates derived from these models should be interpreted with caution, as they may be misleading. The IPSS, DIPSS, and DIPSS Plus were developed before JAK inhibitors became clinically available; given that JAK inhibitors have been associated with improved survival in myelofibrosis, the survival estimates from these models are likely no longer accurate. The MIPSS70+ was validated in a high-risk cohort treated at a tertiary referral center, which may not be representative of the broader patient population, and the model places substantial weight on molecular findings. Furthermore, these systems were validated in primary myelofibrosis and not in myelofibrosis arising from polycythemia vera or essential thrombocythemia. For secondary myelofibrosis, the Myelofibrosis Secondary to PV and ET Prognostic Model (MYSEC-PM) may be used, although it does not incorporate nondriver mutations.

Nonetheless, the clinical variables underlying these models remain highly relevant. In evaluating a patient for a transplant, I consider both the individual clinical features incorporated into these systems and the overall risk category that each system assigns. Accordingly, I calculate several prognostic scores at the time of initial consultation. A frequent challenge arises when these scores are discordant, with one system assigning low risk and another assigning high risk. For example, a patient with normal blood counts and a high-risk mutation would be classified as high risk by the MIPSS70. In such cases, recommending a transplant can be difficult, both for the physician and for the patient, given the absence of hematologic abnormalities.

The Myelofibrosis Transplant Scoring System (MTSS) is another useful tool. In addition to established risk factors such as age and ASXL1 mutation status, the MTSS incorporates donor type. One practical approach is to use the MIPSS70 and the MTSS together; a patient with intermediate-risk disease according to the MIPSS70 and a low-risk score according to the MTSS is generally a favorable candidate for a transplant.

Additional considerations extend beyond disease risk alone. Timing is one such consideration. There is often a window during which a transplant can be performed with comparable outcomes, depending on patient- and disease-related factors, and shared decision making is essential during this period. Some patients prefer continued observation, while others wish to proceed promptly. Neither approach is inherently correct, and the decision is highly individual. However, this window has limits at both ends of the risk spectrum. For patients with low-risk disease, a transplant is generally premature. Conversely, for patients with very high-risk disease, I recommend proceeding directly to a transplant. Features that typically prompt this recommendation include transfusion-dependent anemia, thrombocytopenia, and a rapidly rising white blood cell count or blast percentage.

Psychosocial factors may also influence the decision. Transplantation requires a dedicated caregiver who is available at all times, often necessitates that both the patient and the caregiver leave their employment for a period, and may require a patient who lives a great distance from a transplant center to relocate. Social workers can frequently assist in identifying caregivers and in applying for financial assistance; nevertheless, these psychosocial demands remain a considerable challenge for many patients.

Outcomes of the transplant itself have improved in several respects. For example, the incidence of graft-vs-host disease has declined with the adoption of newer prophylactic strategies such as posttransplant cyclophosphamide. Moreover, the treatment of steroid-refractory graft-vs-host disease has improved with agents such as ruxolitinib. In addition, an expanded range of antimicrobial agents is now available for the prevention and treatment of infection.

In summary, appropriately selected patients with myelofibrosis can derive substantial benefit from an allogeneic hematopoietic stem cell transplant. Many factors inform the decision of who should undergo a transplant and when, and, in many cases, this determination requires a shared decision-making process between the physician and the patient. Importantly, outcomes following a transplant continue to improve, including among patients with high-risk disease, and transplantation represents a realistic and increasingly successful treatment approach for many patients with myelofibrosis.