Role of molecular alterations in transplantation decisions for patients with primary myelofibrosis

Abstract

The aim of our study was to analyze the potential survival benefit associated with hematopoietic stem cell transplantation (HSCT) according to clinicobiological scores, which incorporate mutation-enhanced international prognostic score system (MIPSS) to facilitate decision-making in this context. One transplant (n = 241) and 1 nontransplant cohort (n = 239) were used to test the hypothesis that patients with primary myelofibrosis with higher risk molecular score benefit from HSCT. A weighted propensity score was applied to balance confounding factors with the transplanted cohort as reference. Weighted Cox proportional hazard models and logistic regression analyses were performed. Overall, 105 patients who did not receive transplant could be matched to the 239 patients who did receive transplants. HSCT was associated with a higher 6-year overall survival rate in intermediate-2 (60.1% vs 41.5%) and high-risk DIPSS patients (44.4% vs 6.55%), high-risk MIPSS70 (46.5% vs 23.9%), high-risk (73.2% vs 39.7%) or very high-risk MIPSS70+V2 (51.8% vs 24%). Patients with intermediate MIPSS70 scores have an advantage of survival with HSCT only when their myelofibrosis transplant scoring system (MTSS) were low or intermediate. Patients who received transplant had an increased mortality risk the first year, but a significant benefit with HSCT after the 1-year landmark was observed in higher risk patients. This study confirms that, similar to DIPSS, MIPSS70 and MIPSS70+V2 risk score in addition to MTSS can be used to determine which patients with primary myelofibrosis have survival benefit from HSCT over non-HSCT strategies.

Graphical abstract

Figure 1.

Matching cohorts and reclassification between DIPSS and MIPSS70. (A) Love plot showing how matching largely erased the differences between the 2 groups. (B) Sankey diagram represents the reclassification of DIPSS categories into MIPSS70 scoring system.

Figure 2.

OS according to HSCT and DIPSS classification. (A) OS in the global cohort was depicted according to HSCT and (B) with a landmark at 1-year of follow-up. (C) OS was represented according to DIPSS categories and HSCT (supplemental Figure 4 for landmark analysis). The numbers of individuals at risk for the nontransplant cohort are fractions because of the inclusion of weighting in the analysis.

Figure 3.

Impact of HSCT on molecular subtypes. (A) OS was depicted according to HSCT and MIPSS70, and (B) the presence of HMR mutation (landmark analysis in supplemental Figure 5). OS was represented in patients with (C) MIPSS70 intermediate and (D) high according to the MTSS score for patients who received transplantation. The number of individuals at risk for the nontransplanted cohort are fractions because of the inclusion of weighting in the analysis.

Figure 4.

Impact of HSCT in every category of prognostic scoring systems and according to the presence of HMR. Forest plot representing the hazard ratio (HR) of HSCT in OS in each category. Analyses were performed in the 2 periods of the landmark analysis: 0 to 1 year (left panel) and 1 to 8 years (right panel). P values were corrected using the Bonferroni-Holm method. For some subgroups, results were not reported (NA) because of an estimation error.

Figure 5.

Subanalysis of patients with available cytogenetics. A new matching was performed in patients with available karyotype. (A) Sankey diagram represents the reclassification of DIPSS categories into MIPSS70+V2 scoring system. (B) OS was depicted according to HSCT and MIPSS70+V2 (supplemental Figure 10 for landmark analysis). The numbers of individuals at risk in the nontransplant cohort are fractions because of the inclusion of weighting in the analysis.