Researchers have developed a new risk model for primary myelodysplastic syndromes (MDS) that integrates genetic and clinical information.
The research team considered the current standard for prognostication—the revised International Prognostic Scoring System (IPSS-R)—to be too complex, limited to newly diagnosed cases, and missing information on mutations and age.
So they devised a “simpler and more contemporary” prognostic system, the Mayo Alliance Prognostic Model for MDS.
The team, from the Mayo Clinic in Rochester, Minnesota, and the National Taiwan University Hospital (NTUH), described the new model in Mayo Clinic Proceedings.
Lead author Ayalew Tefferi, MD, of the Mayo Clinic, said the new model “is not an enhancement of the international prognostic scoring system tool, it's a complete makeover."
The team analyzed mutation information from 357 patients with primary MDS or leukemic transformation treated at the Mayo Clinic from the end of December 1994 through mid-December 2017.
The patients were a median age of 74 and 70% were males.
They compared the Mayo patients to 328 NTUH patients, who were a median age of 66 and 65% were males.
Multivariate analysis of the Mayo cohort identified the following as predictors of inferior overall survival:
- Monosomal karyotype (hazard ratio [HR], 5.2; 95% CI, 3.1-8.6)
- Non-monosomal karyotype abnormalities other than single/double del(5q) (HR, 1.8; 95% CI, 1.3-2.6)
- RUNX1 (HR, 2.0; 95% CI, 1.2-3.1)
- ASXL1 (HR, 1.7; 95% CI, 1.2-2.3) mutations
- Absence of SF3B1 mutations (HR, 1.6; 95% CI, 1.1-2.4)
- Age greater than 70 years (HR, 2.2; 95% CI, 1.6-3.1)
- Hemoglobin level less than 8 g/dL in women or less than 9 g/dL in men (HR, 2.3; 95% CI, 1.7-3.1)
- Platelet count less than 75 x 109/L (HR, 1.5; 95% CI, 1.1-2.1)
- 10% or more bone marrow blasts (HR, 1.7; 95% CI, 1.1-2.8)
They then provided values to reflect the prognostic contribution of each of the above predictors and devised the new 4-tiered Mayo prognostic model.
Median 5-year overall survival rates in the 4 categories in the Mayo model were 73% (low risk), 34% (intermediate-1), 7% (intermediate-2), and 0% (high risk; 9-month median survival).
The team then validated the Mayo alliance model by using the NTUH cohort and compared it to the IPSS-R.
The investigators were able to confirm superior predictive accuracy of their model and a substantial discordance between the the Mayo model and the IPSS-R in terms of the pattern of risk distribution.
Examples of discordance included:
- More than 25% of patients belonging to the high-risk category according to the Mayo alliance model were classified as IPSS-R low or intermediate risk
- Almost 50% of patients with intermediate-2 risk category according to the Mayo alliance model were classified as IPSS-R very low or low risk
- Almost 50% of patients with IPSS-R very low risk were classified as intermediate-2 or intermediate-1 risk according to the Mayo alliance model
The authors wrote that this “suggests a fundamental and not incremental advantage for the new Mayo alliance model.”