The prognostic contribution of CBL, NRAS, KRAS, RUNX1, and TP53 mutations to mutation-enhanced international prognostic score systems (MIPSS70/plus/plus v2.0) for primary myelofibrosis

Contemporary risk models in primary myelofibrosis (PMF) include the mutation (MIPSS70) and mutation/karyotype enhanced (MIPSS70 plus/v2.0) international prognostic scoring systems. High molecular risk (HMR) mutations incorporated in one or both of these models include ASXL1, SRSF2, EZH2, IDH1/2, and U2AF1Q157; the current study examines additional prognostic contribution from more recently described HMR mutations, including CBL, NRAS, KRAS, RUNX1, and TP53. In a cohort of 363 informative cases (median age 58 years; 60% males), mutations included JAK2 61%, CALR 24%, MPL 6%, ASXL1 29%, SRSF2 10%, U2AF1Q157 5%, EZH2 10%, IDH1/2 4%, TP53 5%, CBL 5%, NRAS 7%, KRAS 4%, and RUNX1 4%. At a median follow-up of 4.6 years, 135 (37%) deaths and 42 (11.6%) leukemic transformations were recorded. Univariate analysis confirmed significant survival impact from the original MIPSS70/plus/v2.0 HMR mutations as well as CBL (HR 2.8; p < .001), NRAS (HR 2.4; p < .001), KRAS (HR 2.1; p = .01), and TP53 (HR 2.4; p = .004), but not RUNX1 mutations (HR 1.8; p = .08). Multivariate analysis (MVA) that included both the original and more recently described HMR mutations confirmed independent prognostic contribution from ASXL1 (HR 1.8; p = .007), SRSF2 (HR 4.3; p < .001), U2AF1Q157 (HR 2.9, p = .004), and EZH2 (HR 2.4; p < .001), but not from IDH1/2 (p = .3), TP53 (p = .2), CBL (p = .3), NRAS (p = .8) or KRAS (p = .2) mutations. The lack of additional prognostic value from CBL, NRAS, KRAS, RUNX1, and TP53 was further demonstrated in the setting of (i) MVA of mutations and karyotype, (ii) MVA of MIPSS70/plus/v2.0 composite scores and each one of the recently described HMR mutations, except TP53, and iii) modified MIPSS70/plus/plus v2.0 that included CBL, NRAS, KRAS, and TP53 as part of the HMR constituency, operationally referred to as "HMR+" category. Furthermore, "HMR+" enhancement of MIPSS70/plus/plus v2.0 did not result in improved model performance, as measured by C-statistics. We conclude that prognostic integrity of MIPSS70/plus/plus v2.0, as well as their genetic components, was sustained and their value not significantly upgraded by the inclusion of more recently described HMR mutations, including CBL, NRAS, KRAS, and RUNX1. Additional studies are needed to clarify the apparent additional prognostic value of TP53 mutation and its allelic state.