Can the Use of Bone Marrow Parameters Improve the Efficacy of Risk Prediction Scores in Chronic Myeloid Leukemia in Imatinib Era?

Abstract

Many attempts have been made to develop risk prediction scores for chronic myeloid leukemia in chronic phase (CML-CP) to identify the subgroup with a poorer response to therapy to enable early intensification of treatment. Because the bone marrow (BM) provides a more sensitive reflection of the disease process, we hypothesized that using BM parameters in the Sokal and European Treatment and Outcome Study (EUTOS) risk scores could improve their efficacy in an imatinib-treated population. We analyzed cases of CML-CP for their response and survival outcomes with imatinib using risk groupings determined by the Sokal and EUTOS scores using peripheral blood (PB) or BM parameters (Sokal-PB, Sokal-BM, EUTOS-PB, and EUTOS-BM). A total of 371 cases were analyzed. The concordance for risk groups was greater for the EUTOS scores (81.9%) than for the Sokal scores (68.1%) using PB versus BM parameters. For all 4 risk scores, the predictive efficacy was statistically significant. EUTOS-PB and EUTOS-BM could better prognosticate for progression-free survival (PFS) and overall survival (OS) between the low- and high-risk groups (P< .0001). However, with the Sokal risk score, the use of BM parameters improved the prognostic capacity for PFS between the low- and intermediate-risk groups, with a statistically significant difference (P= .025), but not for OS (P= .88). The use of BM parameters, a simple method that is feasible in routine clinical practice could improve the prognostic efficacy of the Sokal score for PFS but not for OS in low- and intermediate-risk groups. Further research to improve the sensitivity of risk scores for CML-CP prognosis and attempts at risk-directed therapy is warranted.