Prognostic scores for patients with chronic myeloid leukemia under particular consideration of competing causes of death

Abstract

Nowadays in many fields of medicine, prognostic scores are used to predict the outcome for individual patients. In chronic myeloid leukemia (CML), the Sokal, the Euro, and the EUTOS score are established prognostic scores which were addressed by the CML management recommendations of the European LeukemiaNet. This review provides a general definition of prognostic scores and explains their meaning. Main differences between the Sokal, the Euro, and the EUTOS score are highlighted. Due to the therapeutic success of tyrosine kinase inhibitors, the proportion of patients with causes of death unrelated to CML is growing. To assess the potential of a drug to prevent dying of CML, causes of death unrelated to CML need to be considered as competing risks. Supported by data of patients randomized to imatinib-based treatments within the German CML study IV, this review also explores the prognostic performance of the established scores if the primary event is death due to CML only and explains the implicit statistical particularities when treating other causes of death as competing risks. In the presence of competing risks, the application of both the cause-specific hazard model and the subdistribution hazard model is recommended when investigating the influence of prognostic factors on the event of interest. Another purpose of this work is to foster the ability of hematologists to interpret the outcome of a cause-specific hazard and a subdistribution hazard model and to understand the differences between them.

PTCH1 is a reliable marker for predicting imatinib response in chronic myeloid leukemia patients in chronic phase

“…We employed the cumulative incidence procedure and the Fine-Gray test to estimate and compare the probabilities of IFFS, PFS, and CML OS for the high and low gene expression groups [41]. Non-CML-related deaths and changes from imatinib due to toxicity were considered competing risks in the analysis of IFFS.…”

“…In order to evaluate a prognostic test, it is important to test its performance in a cohort as similar as possible to the real scenario [28]. In addition, because CML is a chronic disease, competing risk has been accepted as a better option for analysis [29]. Finally, employing more control genes may produce a more consistent result, but that strategy is quite time-consuming, which restricts its applicability [30].…”

Prognostic Scoring Systems in CML ? Improvement Impossible?

“…Clearly, these attempts were successful or not depending statistical methods used (mostly Cox multivariate analysis and more recently causespecific or sub distribution hazard models) and the numbers of patients included into retrospective studies designed for prognostic score estimation. Importantly, in addition to that these groups of patients had to be treated with homogenous treatment able to achieve similar results [1]. The result of this approach is usually the number which value allocates the individual patient to the specific prognostic group ("low", "intermediate" or "high" risk).…”

“…Various strategies were proposed: 1/ to use different endpoints because more CML patients treated successfully with TKI are dying from different causes than CML. Logically the German group of statisticians analyzed with this aim 1236 patients treated within the German CML Study IV and find out that the best model for prediction of CML-related deaths may be based on age, gender and spleen size [1]; 2/ to combine existing scoring systems with the results of early molecular response (proved to correlate with further response and survival in may trials) [14]; 3/ combination the various existing scores within a new one. The group of Korean scientists recently confirmed different probabilities of achieving complete cytogenetic response at 12 month for the groups of patients with combination of low scores and for the patients who have all three scores (Sokal, Euro and EUTOS) in the high-risk group categories [15]; 4/ to use novel biological or molecular markers that may further improve the prognostication [16,17].…”

Differential impact of additional chromosomal abnormalities in myeloid vs lymphoid blast phase of chronic myelogenous leukemia in the era of tyrosine kinase inhibitor therapy

“…5,6 The long-term cumulative probability of progression of CML from CP to BP is only about 5%. 7,8 Despite the revolutionary progress in the treatment of patients with CML-CP, CML-BP remains a therapeutic challenge. In general, CML-BP is a fatal disease in the era of TKI therapy, with a median survival of only 6–10 months.…”