Predictive Factors for Overall Survival in Chronic Myeloid Leukemia Patients: An Analysis By the Gimema Cml Italian Study

Introduction

Tyrosine kinase inhibitors (TKIs) have dramatically changed the outcome of chronic phase chronic myeloid leukemia patients (CP-CML),improving the long-term outcome; indeed, life expectancy is now very close to that of age matched individuals in the general population. Imatinib (IMA) the first generation TKI, increased the overall survival(OS)by more than 80%. Second-generation TKIs, (2gen TKIs) used in first line dasatinib and nilotinib, induced faster molecular responses, rapidly reducing the disease burden, but did not change the OS of CP-CML newly diagnosed patients. Most of the available data reported were extrapolated from sponsored clinical trials. The aim of this analysis is to detail and analyze the prognostic features influencing the OS in a large Italian CML cohort of patients prospectively enrolled in the GIMEMA CML Italian study.

Methods

Relevant clinical, demographic, biological and treatment-related information were web-based collected during a multicenter,observational Italian study that enrolled consecutive patients in each disease phase, at 68 Italian hematologic centers belonging to the GIMEMA network from January 2013 to June 2020. We analyzed prognostic factors influencing the OS by Kaplan Meier method and Cox multivariable models.

Results

A cohort of 1206 patients was prospectively analyzed, 608 of them received frontline IMA and 598 2genTKIs. Median age in the IMA cohort was 69 years (range 58-77) vs 52 years in the 2genTKIs cohort (range 41-63). The male/female ratio was 1.7 in the IMA group and 1.35 in the 2genTKIs cohort. Ninety-eight percent of patients were in CP. Results of molecular analysis of the BCR-ABL transcript at baseline showed: b2a2 in 33.1 % of patients and b3a2 in 59.9%, while an atypical transcript was found in 7%. The cytogenetic analysis at baseline showed major and minor additional aberrations in 5.7% and 1.6% of patients respectively. In the IMA cohort,according to the Sokal score, 27.7%, 57.3% and 15% of patients were stratified as low, intermediate and high risk, whereas according to the ELTS score 51.3%, 35.5% and 13.3%, of patients were classified as low, intermediate and high risk. In the 2genTKIs cohort, according to the Sokal score, 44.8%, 34.5% and 20.8%, were low, intermediate and high risk, respectively, whereas according to the ELTS score, 66.9%, 22% and 11% were assigned to the respective risk groups.The Charlson comorbidity index in the IMA cohort was 2-3 and 4-5 in 74% and 26% of patients respectively; in the 2genTKIs cohort the score was 2-3 in 89% and 4-5 only in 10% of patients. Overall, median follow-up of the whole population was 24.7 months (range 13.3-39.3).Seventy-three patients (6.1%) in the overall population died, the majority of them in the IMA cohort: 56 patients (9.2%), at median age of 80.5 years,11/608 (1.8%) due to CML-related causes. Conversely,in the 2genTKIs cohort only 17 patients (2.8%) died, at a median age of 62 years, 10/598 (1.7%) for CML-related causes. Estimated 60-months OS of the overall population was 86.4% (95% CI 81.3-90.2): 75.8% (95% CI 64.5-84) in the IMA cohort and 93.8% (95% CI 87.5-97) in the 2genTKIs group (p<0.0001). The ELTS score provides a better stratification of 60 months OS both in the IMA (OS 60-months 91.4%, 65.2%, 48.7% in low, intermediate and high risk, p<0.0001) (Fig.1a) and the 2genTKIs subgroups (OS 60-months OS 95.4%, 92%, 87.9% in low, intermediate and high risk, p=0.0013)(Fig.1b). An adjusted Cox model on the entire population showed that prognostic factors influencing OS are: ELTS score (high risk vs low HR= 5.2, 95%CI 2.7-10.03, p<0.0001),the type of TKI (2genTKIs vs IMA HR= 0.46; 95% CI 0.24-0.87, p=0.018), age (HR=1.03 per year, 95%CI 1.00-1.05, p=0.025) and the Charlson index (4-5 vs 2-3, HR= 1.75, 95%CI 1.43-2.1, p<0.0001).

Conclusions

In this first analysis of our study different clinical behaviors were observed among Italian hematologists, who prevalently prescribed IMA to older patients,with more comorbidities, as compared to 2genTKIs.These differences explain a better OS for patients treated with 2genTKIs vs IMA, however, the risk of death for CML related causes is quite similar between the two groups, all the differences being attributable to other causes of death.Prognostic baseline features associated to an increased OS confirmed that, in addition to age, the ELTS score and the comorbidities are the main clinical factors that independently influence the long-term OS