CML Patients Who Fail to Achieve ≤ 1% BCR::ABL1 at 12 Months or ≤ 0.1% BCR::ABL1 at 24 Months (major molecular response, MMR) Show Significantly Inferior Survival, but Not Due to CML-Related Deaths: Data from an Italian Real-Life Observational Study

Introduction

The achievement of CCyR (complete cytogenetic response) or of the corresponding BCR::ABL1 ≤1% molecular value is generally accepted as the goal to be achieved for a good survival, but the time to achieve this goal and the relevance for survival of achieving deeper molecular responses such as MMR (BCR::ABL1≤0.1% value) are still under discussion. In order to clarify these aspects we studied the database of the Italian CML Network, that prospectively enroll all adult CML patients from 68 Italian hematology centers since January 2013 (https://www.epiclin.it/lmc).

Methods and Results

A total of 1433 patients with updated follow-up data were included in this analysis. Baseline information included sociodemographic, clinical, and standard laboratory data. 702 patients were treated with first-line imatinib and 731 with second-generation (2^ndG) TKIs (414 with nilotinib, 312 with dasatinib and 5 with bosutinib). The two cohorts show a relevant difference in median age: 70 years for the imatinib cohort and 52 years for the 2^ndG cohort. The ELTS risk score stratification is also different: in the imatinib cohort, low risk (LR) patients are 51%, intermediate risk (IR) 35% and high risk (HR) 14%, whereas in the 2^ndG TKIs cohort, LR patients are 68%, IR 21% and HR 11%.

The 5-year survival for the overall population is 88% with 22 of 133 deaths (16.5%) due to CML-related causes (progression to blast crisis). Due to the difference in median age, the 5-year survival in the imatinib arm is 82%, with 9 of 98 total deaths (9%) due to CML, while the 5-year survival in the 2^ndG TKI arm is 94%, with 13 of 35 (37%) deaths due to CML. All but 4 progressions and most CML-related deaths occur within the first 2 years of therapy. 11 of 22 (50%) CML-related deaths occur in the HR group (175 patients, 12% of the total), compared to 8 in the IR group (404 patients, 28%) and only 3 in the LR group (852 patients, 60%). Therefore, regardless of the treatment received, the risk of CML-related death is approximately 6% in the HR group, 1.98% in the IR group, and only 0.35% in the LR group. The ELTS risk score at diagnosis discriminates survival probabilities very well in both treatment cohorts: in the imatinib cohort, survival over 5 years was LR 92%, IR 74% and HR 66% and in the 2^ndG TKIs treated cohort survival was LR 97%, IR 91%and HR 82%. In the imatinib-treated cohort, molecular response is associated with a statistically significant inferior survival for those patients not achieving BCR::ABL1 ≤ 1% at 12 months (P value <0.0001) or ≤ 0.1% at 24 months (P value = 0.0002), but the occurrence of CML-related deaths is not significantly associated with molecular response. Although less marked, also in the 2^ndG TKI group, survival is significantly inferior in patients who do not achieve BCR::ABL1 ≤ 1% at 12 and 24 months, but again no significant association with CML-related deaths is found.

Conclusions

Our real-world data, prospectively collected in a large number of CML patients treated with first-line imatinib or 2^ndG TKIs, show that 1) only a minority (less than one-fifth) of the deaths observed over a 5-year period are CML-related; 2) half of the CML-related deaths occur in high-risk ELTS patients, who represent only 12% of all CML cases in the Italian population, while the risk of dying from CML-related causes is marginal (0.35%) in low-risk ELTS patients, who represent approximately two-thirds of all CML patients; 3) 2^ndG TKIs are not able to prevent the occurrence of progression and subsequent CML-related deaths as more CML-related deaths were observed in patients treated with 2^ndG TKIs; 4) the ELTS scoring system is highly predictive of the probability of survival in CML patients, regardless of the first-line treatment chosen; 5) Failure to achieve molecular milestones represented by BCR::ABL1 ≤1% at 12 months or ≤0.1% at 24 months, is associated with statistically significant worse survival, although not due to CML-related deaths. The reasons for this association are unclear at this time and are the subject of further investigation.