Real-World Data from Long Term Follow-up in Chronic Myeloid Leukemia (CML) Patients (PTS) Treated with Imatinib. Predictive Factors for Achievement of Deep Molecular Responses and for Failure

Abstract

BackgroundResponse to TKI is the most important predictor of survival in CML and its early acquisition predicts the achievement of deep molecular responses (MR) (Baccarani et al. Blood 2013). In current ELN recommendations, the optimal time point to achieve Major Molecular Response (MMR) is defined at 12 months (mo) after diagnosis of CML however, not achieving MMR is not a failure criterion at any time point (Saussele et al, Blood 2014). There are controversies about when CML therapy must be changed in pts who do not achieve MMR after 12 mo of treatment. We wonder if pts achieving late MMR are less likely to achieve deep MR or are at risk of progression to accelerated phase (AP) or blast crisis (BC).ObjectivesWe aim to identify predictive factors for achievement of deep MR and improvement of failure-free survival (FFS). Other objective was to determine if delay in MMR acquisition predicts outcomes in order to justify an optimal landmark for changing TKI reducing the risk of failure.MethodsBetween 2000 and 2015, 122 consecutive adult pts with CP CML were treated with various TKI modalities. Adherence (ADH) monitoring occurred in every visit and was reported in medical records, optimal ADH defined as >90% of dose completed. Cytogenetic and MR were defined according to ELN definitions and followed the procedures as described elsewhere (Cross et al. Leukemia 2012). In particular, BCR-ABL<1% corresponds to complete cytogenetic remission (CCyR) (Lauseker et al. 2014). MMR: BCR-ABL transcript <0.1% IS. Confirmed MR4.0: BCR-ABL detection ≤0.01% IS in two consecutive analyses. Failure was defined as loss of CCyR, AP or BC. Different models were performed for CCyR at 6mo, MMR at 12mo and MMR at 24mo as predictors of MR 4.0; in all cases age and Sokal were used as covariates. Failure-risk was evaluated with Cox regression including in the model: Time to MMR, ADH and Sokal as covariates.ResultsFrom 122 pts diagnosed with CP CML, 110 pts received 1st line imatinib (IM) 400mg daily and constitute the basis of the analysis. Median age was 48 years (14-82), 65% were males, 82% low Sokal score. With a median follow-up of 107 mo (IQR 69.5-141) 82% pts are still on initial IM, 4.5% died and 1.8% were lost to follow-up. A total of 20 pts (18%) discontinued Imatinib, 9% due to intolerance and 9% due to failure. At 15 years, FFS was 88% and OS 95%.The proportion of pts treated with IM who achieved CCyR at 6mo and MMR at 12mo was 84% and 38% respectively. Median-time to achieve MMR and MR4.0 was 2.6 and 3.6 years respectively. By logistic regression analysis CCyR at 6mo OR 5,6 (95% CI 1,6-19,00), MMR at 12mo OR 5,3 (95% CI 1,4-21,0), ADH OR 7,0 (95% CI 1,3-37,0) and not MMR at 24mo OR 2,5 (95% CI 0,9-7,2) resulted independent predictors for achieving MR4.0 and confirm earlier observations. In Cox model of proportional hazards for risk of failure, time to achieve MMR HR 1,02 (95% CI 1,01-1,04), ADH HR 6.1 (95% CI 1,38-26,8), and Sokal HR 4,7 (95% CI 1,12-19,3) were independent predictors of failure. Time to achieve MMR is a continuous variable, so the HR associated with it, is equivalent to a 2% monthly increase in the risk of failure. To calculate HR in other time intervals, the product of the coefficient model (B) and time of interest (in mo) were used as an exponent of the base of the natural logarithm (e). Not reaching MMR at 12 mo correlates with higher risk of failure: hazard-risk (HR) 1.30, (95% CI 1.28-1.30) and the risk increases with every passing year not achieving MMR (Table 1). Table 1Time to MMR as predictor of progressionNo MMR achievement in different time intervalsHRCI 95%After 12 mo of TKI1,30(1,3-1,3)After 24 mo of TKI1,70(1,7-1,7)After 30 mo of TKI2,21(2,2-2,2)After 60 mo of TKI3,74(3,7-3,8)ConclusionsThe prognostic value of CCyR at 6mo and MMR at 12mo was confirmed. The association between early MMR and deep MR was confirmed. In this long-term follow-up Latin American real-world cohort of patients ADH was also an early predictor for achievement of deep MR. Although good responses are obtained with IM, some patients still progress, this is the reason why it is necessary to identify those who will relapse so as to consider a change in treatment to another TKI at a convenient time point. Time to achieve MMR, ADH, and Sokal Risk score were independent predictors of failure. Not achievement of MMR at 12mo correlates with increased risk of failure. The risk increases annually in proportion to the time delay. Patients who are delayed in reaching MMR should be controlled closely. DisclosuresPavlovsky:Novartis: Honoraria, Speakers Bureau; Janssen: Honoraria, Speakers Bureau. Pavlovsky:Novartis - Bristol: Speakers Bureau.