4425 Adherence to imatinib therapy has proven to be a major determinant of treatment results, but the degree of impact and the determinants of nonadherence are still contradictory. There is no information regarding adherence to imatinib therapy in the Brazilian public health system. The aims of this study were to identify the characteristics related to treatment interruptions and nonadherence and to examine how these interruptions affect treatment responses and survival. Materials and Methods We conducted a retrospective study in a cohort of patients (pts) with CP-CML enrolled in 14 Hematology centers in South Brazil. All pts received imatinib 400mg as first or second-line therapy. Early-imatinib treatment was considered when imatinib started before 12 months (mo) from diagnosis. Patient evaluation and response criteria followed the ELN recommendations. The ACE-27 (Adult Comorbidity Evaluation-27) is a 27 item comorbidity index for patients with cancer and assign weights from 1 to 3 based on the dysfunction grade of each condition (mild, moderate and severe, respectively). An ACE-27 score was applied to each patient. Imatinib suspensions were considered if superior to 20 days at any point during therapy. Two levels of analysis were performed: all kinds of interruptions (nonadherence and toxicity) and only nonadherence ones. Information for nonadherence was taken from medical and pharmacy registers (pt self-report, missing scheduled appointments and pill counts). Results We analyzed data from 185 pts with CP-CML diagnosed since 1990. The median age at diagnosis was 48 yr (4 – 85) and 55% were male. The median time from diagnosis to imatinib was 7 mo (0 – 178) and 71% pts were early-imatinib treated. Prior therapy with interferon was used in 70% pts. The median of follow-up was 47 mo. Treatment interruption was observed in 63/185 patients (34%) and was related to toxicities in 35/63 pts (55%) and to nonadherence in 28/63 pts (45%). The adherence rate was 85%. In a multivariate analysis, only late-onset imatinib treatment (Odds Ratio [OR]=36,05; p<0,001) and severe comorbidity (OR=27,05; p=0,03) were associated with higher risk of interrupting imatinib for any reason. The only variable associated with nonadherence was late-onset imatinib treatment (OR=14,76; p<0,001). Although not statistical significant, male and comorbidity showed a tendency to be linked with nonadherence (Table 1). Nonadherent pts, compared with adherent ones, had lower complete cytogenetic response (CCyR) rates at 12 mo (39% and 65%; p=0,004, respectively; Figure 1) and lower major molecular response (MMR) rates at 18 mo (9,5% and 35%; p=0,002, respectively; Figure 2). Finally, treatment interruption had a relevant negative impact on EFS in 4 yr. In the group that had treatment interruption, EFS was 52%, compared with 78,5% in the group without interruptions (p=0,002; Figure 3). Analysis performed only on nonadherence treatment interruption groups showed no significant difference (53% and 71%; p=0,15, respectively). Conclusions In this cohort, a substantial proportion of pts failed to take imatinib properly, decreasing the chances of disease control. The late onset of imatinib therapy correlates with lower adherence, so front-line imatinib therapy should be started as soon as possible. Special attention should be given to pts with severe comorbidities, as they are more prone to suffer side effects or to lack adherence. Finally, pts who interrupted treatment had lower CCyR, MMR and EFS. Clinical and patient characteristics related to nonadherence Disclosures: No relevant conflicts of interest to declare.

4487 Background: Imatinib treatment for Chronic Myeloid Leukemia (CML) was first introduced in Brazil in 2001, initially used as second line therapy for patients resistant or intolerant to interferon (IFN). In 2008 imatinib was adopted as front-line therapy for chronic-phase (CP-CML) and clinical experience is improving since then, but little is known about the result of its introduction in our clinical practice. Aims: To evaluate the impact of imatinib treatment in the outcomes of a cohort of CP-CML and the prognostic significance of Sokal and Hasford scores and late-onset treatment. Methods: We conducted a retrospective study in a cohort of patients with CP-CML from a south Brazilian database. All patients received imatinib 400mg as first or second-line therapy. Patient evaluation and response criteria followed the European LeukemiaNet. The outcomes were response to treatment, event-free survival (EFS) and overall survival (OS). Results: We analyzed data from 185 pts with CP-CML diagnosed since 1990. The median age at diagnosis was 48 years (4 – 85). The median time from diagnosis to imatinib was 7 months (0 – 178) and 29% of pts had more than 12 months lapse. Prior therapy with IFN was used in 70% pts. All pts had a minimum follow-up of 12 months. At baseline, 57 pts (31%) were in complete hematological response (CHR) due to the use of previous treatment. Of the 127 pts not in CHR at baseline, 98% achieved CHR early during imatinib treatment. 177 pts had cytogenetic evaluation during treatment and 9 pts were in complete cytogenetic response (CCyR) due to the use of previous IFN. Of the 168 pts not in CCyR at baseline, 86,4% achieved a major cytogenetic response (MCyR) during imatinib treatment (84% had a CCyR and 2,4% had a partial cytogenetic response). The rate of pts achieving MCyR any point during treatment differed significantly in the low, intermediate and high risk Sokal score groups (97%, 81% and 78% respectively, P=0,04), but not in the Hasford score groups (90%, 85% and 72%, P=0,22). Minor cytogenetic response was seen in 3,6% of pts, minimal cytogenetic response in 6% and 9,5% had no cytogenetic response. The median time to a MCyR was 9 months, with 62% of pts achieving MCyR at 12 months. The rate of pts achieving MCyR in 12 months differed significantly between pts who start imatinib before 12 months from diagnosis (68%) and those late treated (47%, P=0,02). Evaluation of minimal residual disease at the molecular level was available for 155 pts: 25,5% of pts had a complete molecular response (CMR), 43% had a major molecular response (MMR) and 2 pts were in MMR at baseline due to previous IFN. The projected EFS and OS rates at 4 years were, respectively, 68% and 92% after a median follow-up time of 4 years. The rate of EFS differed significantly in the low, intermediate and high risk Sokal score groups (80%, 66% and 52% respectively, P=0,04), but not in the Hasford score groups (78%, 62% and 44%, P=0,09). During treatment with imatinib, 120 pts (65%) had a register of any grade hematologic adverse event (21% being grades 3 or 4) and 165 pts (90%) had a register of any grade nonhematologic adverse event (9,3% being grades 3 or 4). Of the 185 pts who received treatment, 134 (72%) continue to receive imatinib and 51 (28%) discontinued treatment. The reasons for discontinuation were: 11 (6%) pts had drug-related adverse events (3 [1,6%] hematologic and 8 [4,3%] nonhematologic), 17 (9,2%) had disease progression (5 [2,7%] loss of CHR, 10 [5,4%] loss of CCyR, 2 [1%] had progression to accelerate or blastic phase), 22 (11,9%) had treatment failure (3 [1,6%] had no CHR, 13 [7%] had no CCyR and 6 [3,2%] had no MMR), 1 pt (0,5%) discontinued due to comorbidity. For the 51 (100%) pts that discontinued imatinib, 31 (61%) switched to dasatinib, 17 (33%) to nilotinib, 1 (2%) to hydroxyurea, 1 (2%) to other treatment and 1 (2%) remained without treatment. Sixteen pts (8,6%) died during imatinib treatment or during long-term follow-up after discontinuation of imatinib. Conclusions: In our population of CP-CML pts treated with imatinib, a majority of patients achieved complete cytogenetic and major molecular responses, with a prolonged of OS and EFS and good safety profile. Sokal score showed better prognosis prediction than Hasford. Early onset of imatinib therapy led to better outcomes and justifies imatinib as front-line treatment of our patients. Disclosures: No relevant conflicts of interest to declare.

Background Treatment of chronic myeloid leukemia with imatinib leads to disease remission in a majority of patients (pts), but in a subgroup of pts controlling the disease remains a challenge. One of the proposed prognostic factors for identifying this subset of pts is the treatment response in the first months of imatinib therapy. The definitions of response, warning signs and failure to treatment, with the identification of pts at greater risk of progression or failure, were revised and published in 2013 by the European Leukemia Net group. Aims To apply the new recomendations of ELN2013 in order to verify the evolution of the patients according to the response criteria at 12 months (mo) of imatinib therapy. Methods Retrospective study in a cohort of patients with CP-CML from a southern Brazilian database. All patients received imatinib 400mg as first or second-line therapy. Patient evaluation and response criteria followed the ELN2013 criteria and were classified in 3 groups according to the cytogenetic and molecular responses at 12 mo: optimal, warning and failure. An event was defined as any of the following while on therapy: death from any cause, loss of complete hematologic response, loss of complete cytogenetic response, discontinuation of therapy for toxicity or lack of efficacy, or progression to accelerated phase or blastic phase. Results Data from 98 pts with a median follow-up (FU) time of 40 mo was analysed. At 12 mo of imatinib therapy, 35 pts (36%) had optimal response according to the ELN2013 criteria. 26 (26,5%) were classified as warning and 37 (37,5%) as failure. Within the pts with optimal response, only 4 (11%) have had any event, with a median time-to-event (TTE) of 36 mo; in the warning group the event rate and TTE were 3 (12%) and 28 mo, respectively. There were no significant differences in the results between these two groups. On the other hand, within the pts with failure, 24 (65%) presented an event during FU with a TTE of 19 mo. This difference was significant compared to the two previous groups. Interestingly, within the warning group, 12 pts (46%) developed major molecular response (MMR) at any point during FU, while only 12 (32%) pts achieved RMM in the failure group. Conclusion In this cohort of CP-CML pts is established the diference in the event rate between the pts who achieve optimal or warning responses at 12 mo of imatinib treatment when compared with those with failure criteria, according to the new ELN2013 recomendations. Despite having some late responders, atention must be paid in pts in the failure group as they are in greater risk of progression or failure. As demonstrated in this analysis, the ELN2013 criteria can be applied in this population in the clinical practice. Disclosures: No relevant conflicts of interest to declare.

Sustained deep molecular response (MR4.5) after imatinib treatment defines a subgroup of patients with chronic myeloid leukemia (CML) with better outcome and that probably would be able to stop treatment in the future, according to results of clinical TKI discontinuation trials. Most of these trials showed that patients with a long-term imatinib treatment and low Sokal risk have a higher probability of maintain a deep molecular remission after stopping treatment. OBJECTIVES The main objective is to review the molecular responses, overall survival and event free survival of CP CML patients that have been treated with imatinib in 14 hematology centers in South Brazil. Using our data basis we also would like to see how many of them present long-term imatinib treatment, sustained deep molecular remission and correlate these findings with the Sokal risk groups. These data would allow us to predict patient profile that could be able to discontinue the treatment in the future in a prospective clinical trial. PATIENTS AND METHODS This is a retrospective study in a cohort of pts with chronic myeloid leukemia chronic phase (CP) that have been treated in 14 hematology centers in South Brazil. All pts received imatinib 400mg as first or second-line therapy. Patient evaluation and response criteria followed the ELN recommendations. MR(4.5) was defined as ≥ 4.5 log reduction of BCR-ABL on the international scale (IS) and determined by reverse transcriptase polymerase chain reaction. All tests were performed at a central standardized according to ELN. Event-free-survival (EFS) was measured from the start of imatinib to the date of any of the following events while on therapy: death from any cause, loss of complete hematologic response, loss of complete cytogenetic response, discontinuation of therapy for toxicity or lack of efficacy, or progression to accelerated phase or blastic phase. Overall survival (OS) was measured from the start of imatinib until death of any cause or to the date patient was last known to be alive. RESULTS Data from 474 patients was analyzed. After a median observation time of 46 months, 5-year overall survival (OS) was 86%, 5-year event-free-survival was 53%. Of the 474 patients, 258 had adequate PCR evaluations during treatment. 118 of 258 (45,7%) patients achieved MR(4.5) and 69 of 258 (27%) had sustained response for at least two years after a minimum time of treatment of 3 years. The cumulative incidence of MR(4.5) after 9 years was 76% (median, 3 years). In the group that achieved MR(4.5), there was only 1 (0,8%) death and 1 (1,1%) progression compared to 8 deaths (5,7%) and 8 progressions (7,5%) in the group without MR(4.5); these differences were significant with p=0,03 and 0,02 respectively. In the subgroup of 69 patients that had had been treated with imatinib for 3 year or more and sustained deep response (RM4,5) for at least two years, 21 pts had low Sokal risk, 7 pts intermediate Sokal risk and only 4 pts a high Sokal risk. Unfortunately, in 37 pts the Sokal risk could not be accessed due to missing information. CONCLUSION In our series MR(4.5) is reached in the majority of patients with long-term imatinib treatment. MR(4.5) is a predictor of outcome with only one disease progression and one death due to CML in this group of patients. Unfortunatly PCR are not available for all patients in our clinical practice, but this situation are improving. Regarding the 69 patients with TKI discontinuation trial criteria, we find out that 21 patients fulfill such criteria. In the future, according to the results of current stop trials it could be possible include this selected group of CP CML patients in a prospective clinical trial. Disclosures No relevant conflicts of interest to declare.

4423 Background Treatment of chronic myeloid leukemia with imatinib leads to disease remission in a majority of patient, but in some patients (pts) controlling the disease remains a challenge. One of the proposed prognostic factors for identifying this subset of pts is the treatment response in the first months of therapy. Objectives We conducted a study to evaluate the importance of the early complete cytogenetic response (CCyR) and the factors associated with its achievement. Methods This is a retrospective study in a cohort of pts with chronic-phase chronic myeloid leukemia (CP-CML) enrolled in 14 Hematology centers in South Brazil. All pts received imatinib 400mg as first or second-line therapy. Early-imatinib treatment was considered when imatinib started before 12 months (mo) from diagnosis. Patient evaluation and response criteria followed the ELN recommendations. The ACE-27 (Adult Comorbidity Evaluation-27) is a 27 item comorbidity index for patients with cancer and assign weights from 1 to 3 based on the dysfunction grade of each condition (mild, moderate and severe, respectively). An ACE-27 score was applied to each patient. Imatinib suspensions were considered if superior to 20 days at any point during therapy. EFS was measured from the start of imatinib to the date of any of the following events while on therapy: death from any cause, loss of complete hematologic response, loss of complete cytogenetic response, discontinuation of therapy for toxicity or lack of efficacy, or progression to accelerated phase or blastic phase. Results We analyzed data from 450 pts with CP-CML diagnosed since 1990. The median age at diagnosis was 48 yr (4 – 85) and 55% were male. The median time from diagnosis to imatinib was 7 mo (0 – 178) and 71% pts were early-imatinib treated. Prior therapy with interferon was used in 60% pts. The median of follow-up was 47 mo. With 6 months of imatinib therapy, 198 pts (44%) achieved CCyR. In this group, the four year cumulative incidence of events was 33 (17%) and the EFS was 75,5%. 252 (56%) were not in CCyR at 6 months of therapy. In this group, a greater proportion of cumulative of events was observed: 86 (34%), and the EFS was 62,3%. This difference was significant (P=0,03; Figure 1). In this group of pts, 63% achieved CCyR after 6 months any time during follow up and the median time for CCyR in these pts was 17 months. The chance of achieving major molecular response (MMR) during follow up was 79% for the pts with CCyR at 6 months compared to 53% for the group with no CCyR at 6 months (P<0,001). Some factors were associated with reduced chance of CCyR at 6 months. In a multivariate analysis, the pts with late-onset imatinib treatment (more than 12 mo from diagnosis) had a CCyR rate of 31%, in contrast, the pts who started imatinib before 12 mo had a rate of 50% (P=0,02). The pts with good adherence to treatment had greater CCyR rate than those with poor adherence (interruption greater than 20 days), 51,4% and 35%, respectively (P=0,04). Comorbidity measured by ACE-27 score also influenced the CCyR rates at 6 months: 54% of score 0 (no comorbidity) patients achieved CCyR, compared to 30% of pts with score 1 (mild comorbidity), 33% of pts with score 2 (moderate) and 47% of pts with score 3 (severe) (P=0,009). The greater CCyR rate in the severe comorbidity group probably lacks significance due to the reduced number of pts in this group (22). Conclusions A great proportion of pts achieve CCyR after 6 months of imatinib therapy, nevertheless, the pts who achieve CCyR by 6 months of therapy have greater proportion of major molecular response and event-free survival. Imatinib therapy should be started as soon as possible and additional efforts must be taken to avoid nonadherence. Finally, special attention should be given to pts with comorbidities as their results tend to be worse. Disclosures: No relevant conflicts of interest to declare.