Treatment Response Patterns and Related Prognostic Indicators in Chronic Myeloid Leukemia (CML) Patients: A Multi-Country Medical Record Review Study

Abstract

Abstract 3131 Objective:To assess treatment responses and prognostic indicators of response among patients with chronic myeloid leukemia (CML) in multiple countries. Methods:Physicians in United States (N=60), United Kingdom (N=45), Germany (N=49), and Japan (N=60) were recruited to abstract data retrospectively from medical records of patients diagnosed with CML between 1/1/2005 and 12/31/2009. Patients selected were age ≥18 years and in chronic phase at the time of diagnosis and were Philadelphia chromosome and/or BCR-ABL positive. Patients were also required to have received 1st-line treatment with imatinib and not be enrolled in a randomized clinical trial during the study period. We included a subgroup who had also received 2nd-line therapy with nilotinib or dasatinib. We assessed rates of complete hematological response (CHR) at 3 months (mos), complete cytogenetic response (CCyR) at 12 mos, and complete or major molecular response (MMR) at 18 mos, stratified by line of therapy. We used multivariable logistic regression models to evaluate the independent prognostic effects of age, gender, Sokal score, spleen size, race, Charlson comorbidity index score (CCIS), dose, and time to initiation of therapy. For patients on 1st line therapy, only those in chronic phase at the time of imatinib initiation and with non-missing data on the variables of interest (n=751) were included in the regression analyses. Results:Data were abstracted from medical records of 1, 063 patients in US (N=300), Japan (N=300), Germany (N=243), and in the UK (N=220). The average patient age was 55 years and a majority were male (60%). Response rates by line of therapy are listed in Table 1.Table 1:Response Rates Stratified by line of TherapyLine of TherapyTreatment (N)CHR at 3 months n (%)CCyR at 12 months n (%)MMR at 18 months n (%)1st LineImatinib (N=1,063)563 (53%)542 (53.1%)257 (25.1%)2nd LineDasatinib (N=148)72 (48.7%)44 (31.7%)25 (17.7%)Nilotinib (N=113)48 (42.3%)25 (22.9%)21 (18.9%)The proportion of patients achieving 1st-line response was similar across countries except that a greater percentage of patients from the UK obtained CHR at 3 mos (70.5%) compared to other countries (range: 46% [US] to [50.6% [Germany]) and MMR at 12 mos (55%) compared to the other countries (range: 26% [Japan] to 40% [Germany]). In 1st-line therapy, age ≤45 years (odds ratio [OR]=1.57, 95% CI=1.05–2.33 reference [ref] age 45–65) and low Sokal score (OR=1.64, 95% CI=1.16–2.33; ref intermediate) were associated with a greater likelihood of CHR response at 3 mos while Black race (OR=0.40, 95% CI=0.16–0.98; ref White race) and a CCIS score of 2 or higher (OR=0.48, 95% CI=0.25–0.75; ref: CCIS 0) were associated with a lower likelihood of CHR response at 3 mos. Age ≤45 years (OR=1.49, 95% CI=1.02–2.17) and Black race (OR=0.33, 95% CI=0.14–0.77) had similar associations with CCyR at 12 mos as for CHR at 3 mos. We did not find any significant prognostic indicators of MMR at 18 mos.In 2nd line therapy, choice of dasatinib as the 2nd-line drug (OR=2.77, 95% CI=1.19–6.43; ref: nilotinib), CHR to 1st-line therapy (OR=5.97, 95% CI=1.72–20.71; ref: no CHR), and low Sokal score (OR=7.93, 95% CI=2.93–21.51) were indicative of higher likelihood of CHR. Treatment with dasatinib (OR=7.43, 95% CI=2.36–23.40) and initial dose > 100mg for dasatinib or 600mg for nilotinib (OR=3.25, 95% CI=1.11–9.51; ref: dose ≤ 100mg for dasatinib or ≤ 600mg for nilotinib) predicted a greater likelihood of CCyR at 12 mos. We found no variables prognostic of MMR at 18 mos. Conclusions:We found several factors in addition to Sokal score to be prognostic for clinical response in 1st-line therapy, including younger age, White race, and no comorbidity. In addition, in 2nd-line therapy, the likelihood of response was higher in patients treated with dasatinib than in those who received nilotinib. To our knowledge, this difference has not been reported previously and could reflect a real difference in outcomes. However, it may also be due to residual confounding by factors that we could not control in a retrospective observational study. Disclosures:Trask:Pfizer: Employment. Mitra:Pfizer: Consultancy. Iyer:Pfizer: Employment. Candrilli:Pfizer: Consultancy. Kaye:Pfizer: Consultancy.