만성골수성백혈병 환자에서 이매티닙의 노출-반응 관계 분석

Abstract

Imatinib is still the most commonly used for first line therapy of chronic myeloid leukemia (CML), although new generation tyrosine kinase inhibitors (TKIs) have been developed. At the approved dose of imatinib 400 mg/day for the treatment of chronic phase CML (CP-CML), imatinib can elicit adverse drug reactions (ADRs), which leads to low compliance or dose discontinuation. In advance to improve long-term treatment outcomes, it is necessary to prevent ADRs of imatinib. The purpose of this study was to assess the imatinib exposure–safety response relationship and to propose a weight-based dosage regimen of imatinib in patients with CP-CML.

The 48-week data were collected from a control arm of a multinational, randomized phase 3 involving the patients newly diagnosed with CP-CML who received imatinib at starting dose of 400 mg/day. Imatinib exposure was estimated as the initial dose (400 mg) adjusted for patients body weight (BW). A major safety parameter was the occurrence of dose-limiting toxicity (DLT) whose definition was the incidence of grade 3+ ADRs according to the NCI-CTCAE ver 3.0. A logistic regression analysis was used to determine the imatinib exposure–safety response relationships and the covariates affecting the safety response. The receiver operating characteristic (ROC) curve of the Dose/BW for DLT occurrence was generated. The ROC curve was used to determine the cut-off distinguishing the low and high DLT subgroups. Kaplan–Meier (K–M) analysis was performed to compare the DLT occurrence between the low and high DLT subgroups by using the log-rank test. The rates of DLT among the groups of patients divided into quintiles of Dose/BW were compared by using the chi-squared test. A p-value less than 0.05 was considered statistically significant.

A total of 81 Asian patients were enrolled from 2011 to 2014. There were 39 patients with DLT by 48 weeks of imatinib treatment, 34 patients without DLT for 48 weeks. Eight patients who had a dose interruption not related imatinib were excluded from the analysis. The logistic regression analysis demonstrated a significant positive relationship between the Dose/BW and the occurrence of DLT (p = 0.003). The AUC on the ROC curve was 0.703, indicating fair predictability of DLT. The distinctive cut-off of Dose/BW that separates the high and low risk of DLT was 7 mg/kg/day. In the K–M curve, there was a significant difference between the two subgroups of low- (40.3%) and high-risk (65.8%) DLT (log-rank test, p = 0.027). The covariate related to the DLT was the absolute neutrophil count (ANC) at baseline. The probability of DLT occurrence (P) in the multivariate logistic regression model (p = 0.001) was expressed as:

Logit P = ln (P/1-P) = 0.509×(Dose/BW) – 0.023×(ANC) - 2.976

The rates of DLT from the first to the fifth quintiles of Dose/BW were 33 (Q1), 38 (Q2), 47 (Q3), 63 (Q4) and 73% (Q5), respectively. The first quintile group (mean Dose/BW = 5.02 mg/kg) of Dose/BW had the lowest DLT rate (Chi2 test, p = 0.006). Of the 67 patients with molecular response data, 24 patients (35.8%) had achieved MMR by 48 weeks. As a result of logistic regression analysis, there was no significant relationship between Dose/BW and MMR achievement (p = 0.282).

The imatinib Dose/BW–DLT relationship demonstrated that the CP-CML patients with a lower BW had a high likelihood of DLT occurrence (p = 0.003). This study suggests that patients weighing less than 60 kg start the treatment for CP-CML with imatinib 300 mg/day, which would improve compliance as well as the long-term therapeutic outcomes of imatinib.