2 MATERIALS
From 2019 to 2020, the patients with histologically confirmed GIST in Sichuan University Gastrointestinal Surgery Center in Chengdu, China, actively participated in this study. Inclusion criteria are as follows: (1) ≥18year-olds with hepatic and renal functions, adequate hematological, histologically proven GIST; (2) Eastern Cooperative Oncology Group performance status (ECOG PS) ≤2; (3) The duration of imatinib adjuvant therapy at least 29days. The study protocol was approved by the medical ethics review boards and performed following the Declaration of Helsinki. All patients provided written informed consent before registration. This trial was registered with the Chinese Clinical Trial Registry and the registration number was ChiCTR1900020854 (http://www.chictr.org.cn.). Patients underwent follow-up with enhanced computed tomography or magnetic resonance imaging scans every 6 months for the first 2 years, every 1 year for the next 5 years in the process of imatinib adjuvant therapy. Hematological and non-hematological adverse events were documented and graded in light of CTCAE (Common Terminology Criteria for Adverse Events Version 5.0) v5.0. Moreover, patient medical records were reviewed retrospectively.22
Clinical data of patients with GIST, who had been treated with various doses of imatinib (200–600 mg/day) for at least 29 days were collected between December 2019 and November 2020. Within 24±2 h after the last dosage of imatinib, blood samples (at least 3 mL) were collected into heparinized tubes, centrifuged at 3,000 rpm for 10 min at room temperature, and stored at −80°C until analysis. Cminwas measured by a validated liquid chromatography-tandem mass spectrometry (LC-MS/MS) method, specific methods have been previously reported.23 All samples were detected in the Laboratory of Clinical Pharmacology, West China Hospital of Sichuan University (Chengdu, China). In addition, to evaluate the correlations between imatinib Cmin and different dosages, the patients were divided into four groups according to the imatinib Cmin quartiles. The lower quartile (Q1) included the 25% of patients with the lowest imatinib Cmin; Q2 and Q3 indicated the 25 % below and above the median concentration, respectively; Q4 included the 25% of patients with the highest imatinib Cmin.
Statistical analysis
Pearson χ2 or Fisher’s exact test was used for the comparison of categorical variables, while the Wilcoxon rank-sum test was used for ordinal and continuous variables. Recurrence-free survival (RFS) was the primary endpoint of this study, which was defined as the time from the date of the first dose of imatinib to the date of the first documented tumor recurrence assessed by radiographic imaging or mortality resulting from any cause. Patients who were alive and without the disease-specific recurrence were censored at the last follow-up.
We calculated the inverse probability of treatment weighting (IPTW) method using the propensity scores for eliminating the effects of confounding and selection bias by creating a pseudo-sample with the method as reported previously. 24,25 To estimate the IP weights, we calculated each patient’s probability of imatinib dosage adjustment via application of multivariate logistic regression model, with dosage adjustment of imatinib as the outcome and BSA, age, gender, risk stratification, mutation status, primary tumor site, and imatinib trough concentration (Cmin) as covariates. The IP weights were then calculated based on each patient’s estimated probability of whether dose reduction of imatinib, given their actual regimen of imatinib therapy.
To study the effect of dosage of imatinib adjustment on RFS of resected GISTs, weighted Cox proportional hazard regression models were estimated. Hazard ratios (HRs) along with their 95% confidence intervals (CI) were reported. RFS was calculated by the Kaplan–Meier method and compared by log-rank test. Cox proportional hazard model was used to assess the prognostic factors for the IPTW-matched patients. All statistical analyses were performed by using RStudio (https://www.rstudio.com/) in the text of R, version 4.1.2 (R Core Team 2021, Vienna, Austria) (https://www.r-project.org/). All tests were two-sided, P-values less than .05 were considered statistically significant.