Risk assessment based on the Risk Index for Serious Adverse Events (SAEs) in phase I trials.

Abstract

e14032 Background: Although patients (pts) in phase I trials may have expectations about their chance to benefit, they sometimes experience serious adverse events (SAEs). These events impact on their quality of life negatively as well as their clinical management and survival, if once occurred, regardless of timing of occurrence or relation to study drugs. Therefore, paying close attention for pts who are more likely to experience SAEs may be required; however, little is known about which clinical features are related to SAEs in phase I trials. The aim of this study is to construct a risk index for SAEs prediction. Methods: We collected medical records of 418 pts with solid tumor who participated in phase I trials for the first time and were treated at our hospital from January 2011 to December 2015. The time to first SAE (TTS), defined as the time from registration to first SAE, was analyzed. Stepwise Cox regression model was applied to screen out the factors to build a risk index. Internal validation was performed using a bootstrap procedure. Results: Four hundred and seven pts were analyzed for building a risk index. Median age was 63 years (range 22-81). Predominant primary tumor site was gastrointestinal cancer (65.1%), followed by lung cancer (8.4%) and genitourinary cancer (5.7%). A total of 251 pts (61.7%) were treated with small molecule compounds, while 136 pts (33.4%) were treated with antibody therapy including immune therapy. A total of 101 pts (24.8%) experienced SAEs included 18 events related to study drugs. Four independent risk factors (use of antibody therapy, 60 < age < 70, higher serum albumin concentration and lower serum alkaline phosphatase concentration), identified by stepwise Cox regression model, were combined into simple ternary risk index that stratified pts into low- (n = 93), intermediate- (n = 175) and high-risk group (n = 139). Hazard ratios for high-risk and intermediate-risk relative to low-risk group and their 95% confidence intervals were 5.25 (2.64-10.42; high-risk) and 2.00 (0.99-4.07; intermediate-risk). Conclusions: Risk assessment based on the risk index for SAEs may be useful for patient management in phase I trials. Further research is needed to validate this risk index.