1 INTRODUCTION In cancer research, the relation between repeated biomarker measurements and the risk of cancer-related events, such as the progression of the cancer to a stage that requires an intervention, is usually of great interest (Pepe et al., 2001; Putt et al., 2015; Van Hemelrijck et al., 2011). The repeatedly measured biomarkers typically have short-term fluctuations and are intrinsically associated with patients' status, in other words, endogenous (Rizopoulos, 2011).