Optimal summary of information in longitudinal CA125 measurements for efficient ovarian cancer screening

Ovarian cancer mortality rates have changed little over 30 years despite th erapeutic advances. Given the large proportion of cases discovered in late stages and consequent poor prognosis, and the excellent prognosis if discov ered in early stages, screening is an appealing approach to investigate red ucing ovarian cancer mortality, Due to the large scale of screening program s end the low incidence of ovarian cancer, it is paramount that the screeni ng strategy is as efficient as possible. CA125 is a marker for ovarian canc er that shows promise in detecting most ovarian cancers preclinically. Howe ver, its efficiency could be substantially improved if longitudinal CA125 v alues were taken into account in the screening program, We provide a detail ed description, with minimal mathematical notation, of a systematic statist ical calculation to optimally summarize the information in longitudinal CA1 25 values. Based on this calculation we describe a strategy, called the ris k of ovarian cancer algorithm (ROCA), which implements an efficient approac h to screening for ovarian cancer based on longitudinal CA125 measurements.