Association of patient-reported outcomes and ovarian cancer recurrence.

Background: The vast majority of patients with advanced ovarian cancer experience disease recurrence after primary treatment.

Objective: To explore the diagnostic accuracy of repeated measurement of patient-reported outcomes and quality-of-life scores in relation to ovarian cancer recurrence.

Methods: Patients with ovarian cancer were recruited to the PROMova study by the end of their primary treatment at eight centers in Denmark. The purpose of the PROMova study was to explore the applicability of repeated use of patient-reported outcomes, which consisted of the European Organization for Research and Treatment of Cancer generic questionnaire and the ovarian specific questionnaire. The patient-reported outcomes were completed 3, 6, 9, 12, and 15 months after enrollment or until recurrence. The 3-month interval between completions was the period in which recurrence was assessed. Imaging and the biomarker CA125 were used as reference modality for recurrence. Mixed effects logistic regression was used to investigate the association between mean patient-reported outcome scores and recurrence. Receiver operating curves were used to establish cut-off scores. The diagnostic accuracy of patient-reported outcomes, including sensitivity, specificity, and positive and negative predictive values was estimated based on the Youden index. For combined scales, diagnostic accuracy was investigated based on multivariate analysis.

Results: The analysis included 196 patients with an overall recurrence rate of 50.5% and an overall mean time to recurrence of 302 days. With imaging as reference, patients with recurrence reported significantly lower global health, worse physical functioning, and more abdominal symptoms preceding recurrence. With CA125 as reference, global health, physical and emotional functioning were impaired. Despite the worsening of a number of symptoms prior to recurrence whichever reference modality was applied, the patient-reported outcome scores did not provide adequate diagnostic accuracy.

Conclusion: Repeated use of patient-reported outcomes during surveillance of ovarian cancer was not of diagnostic value. Future efforts should be directed at improving the administration of patient-reported outcomes as well as exploring the potential of using these outcomes as an indicator of clinical relevance.