Predictors of pretreatment CA125 at ovarian cancer diagnosis: a pooled analysis in the Ovarian Cancer Association Consortium

Ana Babic, Daniel W. Cramer, Linda E. Kelemen, Martin Köbel, Helen Steed, Penelope M. Webb, Sharon E. Johnatty, Anna deFazio, Diether Lambrechts, Marc T. Goodman, Florian Heitz, Keitaro Matsuo, Satoyo Hosono, Beth Y. Karlan, Allan Jensen, Susanne K. Kjær, Ellen L. Goode, Tanja Pejovic, Melissa Moffitt, Estrid HøgdallClaus Høgdall, Iain McNeish, Kathryn L. Terry

Research output: Contribution to journalArticlepeer-review

8 Scopus citations


Purpose: Cancer antigen 125 (CA125) is a glycoprotein expressed by epithelial cells of several normal tissue types and overexpressed by several epithelial cancers. Serum CA125 levels are mostly used as an aid in the diagnosis of ovarian cancer patients, to monitor response to treatment and detect cancer recurrence. Besides tumor characteristics, CA125 levels are also influenced by several epidemiologic factors, such as age, parity, and oral contraceptive use. Identifying factors that influence CA125 levels in ovarian cancer patients could aid in the interpretation of CA125 values for individuals. Methods: We evaluated predictors of pretreatment CA125 in 13 studies participating in the Ovarian Cancer Association Consortium. This analysis included a total of 5,091 women with invasive epithelial ovarian cancer with pretreatment CA125 measurements. We used probit scores to account for variability in CA125 between studies and linear regression to estimate the association between epidemiologic factors and tumor characteristics and pretreatment CA125 levels. Results: In age-adjusted models, older age, history of pregnancy, history of tubal ligation, family history of breast cancer, and family history of ovarian cancer were associated with higher CA125 levels while endometriosis was associated with lower CA125 levels. After adjusting for tumor-related characteristics (stage, histology, grade), body mass index (BMI) higher than 30 kg/m2 was associated with 10% (95% CI 2, 19%) higher CA125 levels, while race (non-white vs. white) was associated with 15% (95% CI 4, 27%) higher CA125 levels. Conclusion: Our results suggest that high BMI and race may influence CA125 levels independent of tumor characteristics. Validation is needed in studies that use a single assay for CA125 measurement and have a diverse study population.

Original languageEnglish (US)
Pages (from-to)459-468
Number of pages10
JournalCancer Causes and Control
Issue number5
StatePublished - May 1 2017


  • Biomarker
  • CA125
  • Ovarian cancer
  • Predictors
  • Prognosis

ASJC Scopus subject areas

  • Oncology
  • Cancer Research


Dive into the research topics of 'Predictors of pretreatment CA125 at ovarian cancer diagnosis: a pooled analysis in the Ovarian Cancer Association Consortium'. Together they form a unique fingerprint.

Cite this