Background: Biomarkers that aid in the differential diagnosis of malignant pelvic masses from benign ones prior to surgery are needed in order to triage women with malignant masses to appropriate specialist care. Because high albumin-adjusted serum calcium predicted ovarian cancer among women without evidence of disease, we hypothesized that it might predict cancer among women with pelvic masses that were evident radiographically.
Methods: We studied a cohort of 514 women with pelvic masses who underwent resection at Wake Forest University Baptist Medical Center from July 2009 through June 2013. We divided patients into a "training" set, to identify associations in the data, and a "testing" set, to confirm them. Data were obtained from medical records. A best fit model was selected using the Akaike Information Criterion.
Results: Albumin-adjusted serum calcium was significantly higher in women with malignant versus benign masses (P = 0.0004). High normocalcemia, i.e., an albumin-adjusted serum calcium ≥ 10 mg/dL, occurred in 53% of women with malignant tumors versus 12% of benign tumors. High normocalcemia was associated with an approximately 14-fold increased risk of malignancy. The best fit model (Overa) included albumin, calcium, and nonlinear terms. Overa achieved an area under the curve of 0.83 with a sensitivity of 72% and specificity of 83%, a positive predictive value of 71% and a negative predictive value of 85%.
Conclusions: A model using serum calcium and serum albumin to predict malignancy in women with pelvic masses has high sensitivity and is economical.
Impact: Our model can help triage women with ovarian cancer to appropriate surgical care.
©2015 American Association for Cancer Research.