Endometrial cancer is the most common gynecologic cancer in the United States. However, no early detection test exists for asymptomatic women at average risk for endometrial cancer.
We sought to identify early detection biomarkers for endometrial cancer using prediagnostic serum.
We performed a nested case-control study of postmenopausal women in the Prostate, Lung, Colorectal, and Ovarian(PLCO) Cancer Screening Trial (n=78,216), including 112 incident endometrial cancer cases and 112 controls. Prediagnostic serum was immunodepleted of high-abundance proteins and digested with sequencing grade porcine trypsin via pressure cycling technology. Quantitative proteomics and phosphoproteomics was performed using high-resolution liquid chromatography-tandem mass spectrometry and highly multiplexed isobaric mass-tag combined with basic reversed-phase liquid chromatography. A set of proteins able to predict cancer status was identified with an integrated score assessed by receiver operator curve analysis.
Mean time from blood draw to endometrial cancer diagnosis was 3.5 years (standard deviation 1.9 years). There were 47 differentially abundant proteins between cases and controls (P<0.05). Protein alterations with high predictive potential were selected by regression analysis and compiled into an aggregate score to determine the ability to predict endometrial cancer. An integrated risk score of six proteins was directly related to disease incidence in cases with blood draw ≤2 years, >2 years to ≤5 years or >5 years prior to cancer diagnosis. The integrated score distinguished cases from controls with an area under the curve of 0.80 [95% Confidence Interval 0.72-0.88].
An integrated score of six proteins using prediagnostic serum from the PLCO Cancer Screening Trial distinguishes postmenopausal endometrial cancer cases from controls. Validation is needed to evaluate whether this test can improve prediction or detection of endometrial cancer among postmenopausal women.