Study
PLCO
(Learn more about this study)
Project ID
PLCO-966
Initial CDAS Request Approval
Apr 8, 2022
Title
Evaluating a multi-biomarker panel for the early detection of ovarian cancer
Summary
Epithelial ovarian cancer (EOC) is usually diagnosed in advanced stages when the cancer has metastasized leading to poor survival outcomes. Women diagnosed with localized ovarian cancer have a 92% 5-year survival rate whereas those diagnosed with advanced stage disease have only a 17% 5-year survival rate. This provides a strong rationale that a screening tool that detects ovarian cancer in its earlier stages would have promise to reduce mortality. Previously in the ovarian cancer arm of the PLCO, the combination of cancer antigen -125 (CA-125) and transvaginal ultrasound (TVUS) was not found to be efficacious in reducing mortality compared with the usual care arm. Additionally, the risk of ovarian malignancy algorithm (ROMA) which integrates CA-125 and human epididymis secretary protein 4 (HE4) with menopausal status to differentiate between low-and high-risk patients with EOC detects only a modest proportion of early stage EOCs. Thus, the need remains to develop an effective screening procedure with high sensitivity, specificity, and positive predictive value (PPV) that meets the stringent criteria needed for population-level screening. Exploring the use of serum-based multiple biomarker panels with high sensitivity and specificity is a potentially valuable line of inquiry to generate a better screening procedure. The Prostate, Lung, Colorectal, and Ovarian (PLCO) study provides a good data resource with multiplicity of biomarkers to explore different biomarker panels. The study will investigate different biomarkers with a particular focus on CA-125, HE4 and apolipoprotein – A1 (ApoA1) and assess how a multiplex assay of these biomarkers can improve early detection of EOCs. We hypothesize that a combination of the selected biomarkers could result in high enough sensitivity and specificity to be a viable screening tool for further evaluation to determine if it leads to effective detection of early stage EOCs.
Aims
The overall goal of this study is to use the subset of the ovarian cancer arm of the PLCO data whose serum was analyzed for 36 biomarkers (n=119 cases, n=952 controls) to develop a screening algorithm with translational importance for the early detection of EOCs. This will be achieved by studying the performance of combinations of candidate biomarkers that will be selected based on availability in the PLCO and evidence from previous studies showing high sensitivity and specificity. The specific aims are to: 1. Determine the association between the selected candidate biomarkers and ovarian cancer. This will be achieved through a nested case-control approach. Multivariable models adjusting for other potential confounders other than the matched variables will be fitted. The sensitivity and specificity of the candidate biomarkers will be assessed using the receiver operating characteristic (ROC) curve analyses and by calculating the area under the curve (AUC). 2. Develop an algorithm of a combination of biomarkers for detecting early EOCs. Different combinations of the biomarkers will be examined for their joint sensitivity and specificity. The ROC curve will be plotted, and the AUC calculated. Different and reasonable cutoff points for different algorithms will be simulated to determine an algorithm with translational importance. 3. Determine the lead time of EOCs diagnosis with this algorithm. Previously validated mathematical models for determining lead time will be used to assess the lead time of EOCs diagnosis.
Collaborators
Dr. Anthony Alberg, Department of Epidemiology and Biostatistics, Arnold School of Public Health University of South Carolina
