|
Activity Number:
|
21
- Advances of Statistical Methodologies in Proteogenomic Research
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Sunday, August 7, 2022 : 2:00 PM to 3:50 PM
|
|
Sponsor:
|
Biometrics Section
|
|
Abstract #323647
|
|
|
Title:
|
Modeling Genomic and Proteomic Signals for the Early Detection of Cancer
|
|
Author(s):
|
Steven James Skates* and Yiling Liu and Wenqing Jiang and Bethan Powell and Scott Lenttz
|
|
Companies:
|
Massachusetts General Hospital and Massachusetts General Hospital and Boston University School of Public Health and Kaiser-Permanente Northern California and Kaiser-Permanente Southern California
|
|
Keywords:
|
longitudinal;
change-point;
Bayes factor;
early detection;
ovarian cancer
|
|
Abstract:
|
Early detection of ovarian cancer to reduce cancer specific mortality remains an elusive goal. Serum CA125 is the most widely used ovarian cancer biomarker. We developed a longitudinal CA125 algorithm and tested it in multiple prospective screening trials. Frequent 3 monthly testing in high risk women shifted detection significantly to earlier stages. However, annual testing in normal risk postmenopausal shifted stage but not sufficiently to reduce ovarian cancer mortality. To enhance early detection in high risk women even further, we added serum HE4 to the CA125 algorithm, integrated BRCA1 and BRCA2 risk models with the serum biomarker algorithm, and implemented this algorithm in a pilot early detection trial. We describe the integration of the genomic risk factors and serum protein biomarkers to derive a longitudinal joint biomarker early detection algorithm built on hierarchical longitudinal change-point models and application of Bayes factors to derive the posterior probability of the patient having a change-point in one or both biomarkers as a surrogate for having ovarian cancer.
|
Authors who are presenting talks have a * after their name.
