Abstract
Association Between Endometrial Cancer Risk Classification and Gene Expression in The Cancer Genome Atlas Database
Obstetrics and gynecology (New York. 1953), Vol.123 Suppl 1(Supplement 1), pp.90S-90S
05/2014
DOI: 10.1097/01.AOG.0000447424.95726.84
Abstract
INTRODUCTION:To investigate if gene expression can differentiate endometrial cancer risk classification based on GOG 99 study in The Cancer Genome Atlas patients.
METHODS:Gene expression was extracted from 271 endometrioid endometrial cancer samples in the TCGA database. Patients were stratified to high risk, high intermediate risk, and low or low intermediate risk based on clinicopathologic parameters from GOG 99. Genes differentially expressed between the classes at univariate significance level of less than .001 were included in the gene signature.
RESULTS:There were 167 patients in the low-risk group, 81 patients in the high intermediate-risk group, and 23 patients in the high-risk group in The Cancer Genome Atlas database. GOG 99 risk classification and FIGO (International Federation of Gynecology and Obstetrics) stage were independently significant for survival in the multivariate analysis (P<.001). A total of 18,048 genes were included in the initial analysis. Four hundred forty-six genes classified high risk, high intermediate risk, and low or low intermediate risk subgroup individually (P=.01) with an accuracy of 58% and a κ coefficient of 0.32. When combining high-risk and high intermediate-risk groups, the gene signature compared with low-risk patients included 608 genes (P=.01) with an accuracy of 69% and a κ coefficient of 0.36 (area under the curve 0.69), closer to the values of the gene signature including only low-risk compared with high-risk patients (accuracy 81%, κ 0.33, area under the curve 0.74).
CONCLUSIONS:Gene expression analysis could differentiate risk groups in endometrioid endometrial cancer. The gene signature grouping high intermediate risk and high risk compared with low risk seems to be more accurate than using all three risk groups. This gene signature may help to identify higher risk patients with endometrial cancer before surgery.
Details
- Title: Subtitle
- Association Between Endometrial Cancer Risk Classification and Gene Expression in The Cancer Genome Atlas Database
- Creators
- Hye Chon - H. Lee Moffitt Cancer Center and Research Institute, Tampa, FLReem Abdallah - Moffitt Cancer CenterJesus Bosquet - University of Iowa, Obstetrics and Gynecology
- Resource Type
- Abstract
- Publication Details
- Obstetrics and gynecology (New York. 1953), Vol.123 Suppl 1(Supplement 1), pp.90S-90S
- Publisher
- The American College of Obstetricians and Gynecologists
- DOI
- 10.1097/01.AOG.0000447424.95726.84
- ISSN
- 0029-7844
- eISSN
- 1873-233X
- Language
- English
- Date published
- 05/2014
- Academic Unit
- Obstetrics and Gynecology
- Record Identifier
- 9984386058702771
Metrics
5 Record Views