| Autor: |
Siddik Sarkar, Sarbar Ali Saha, Poulomi Sarkar, Sarthak Banerjee, Pralay Mitra |
| Rok vydání: |
2022 |
| DOI: |
10.21203/rs.3.rs-1560503/v1 |
| Popis: |
Background Clinopathological parameters such as age, residual tumor, grade and stage are often used to predict the survival of ovarian cancer patients, but still the 5-year survival of high grade serous ovarian cancer remains > 30%. Here we established a molecular gene signature-based scoring system using data of ovarian cancer cohorts that could potentially determine the median overall survival of high-grade serous ovarian cancer patients. Methods The data mining and analysis of raw expression data spanning over HGSOC cohorts (n = 4784) deposited in various data repositories were performed. The feature extraction/ selection tool using Cox, LASSO regression was conducted on training data to obtain predicted genes along with the coefficients that contribute to obtaining molecular prognostic score (mPS). The receiver operator characteristics curve were plotted to study prediction efficiency of mPS . findings: The 20 gene-based mPS predicted the 5-year overall survival with 70% efficiency both in training (n = 491) and test datasets (n = 491) and also applicable in training OTTA-SPOT HGSOC samples (n = 3762). The mPS has significant impact (HR [95%CI] = 6.1 [3.65–10.3]; p |
| Databáze: |
OpenAIRE |
| Externí odkaz: |
|
