Development and Validation of Predictive Risk Scores for Ovarian Clear Cell Carcinoma: A Penalized Regression Model

Development and Validation of Predictive Risk Scores for Ovarian Clear Cell Carcinoma: A Penalized Regression Model

ABSTRACT Background The precision management of ovarian clear cell carcinoma (OCCC) faces limitations due to the absence of personalized prognostic tools. This study aimed to establish predictive risk scores to enable effective stratification and tailored treatment of OCCC. Methods Retrospective dat...

Full description

Saved in:
Bibliographic Details
Main Authors: Shohei Iyoshi, Ryo Emoto, Nobuhisa Yoshikawa, Kosuke Yoshida, Masato Yoshihara, Satoshi Tamauchi, Akira Yokoi, Yusuke Shimizu, Yoshiki Ikeda, Kaoru Niimi, Michiyasu Kawai, Shigeyuki Matsui, Hiroaki Kajiyama
Format: Article
Language:English
Published: Wiley 2025-08-01
Series:Cancer Medicine
Subjects:
clear cell carcinoma
machine learning
ovarian cancer
penalized regression
prediction model
Online Access:https://doi.org/10.1002/cam4.71118
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:ABSTRACT Background The precision management of ovarian clear cell carcinoma (OCCC) faces limitations due to the absence of personalized prognostic tools. This study aimed to establish predictive risk scores to enable effective stratification and tailored treatment of OCCC. Methods Retrospective data from 206 OCCC patients treated between 2004 and 2019 at two hospitals were analyzed. Penalized regression models were utilized to develop three risk scores based on preoperative laboratory results and intraoperative findings. These scores underwent internal and external validation. Results The median follow‐up periods were 65.7 and 44.0 months for the derivation and validation cohorts, respectively. In internal validation with the derivation cohort, all three risk scores effectively identified the high‐risk group for tumor recurrence. Upon validation with the external cohort, Risk Score 3, which incorporated variables selected in most cross‐validations by the penalized regression (Elastic Net), distinctly differentiated the high‐ and low‐risk groups (p = 0.03). Risk Score 2, consisting solely of preoperatively available variables, also demonstrated marginal significance (p = 0.08). Conclusion Our findings underscore the significance and utility of the developed risk scores in tailoring personalized treatments for patients with OCCC.
ISSN:2045-7634

Similar Items