Predictive Value of Machine Learning for Platinum Chemotherapy Responses in Ovarian Cancer: Systematic Review and Meta-Analysis (Preprint)

• DOI: 10.2196/preprints.48527
• Publication Date: 2024-01-22T15:30:31Z
• Authors: Qingyi Wang; Zhuo Chang; Xiaofang Liu; Yunrui Wang; Chuwen Feng; Yunlu Ping; Xiaoling Feng
• Keywords: Concordance

<sec> <title>BACKGROUND</title> Machine learning is a potentially effective method for predicting the response to platinum-based treatment ovarian cancer. However, predictive performance of various machine methods and variables still matter controversy debate. </sec> <title>OBJECTIVE</title> This study aims systematically review relevant literature on value chemotherapy responses in patients with <title>METHODS</title> Following PRISMA (Preferred Reporting Items Systematic Reviews Meta-Analyses) guidelines, we searched PubMed, Embase, Web Science, Cochrane databases studies models therapies cancer published before April 26, 2023. The Prediction Model Risk Bias Assessment tool was used evaluate risk bias included articles. Concordance index (C-index), sensitivity, specificity were prediction investigate platinum <title>RESULTS</title> A total 1749 articles examined, 19 them involving 39 eligible this study. most commonly modeling logistic regression (16/39, 41%), Extreme Gradient Boosting (4/39, 10%), support vector 10%). training cohort reported C-index models, pooled 0.806; validation 12 0.831. Support performed well both cohorts, 0.942 0.879, respectively. sensitivity 0.890, 0.790 cohort. <title>CONCLUSIONS</title> can effectively predict how respond may provide reference development or updating subsequent scoring systems.