Pyroptosis-Related Gene Signatures Enable Robustly Diagnosis, Prognosis and Immune Responses Prediction in Uterine Corpus Endometrial Carcinoma.

Uterine corpus endometrial carcinoma (UCEC) is a gynecological malignancy with poor prognosis and high lethality rates. Pyroptosis, a pro-inflammatory programmed cell death pattern, significantly influences tumor growth, development, and metastasis. We intend to explore whether pyroptosis-related genes can be screened as targets for early detection and patient prognosis. We used nine common machine learning algorithms to build classifiers based on the pyroptosis-related genes, evaluated the classifiers' performance using metrics like the receiver operating characteristic curve (ROC), and verified the results using external datasets. Using Least Absolute Shrinkage and Selection Operator (LASSO) regression analysis, we built a predictive model. ROC and univariate/multivariate Cox analyses were used to assess the model's performance and its independence in predicting patient prognosis. We used a variety of statistical methods and algorithms to investigate the connection between tumor immunity and pyroptosis-related genes. We identified 26 pyroptosis-related genes associated with the diagnosis and prognosis of UCEC. We found the logistic regression classifier performing the best. We then constructed a predictive model based on seven PRGs about . The pyroptosis-related gene risk signature (PRGRS) effectively classified UCEC patients. We demonstrated that PRGRS independently impacted UCEC prognosis and confirmed its expression using qRT-PCR experiments. Furthermore, we found associations between PRGRS and tumor immune response. Our study highlights novel pyroptosis-related gene signatures that may be utilized for early screening and prognosis prediction in UCEC patients, offering potential targets for future research and guidance for personalized anticancer therapies.