Colorectal cancer (CRC) is one of the most common types of cancer, with an estimated 1.8 million new cases annually. In Brazil, around 41 thousand new cases of CRC are estimated between 2020 and 2022, making it a prime opportunity for mortality and survival prediction studies. A recent study sought to evaluate and compare the accuracy of three machine learning algorithms in predicting CRC patient survival in São Paulo, Brazil, based on data from the Hospital Based Cancer Registries of São Paulo state (RHC-SP). After analyzing data from over 31,000 patients, the study found that both Random Forest and XGBoost models provided better performance than neural networks and Naive Bayes models. The top five most important features that influenced patient survival included clinical staging, year of diagnosis, presence of recurrence, surgery, and age.
The company or organization behind the study is not mentioned in the article.
The person(s) behind the study are also not mentioned in the article.
