Therefore, the use of neural networks in this study was not further explored. The objective of the present study is to evaluate and compare the validity of three artificial intelligence algorithms for predicting the survival of patients with CRC treated in São Paulo, the most populous state in Brazil, from 2000 to 2021, based on data from the RHC-SP.įinally, some neural network possibilities were tested for these data, both sequential and more complex models, but the performance was lower than the machine learning algorithms used, obtaining lower accuracy than the Random Forest and XGBoost models in all tests performed. The Hospital Based Cancer Registries of São Paulo state (RHC-SP), based at the Fundação Oncocentro do Estado de São Paulo (FOSP), covers a population of approximately 30 million inhabitants, with 33,000 cases of colorectal cancer, configuring a unique opportunity to carry out of mortality or survival prediction studies for Brazilian patients. In Brazil, around 41 thousand new cases are estimated between 20 10. It is estimated that approximately 10% of cancer cases in the world in 2020 will be in the colon or rectum, corresponding to approximately 1.8 million new cases annually 9, with an increasing trend in both genders. In recent years, cancer registry data, such as the US Surveillance, Epidemiology and End Results (SEER), have been used to predict mortality or survival in the US using artificial intelligence 5, 7.Ĭolorectal cancer (CRC) is among the ten most incidents in the world 8. Since models using machine learning do not provide structure and parameters in an explicit and easily interpretable way, it becomes crucial to test their use and their accuracy with real data. They are quickly and easily adaptable to new realities and their use has been tested in cancer studies 6. Currently, artificial intelligence (AI) has been collaborating in the diagnosis of several diseases 2, 3 and in the evaluation of survival 4, the machine learning technique, an application based on artificial intelligence data, in which systems learn and improve automatically without explicit programming 5, has been used in the search for an evaluation that demands fewer human resources, possibly more accurate and perennial survival. The most common statistical models are linear and depart from explicit descriptions of the relationships between data. Such techniques have limitations related to the adaptation of models, changes in the reality, and potential reduction in accuracy over time 1. Additionally, the identification and validation of prognostic factors are important to guide the treatment protocol.Įpidemiological studies have used statistical models, based on pre-established predictors for the prognosis of survival in patients with colorectal cancer (CRC). The analysis of the survival of cancer patients is fundamental for the planning and evaluation of health services. Using part of the dataset to validate our models, the results of the predictors were around 77% of accuracy, with AUC close to 0.86, and the most important column was the clinical staging in all of them. The machine learning models used provided us the predictions and the most important features for each one of the algorithms of the studies. Data were extracted from Hospital Based Cancer Registries of São Paulo, which is coordinated by Fundação Oncocentro de São Paulo, containing patients with colorectal cancer from São Paulo state, Brazil, treated between 20. In this study, five different classifications were performed, considering patients’ survival. The use of machine learning algorithms for cancer studies has grown in recent years, and they can provide important information to medicine, in addition to making predictions based on the data. This increase in cases further intensifies the interest and importance of studies related to the topic, especially using new approaches. In Brazil, the scenery is the same, around 41 thousand new cases were estimated in the last 3 years. Colorectal cancer is one of the most incident types of cancer in the world, with almost 2 million new cases annually.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |