Skip to main content
ARS Home » Plains Area » Houston, Texas » Children's Nutrition Research Center » Research » Publications at this Location » Publication #385717
Research Project: Metabolic and Epigenetic Regulation of Nutritional Metabolism

Location: Children's Nutrition Research Center

Title: Predict multicategory causes of death in lung cancer patients using clinicopathologic factors

Author
item DENG, FEI - SHANGHAI INSTITUTE OF TECHNOLOGY
item ZHOU, HAIJUN - METHODIST HOSPITAL
item LIN, YONG - RUTGERS CANCER INSTITUTE OF NEW JERSEY
item HEIM, JOHN - PRINCETON UNIVERSITY
item SHEN, LANLAN - CHILDREN'S NUTRITION RESEARCH CENTER (CNRC)
item LI, YUAN - FUDAN UNIVERSITY
item ZHANG, LANJING - RUTGERS CANCER INSTITUTE OF NEW JERSEY

Submitted to: Computers in Biology and Medicine
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 11/29/2020
Publication Date: 2/1/2021
Citation: Deng, F., Zhou, H., Lin, Y., Heim, J.A., Shen, L., Li, Y., Zhang, L. 2021. Predict multicategory causes of death in lung cancer patients using clinicopathologic factors. Computers in Biology and Medicine. 129:104161. https://doi.org/10.1016/j.compbiomed.2020.104161.
DOI: https://doi.org/10.1016/j.compbiomed.2020.104161

Interpretive Summary: Lung cancer is the most prevalent cancer and the second leading-cause of cancer deaths among men in the US. It is largely unknown, however, whether multicategory variables (age and clinical/biological variables) can be used to accurately predict non-cancer causes of death among these individuals. We therefore applied a machine learning algorithm with a random forest model to predict survival outcomes in 42,257 qualified lung cancer patients. We showed that random forest model outperforms the standard multinomial logistic regression model in prediction accuracy. Our finding provides a needed strategy to incorporate genetic and epigenetic variables to aid in development of precision medicine (i.e., nutrition and lifestyle to improve health).

Technical Abstract: Random forests (RF) is a widely used machine-learning algorithm, and outperforms many other machine learning algorithms in prediction-accuracy. But it is rarely used for predicting causes of death (COD) in cancer patients. On the other hand, multicategory COD are difficult to classify in lung cancer patients, largely because they have multiple labels (versus binary labels). We tuned RF algorithms to classify 5-category COD among the lung cancer patients in the surveillance, epidemiology and end results-18, whose lung cancers were diagnosed in 2004, for the completeness in their follow-up. The patients were randomly divided into training and validation sets (1:1 and 4:1 sample-splits). We compared the prediction accuracy of the tuned RF and multinomial logistic regression (MLR) models. We included 42,257 qualified lung cancers in the database. The COD were lung cancer (72.41%), other causes or alive (14.43%), non-lung cancer (6.85%), cardiovascular disease (5.35%), and infection (0.96%). The tuned RF model with 300 iterations and 10 variables outperformed the MLR model (accuracy =69.8% vs 64.6%, 1:1 sample-split), while 4:1 sample-split produced lower prediction-accuracy than 1:1 sample-split. The top-10 important factors in the RF model were sex, chemotherapy status, age (65+vs <65 years), radiotherapy status, nodal status, T category, histology type and laterality, all of which except T category and laterality were also important in MLR model. We tuned RF models to predict 5-category CODs in lung cancer patients, and show RF outperforms MLR in prediction accuracy. We also identified the factors associated with these COD.