Long-Term Survival and Second Malignant Tumor Prediction in Pediatric, Adolescent, and Young Adult Cancer Survivors Using Random Survival Forests: A SEER Analysis

• Activity Number: 522 - Life Science Applications of Data Science
• Type: Contributed
• Date/Time: Thursday, August 11, 2022 : 8:30 AM to 10:20 AM
• Sponsor: Section on Statistical Learning and Data Science
• Author(s): Ivy Zhang* and Jun Deng and Gregory Hart and Bo Qin
• Companies: Yale University and Yale University and Bill & Melinda Gates Foundation and Dartmouth College
• Keywords: random survival forest; cancer survivor; second malignancy; long-term outcomes; pediatric cancer; clinical risk

Abstract

As the number of childhood, adolescent, and young adult cancer survivors grows from recent advances in early detection, treatments, and supportive care, there is an increasing need and interest to investigate long-term outcomes such as survival and second malignancy. However, there are currently no machine learning models focused on long-term outcome prediction for children, adolescent, and young adult cancer survivors. The models built in this study aim to fill that gap by achieving the following: (1) predict 30-year survival in these age groups; and (2) predict risk and site of a second tumor within 30 years of the first tumor diagnosis. In the future, survival and second tumor models such as the ones developed in this study could help physicians navigate overwhelming quantities of data by quickly identifying highest risk individuals and ultimately improving cancer survivor outcomes.