What does it actually look like to do cancer research as a high school student?
In this episode of Hope Heals, I share my experience completing my first independent research project through the UCLA Emerging Scientist Program. I explored whether clinical and genomic information collected at diagnosis could be used to predict relapse in acute myeloid leukemia (AML) using machine learning.
I explain what AML is, why relapse is so difficult to predict, how I used publicly available TCGA data, the machine-learning models I tested, and what I discovered from my results. I also discuss one of the biggest questions my research raised: could measurable residual disease (MRD) and information collected throughout treatment improve relapse prediction?
Most importantly, I share what this project taught me about research—and why sometimes, the most valuable result isn't an answer, but a better question.
Here are some resources that guided me during my research, and also for you to learn more:
Dataset: https://www.cbioportal.org/study/clinicalData?id=laml_tcga_pub
https://www.mdpi.com/2073-4409/15/4/338
https://www.mdpi.com/2072-6694/15/13/3512
https://pmc.ncbi.nlm.nih.gov/articles/PMC11286513/
https://onlinelibrary.wiley.com/doi/10.1002/ajh.27625
https://www.catallyst.com/understanding-mrd/
https://www.sciencedirect.com/science/article/abs/pii/S0268960X25000852?via%3Dihub
https://www.insideprecisionmedicine.com/news-and-features/persistent-mutations-throughout-remission-linked-with-poor-survival-in-leukemia/
https://ashpublications.org/blood/article/140/12/1345/485817/Diagnosis-and-management-of-AML-in-adults-2022
https://www.nejm.org/doi/full/10.1056/NEJMoa1516192
https://www.nejm.org/doi/full/10.1056/NEJMoa1301689