Sophie Chen
Caddo Parish Magnet High School
Shreveport, LA
Intraoperative Histological Analysis of Squamous Cell Carcinoma Tumor Margins Using a Convolutional Neural Network
Sophie Chen created a machine learning model she hopes will improve the microscopic identification of benign vs. cancerous tissues during surgery, given that the current tissue analyzation technique is prone to error and can be time-consuming. Her findings could potentially reduce the risk of recurrence and metastasis of squamous cell carcinomas.
View Poster
Sophie Chen, 17, of Shreveport, developed a convolutional neural network (CNN) machine learning model to review microscopic images of tissue during removal of squamous cell carcinoma (SCC) for her Regeneron Science Talent Search computational biology and bioinformatics project. Currently, analyzing tissue during removal of SCC can be time-consuming and is prone to error. Sophie’s CNN model was trained to look at slides of removed tissues and recognize the basic differences between tumor, inflammation, and benign areas.
Her model proved to be robust, fast, and able to accurately identify suspicious tissues from images. Sophie believes it to be an important tool to assist surgeons in determining whether the tumor is fully removed.
Sophie attends Caddo Parish Magnet High School where she participates in speech and debate, plays violin in the school orchestra, and runs a mental health club. She also co-founded a nonprofit that provides access to computer science education to young girls in underserved communities. The daughter of Xinqian Jiang and Shoujun Chen, Sophie wants to pursue a Ph.D. in computational biology to harness the power of A.I. to drive real change.
Beyond the Project
As part of the nonprofit she founded, Sophie collaborated with local groups to lead a six-week introduction to coding during a summer program at her local library, and she started a podcast and blog interviewing women in tech.
FUN FACTS: To blow off steam, Sophie loves singing karaoke, whether it’s “singing along to K-pop and indie or harmonizing to classics from the 2000s.”
