Harvard Catalyst Profiles

Head and Neck Cancers are diagnosed in approximately 60,000 patients in the U.S. annually. The goal of this project is to utilize advanced techniques in machine learning and artificial intelligence, including deep learning neural networks, to predict underlying characteristics of head and neck cancers and patient outcomes based on diagnostic imaging studies. Preliminary work (Kann et al, Nature Scientific Reports, 2018; Kann et al, Journal of Clinical Oncology, 2020) has demonstrated that deep learning can predict lymph node malignancy and extranodal extension, two factors that are critical in determining proper patient management and risk-stratification. The goal of the Student Project will be explore ways to optimize the deep learning platform and experiment with various cutting-edge neural network algorithms. The expectation is that the project will result in a clinically usable AI-application for head and neck cancer patients. Students with an interest in oncology, cancer imaging, computer science, data science, bioinformatics, artificial intelligence are encouraged. The project aims and time commitment are flexible and can be tailored to the availability of the student. There are also several other machine-learning and digital health related projects ongoing in our group, The Mass General Brigham/Harvard Medical School Artificial Intelligence in Medicine Program (https://aim.hms.harvard.edu), including work with MRI and brain tumors, as well as lung cancer, and I would be happy to discuss.