Click here to

Session: Education General ePoster Viewing [Return to Session]

Incorporating Statistical Learning and Artificial Intelligence Education to Medical Physics Graduate Training

C Cardenas*, L Court, University of Texas MD Anderson Cancer Center, Houston, TX


PO-GePV-E-13 (Sunday, 7/25/2021)   [Eastern Time (GMT-4)]

Purpose: To evaluate the value of a web-based educational tool developed to introduce medical physics graduate trainees to novel topics in statistical learning and artificial intelligence (AI).

Methods: A self-paced tutorial-course was containing modules introducing trainees to 1) Python programming, 2) DICOM file operations, 3) Statistical Learning concepts, and 4) Deep Learning concepts was created. The materials were developed such as that no prior programming and/or statistical learning/AI background were required. This course was developed using Python and Jupyter notebooks, with many notebooks covering materials in each of the four topics. Eight first-year medical physics PhD graduate trainees at our institution used this tutorial-course materials over the course of a twelve-week research-tutorial to enhance knowledge for their projects. Weekly-meetings were held to go over materials covered and evaluate research progress. After completion of the self-paced course and research-tutorial, all trainees were asked to complete a survey to assess the value of the educational tool.

Results: All trainees completed the tutorial-course while taking a typical first-year academic course load. 100% of trainees thought the tutorial-course was helpful in developing a strong background in statistical learning and AI. 75% of trainees feel confident that the tutorial has prepared them well to implement AI techniques into their PhD research. Several trainees submitted abstracts on their research-tutorial work for the AAPM annual meeting. One trainee won the Young Investigator Symposium award at the 2020 SWAAPM meeting based on her research-tutorial results.

Conclusion: Training in statistical learning and AI are currently not part of the medical physics graduate education. We developed a structured self-paced tutorial-course to introduce trainees to basic principals in these topics while they worked on individual research projects. The course was positively-received by the trainees and has potential in bridging the gap in knowledge for future medical physicists in these fields.



    Statistical Analysis, Image Analysis, Segmentation


    Education: Knowledge of principles and generalizations

    Contact Email