Machine Learning
The field of radiation therapy is ripe for Artificial Intelligence (AI) and computational modeling due to the large amount of data. These models can be used to automate and improve treatment processes. We are developing AI based methods to automate treatment for brachytherapy of cervical cancer using advanced deep learning models and optimization algorithms.
The above figure shows one of our automated planning methods, which uses a convolutional neural network (CNN) to predict 3D radiation dose from patient-specific imaging inputs. An optimization algorithm is then used to convert this predicted dose to a set of machine parameters, or dwell times (DTs), so that the treatment can be delivered to the patient.
This technology could reduce the time patients spend under anesthesia or discomfort by up to an hour or more, decreasing the use of costly hospital resources and improving patients’ brachytherapy experience. We are currently performing multi-institutional validation at BC Cancer and developing methods to adapt models to clinic-specific practice and patient populations.
In addition to automated planning, we have developed models for brachytherapy decision support, which can help physicians select the most effective device(s) to insert for treatment. Our goal is for all aspects of brachytherapy treatment to be informed by or automated with models trained on prior patient data.
Team Members and Collaborators
Team members: Sandra Meyers, Hannah Williams, Michael Kudla, Roja Zakariaee, Zahra Anjomani, Samantha Lloyd, Haley Patrick
Collaborators: Nuno Vasconcelos, Lance Moore, Aranyo Mitra
Publications
1. Mitra, A., Moore, L.C., Kallis, K., Rash, D.L., Einck, J.P., Nwachukwu, C., Yashar, C.M., Mayadev, J.S., Ayala-Peacock, D.N., Balogun, O.D., Chino, J.P., Contreras, J.A., Zoberi, I., Kisling, K., Zou, J., Vasconcelos, N., Meyers, S.M. Rapid AI Auto-Planning Rivals Manual Expert Planning for Cervical Brachytherapy. Practical Radiation Oncology 2026 S1879-8500(26)00143-8.doi: 10.1016/j.prro.2026.05.002. Online ahead of print.
2. Truong R., Moore, L.C., Mitra, A., Kallis, K., Kisling, K., Vasconcelos, N., Meyers, S.M. Integration of single-click, AI-based brachytherapy auto-planning for cervical cancer within a treatment planning system. Brachytherapy 2025;25(1):206-213.
3. Moore, L.C., Ahern, F., Li, L., Kallis, K., Kisling, K., Cortes, K.G., Nwachukwu, C., Rash, R., Yashar, C., Mayadev, J., Zou, J., Vasconcelos, N., Meyers, S.M. Neural Network Dose Prediction for Cervical Brachytherapy: Overcoming Data Scarcity for Applicator-Specific Models. Medical Physics 2024;51(7):4591-4606.
4. Kallis, K., Moore, L., Cortes, K.G., Brown, D., Mayadev, J., Moore, K.L., Meyers, S.M. Automated treatment planning framework for brachytherapy of cervical cancer using 3D dose predictions. Physics in Medicine and Biology 2023;68(8): 10.1088/1361-6560/acc37c.
5. Cortes, K.G., Kallis, K., Simon, A., Mayadev, J., Meyers, S.M., Moore, K.L. Knowledge-Based Three-Dimensional Dose Prediction for Tandem-And-Ovoid Brachytherapy. Brachytherapy 2022;21(4):532-542.
6. Kallis, K., Mayadev, J., Covele, B., Brown, D., Scanderbeg, D., Simon, A., Frisbie-Firsching, H., Yashar, C.M., Einck, J., Mell, L.K., Moore, K.L., Meyers, S.M. Evaluation of dose differences between intracavitary applicators for cervical brachytherapy using knowledge-based models. Brachytherapy 2021;20(6):1323-1333.
7. Kallis, K., Mayadev, J., Kisling, K., B., Brown, D., Scanderbeg, D., Ray, X., Cortes, K., Simon, A., Yashar, C.M., Einck, J., Mell, L.K., Moore, K.L., Meyers, S.M. Knowledge-based dose prediction models to inform gynecologic brachytherapy needle supplementation for locally advanced cervical cancer. Brachytherapy 2021;20(6):1187-1199.
8. Yusufaly, T.I, Kallis, K., Simon, A., Mayadev, J. Yashar, C.M., Einck, J.P., Mell, L.K., Brown, D., Scanderbeg, D., Hild, S.J., Covele, B., Moore, K.L., Meyers, S.M. A knowledge-based organ dose prediction tool for brachytherapy treatment planning of cervical cancer patients. Brachytherapy 2020;19(5):624-634.