Medical imaging is the cornerstone of modern medicine. An essential problem to solve in medical imaging is how to relate the measured machine data to the actual patient anatomy, which we want to visualize. This process is called reconstruction. In the AI for Oncology Lab, we develop deep learning-based reconstruction algorithms with the purpose to enable new treatment paradigms in radiation oncology and novel imaging protocols in radiology.
In radiation oncology, radiation is delivered in one or several fractions over several weeks to target malignant tissue. To verify the patient position with respect to the treatment plan, a CBCT or MRI (in case of MR-guided radiotherapy) scan is acquired.
Ideally, a CBCT scan could be used to adapt the treatment plan to the current patient anatomy and tumor response. However, this is limited because of the poor soft-tissue contrast, and non-calibrated intensity values which are required to compute the accumulated dose to the tumor and healthy tissue. We develop algorithms that can accurately predict the tissue attenuation values so that these can be used for daily replanning of radiotherapy.
On the other hand, in MRI-guided radiotherapy, an image can be acquired during delivery of radiation. This allows for the opportunity to adapt the treatment to the constantly changing patient. This requires a very fast reconstruction of a highly accelerated image. For this purpose, we have also developed a reconstruction toolkit called direct.
In the same spirit, we develop algorithms that allow to adapt the contrast-enhanced MRI protocol during the acquisition, such that during the acquisition algorithms can decide whether to acquire additional MR sequences, improving the diagnostic performance for this individual patient.
- P. Putzky, D. Karkalousos, J. Teuwen, N. Moriakov, B. Bakker, M. Caan, M. Welling, "i-RIM applied to the fastMRI challenge", 2019
- K. Michielsen, N. Moriakov, J. Teuwen, I. Sechopoulos, "Deep Learning-based Initialization of Iterative Reconstruction for Breast Tomosynthesis", 2020
- Y. Beauferris, J. Teuwen, D. Karkalousos, N. Moriakov, M. Caan, G. Yiasemis, L. Rodrigues, A. Lopes, H. Pedrini, L. Rittner, M. Dannecker, V. Studenyak, F. Gröger, D. Vyas, S. Faghih-Roohi, A. K. Jethi, J. C. Raju, M. Sivaprakasam, M. Lasby, N. Nogovitsyn, W. Loos, R. Frayne, R. Souza, "Multi-Coil MRI Reconstruction Challenge – Assessing Brain MRI Reconstruction Models and their Generalizability to Varying Coil Configurations", 2020
- P. M. Johnson, G. Jeong, K. Hammernik, J. Schlemper, C. Qin, J. Duan, D. Rueckert, J. Lee, N. Pezzotti, E. De Weerdt, S. Yousefi, M. S. Elmahdy, J. H. F. Van Gemert, C. Schülke, M. Doneva, T. Nielsen, S. Kastryulin, B. P. F. Lelieveldt, M. J. P. Van Osch, M. Staring, E. Z. Chen, P. Wang, X. Chen, T. Chen, V. M. Patel, S. Sun, H. Shin, Y. Jun, T. Eo, S. Kim, T. Kim, D. Hwang, P. Putzky, D. Karkalousos, J. Teuwen, N. Moriakov, B. Bakker, M. Caan, M. Welling, M. J. Muckley, F. Knoll, "Evaluation of the Robustness of Learned MR Image Reconstruction to Systematic Deviations Between Training and Test Data for the Models from the fastMRI Challenge", Machine Learning for Medical Image Reconstruction, 2021
- G. Yiasemis, J. Sonke, C. Sánchez, J. Teuwen, "Recurrent Variational Network: A Deep Learning Inverse Problem Solver applied to the task of Accelerated MRI Reconstruction", 2021
- N. Moriakov, J. Sonke, J. Teuwen, "LIRE: Learned Invertible Reconstruction for Cone Beam CT", 2022
- G. Yiasemis, C. I. Sánchez, J. Sonke, J. Teuwen, "On Retrospective k-space Subsampling schemes For Deep MRI Reconstruction", 2023