Pancreatic cancer is among the deadliest types of cancer, and effective and safe treatment remains challenging. MR-guided radiotherapy, a noninvasive, low-toxic, local treatment, has recently emerged as a new treatment option for inoperable pancreatic cancer. The aim of this thesis was to explore several technical developments for the planning and delivery of radiotherapy for pancreatic cancer on the 1.5 T MR-Linac. These developments are needed to transform contemporary MR-guided radiotherapy into a next-generation treatment for pancreatic cancer, which extensively capitalizes on the capabilities of the 1.5 T MR-Linac.
In the first part of this thesis, we focused on (online adaptive) treatment planning. The conventional MRI sequences for online treatment adaptation are prone to blurring artifacts due to respiratory motion, but we found that motion-robust radial MRI sequences can result in much sharper, higher quality images, potentially increasing delineation accuracy and speed. Next, we founded a worldwide consortium of MR-Linac centers to create a consensus protocol on treatment planning for pancreatic cancer. We hope that this collaboration will harmonize the large variations currently found in protocols and strategies between centers around the world that treat pancreatic cancer on the 1.5 T MR-Linac.
The second part revolved around intrafraction motion and treatment delivery. The onboard imaging system of the MR-Linac allowed us to fully characterize pancreatic tumor motion during treatment, in terms of respiratory motion, tumor drift, and their respective variations on a day-to-day and patient-to-patient basis. The measured intrafraction motion was used for retrospective dose calculation, thereby allowing us to evaluate whether the tumor motion during irradiation had lead to dosimetric errors such as tumor underdosage or OAR overdosage. We found that this dosimetric impact is generally modest, but small drifts are potentially much more problematic than large respiratory amplitudes. We further developed intrafraction dose accumulation by designing a 3D dynamic MRI scan capable of capturing complete volumetric information in real-time, allowing for advanced deformable dose accumulation. Lastly, we report on the first patients, all with upper abdominal tumors, treated with active motion management on the 1.5 T MR-Linac. This system is capable of fully MR-based respiratory gating, and is enhanced with baseline drift correction capabilities. We showed through dosimetric analysis that this system improves congruence between the planned dose and delivered dose.
To conclude, MR-guided radiotherapy has opened the way for a new, safe, local treatment for inoperable pancreatic cancer. The full potential of the MR-Linac, however, has been far from realized. This thesis has shown that MRI is an incredibly versatile tool that can be used in every facet of the online adaptive workflow, allowing us to gain more insight into the treatment and ultimately optimizing the planning and delivery. Ablative radiotherapy for pancreatic cancer with the 1.5 T MR-Linac is one of the most intricate procedures in external beam radiotherapy, and the ongoing developments will only further increase complexity. Therefore, the informal dissemination of clinical experiences and treatment techniques, as well as collaboration between MR-Linac centers, remains crucial for further advancement of radiotherapy for pancreatic cancer.