Developing and evaluating deformable dose accumulation for image-guided radiotherapy: "Dose were the DACE"

Radiotherapy aims to deliver a destructive radiation dose to cancer tumours while limiting the dose to healthy surrounding organs.

We explore how MRI scans acquired during radiotherapy –using a special machine that combines MRI and radiation delivery developed in the UMC Utrecht– can be used to improve cancer treatments. By analysing how the anatomy moves during treatment, such as the bladder filling or gas passing through the rectum, we estimate how these changes affect the radiation dose received by the patient. Taking the actually delivered dose into account can improve the remainder of the treatment.

In this work, we simulate both the typical motions seen in prostate cancer patients (and how these would show on MRI scans) and the radiotherapy delivery itself. This lets us test different ways to estimate the actual delivered dose, including a new method we developed. Our method incorporates information about organ positions identified by clinicians and often gives more accurate results.

We also examine how certain we can be about these estimates, which is critical for making clinical decisions. Our simulations help us compare different approaches, including again a new method we developed. This method provides uncertainty estimates that are easy to interpret and to compare to the uncertainty of assuming that the planned dose is delivered. Overall, our approach is more reliable than the alternative of assuming no uncertainty.

Our work shows that using the MRI scans acquired during treatment can lead to more accurate and personalized radiotherapy, potentially improving outcomes for prostate cancer patients.