Fraunhofer Institute for Medical Image Computing MEVIS in Bremen have developed new methods for facilitating and accelerating adjustments to radiation therapy treatments.

Radiation therapy requires a thorough treatment plan to destroy a tumor while keeping healthy tissue untouched. Due to weight loss and tumor shrinkage, the location and size of a tumor can change over time, deeming the initial dose of treatment ineffective. This results in treatment adjustments, if the current treatment is no longer effective.

In treatment adjustment doctors take CT control images to ensure that tumors are re-targeted. These images allow doctors to make sure that sensitive organs haven’t shifted in the body and don’t accidentally get in the path of radiation and are damaged.

Because adjusting treatment can be tedious and complex, MEVIS researchers have developed algorithms that automatically align different images of the same patient. The program corrects the different positions patients assume during radiation sessions. It also distorts and shifts the images, if necessary, to align the structures. This eases determining how an ulcer changes during the course of therapy. MEVIS experts and the University Clinic in Dresden further developed and evaluated an algorithm to register lung images. It displays the lung precisely in different breathing phases.

The tool helps doctors to determine if the initial plan is still accurate and shows the uncertainty that arises from patient motion during sessions across several weeks of therapy. 3D depictions show a series of images. The bigger the movements during irradiation, the blurrier the images compared to the reference. Such illustrations could help decide if the radiation plan should be readjusted.

To plan radiation therapy, physicians contour the organs and the tumor as precisely as possible and plot their shapes. The computer gives suggestions on how the contours should look, but in practice, doctors need to adjust and correct these suggestions in a tedious process. The tool transfers the contours of the initial therapy plan to the current situation using the image registration results of the current CT image. Inaccurate contours ‘snap’ quickly into place with the help of a snapping tool. A contour propagation method for head and neck data was tested in cooperation with the Ludwig-Maximilians-Universität in Munich. The results show that doctors needed only half as much time for post-processing.