ESTRO 2024 - Abstract Book

S3804

Physics - Image acquisition and processing

ESTRO 2024

Analysis of motion-compensated MRI efficiency in lung cancer patients based on a 1D-navigator signal

Frédérique P.D. van Gameren 1 , Katrinus Keijnemans 1 , Pim T.S. Borman 1 , Cornelis A.T. van den Berg 2 , Martin F. Fast 1 , Astrid L.H.M.W. van Lier 1 1 University Medical Center, Department of Radiotherapy, Utrecht, Netherlands. 2 University Medical Center, Department of Computational Imaging, Utrecht, Netherlands

Purpose/Objective:

Respiratory motion is an important source of uncertainty that affects imaging and delivery for abdominal and thoracic external-beam radiotherapy. Motion-compensated imaging is aimed at obtaining sharp images in a predefined respiratory phase at the cost of longer acquisition times. These images can be used to ease delineation. In this study, we quantify unguided respiratory motion characteristics in lung cancer patients for motion-compensated MRI. We gain insights in their correlation with imaging efficiency, which can be used for motion informed imaging as well as for gated radiotherapy treatment purposes.

Material/Methods:

Eleven datasets were acquired in five stage III lung cancer patients (age: 52-73y, female/male: 3/2), of which eight were acquired on a 1.5T MR-sim (Philips Healthcare, the Netherlands) and three on a 1.5T Unity MR-linac (Elekta AB, Sweden). An unguided respiratory-triggered T 2 -weighted TSE-MVXD (PROPELLOR) sequence¹ was used in combination with a 1D respiratory pencil-beam navigator (1D-RNAV) acquisition on the liver-lung interface. For each scan, a motion-compensation window with a width of 5mm on the end-exhale phase of the 1D-RNAV signal was predefined. One boundary of the window was positioned at the most cranial respiratory position detected during the first 30s of the scan. The other boundary was positioned 5mm caudally (Figure 1). The imaging sequence was triggered to acquire data whenever the 1D-RNAV signal was within the window, resulting in a motion-compensated image. The expected imaging efficiency was determined for each dataset by dividing the number of datapoints within the window by the total number of 1D-RNAV datapoints. Furthermore, an alternative position of the window (width 5mm) was defined retrospectively by optimizing for the maximum imaging efficiency for each 1D-RNAV signal.