A Self-Sorting Coronal 4D-MRI Method for Daily Image Guidance of Liver Lesions on an MR-LINAC

Tessa N. van de Lindt, Jan Jakob Sonke*, Marlies Nowee, Edwin Jansen, Vivian van Pelt, Uulke van der Heide, Martin Fast

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

22 Citations (Scopus)

Abstract

Purpose: Novel hybrid MR-LINAC devices provide MRI's superior soft-tissue contrast in the treatment room and thus have the potential to increase accuracy of liver stereotactic body radiation therapy (SBRT). Requirements for daily position verification using 4-dimensional MRI include tumor visibility and short acquisition-reconstruction time (preferably <5 min). The proposed method provides fast acquisition-reconstruction time and the flexibility to vary T1- and T2-weighting, using standard imaging sequences for straightforward implementation on an MR-LINAC. Methods and Materials: Images were acquired using a coronal 2-dimensional, multislice, single-shot turbo spin-echo (TSE) and turbo field-echo (TFE) sequence, which were repeated 30 times. An image-based self-sorting signal (ImS) was extracted from the data, and rigid registration of the diaphragm per slice position was performed and corrected for amplitude variation in the anteroposterior direction. Data were sorted into 10 bins according to amplitude and phase. ImS was validated in 4 healthy volunteers against a navigator signal. Positional variations within bins, missing data, and smoothness of the liver dome were compared between amplitude and phase binning in 10 volunteers. Tumor contrast and registration were investigated in 3 patients. Results: Each ImS was found to be in excellent agreement with the navigator signal with a correlation coefficient of >0.95 and binning differences of <1 bin. Better liver dome smoothness per bin in case of amplitude binning compared with that in phase binning (2.0-2.6 mm vs 2.4-3.7 mm, respectively) is a tradeoff for more missing data (3.5%-17.5% vs 3.5%-4.7%, respectively). Liver lesions were visible in almost all coronal TSE and TFE images, but the lesion boundary was better defined in the TSE images. Rigid registrations could be performed on the tumor area. Conclusions: An efficient self-sorted 4-dimensional MRI method was developed and validated using standard sequences and fast reconstruction on a LINAC-integrated MRI scanner providing good tumor visibility for daily image-guided liver stereotactic body radiation therapy.

Original languageEnglish
Pages (from-to)875-884
Number of pages10
JournalInternational Journal of Radiation Oncology Biology Physics
Volume102
Issue number4
DOIs
Publication statusPublished - 15 Nov 2018

Fingerprint

Dive into the research topics of 'A Self-Sorting Coronal 4D-MRI Method for Daily Image Guidance of Liver Lesions on an MR-LINAC'. Together they form a unique fingerprint.

Cite this