ESTRO 2024 - Abstract Book

S3874

Physics - Image acquisition and processing

ESTRO 2024

Purpose/Objective:

There is a need for robust biomarkers of prostate tumour hypoxia, as this negatively affects the outcomes of prostate cancer treated with radiotherapy [1]. The relaxation rate R2* is a potential MR biomarker of hypoxia, based on its sensitivity to deoxyhaemoglobin [2,3]. The purpose of this study was to develop a blood oxygenation level dependent (BOLD) MRI sequence on an MR Linac, and validate it in healthy volunteers and patients undergoing external beam radiotherapy for prostate cancer.

Material/Methods:

A multi-echo gradient-echo sequence (voxel size = 1.2 x 1.2 x 5mm, flip angle = 80°, TR = 1367ms, first TE = 4.6ms, TE spacing = 4.6ms, number of echoes = 12, number of slices = 22, acquisition time = 5:26) underwent testing on a Leeds Test Object TO5 T1/T2 contrast phantom. The goodness of fit of the observed R2* decay modelled against the theoretical decay, as well as its ability to order the phantom’s 12 vials by the R2 of the gels they contained, were assessed. Five healthy volunteers were imaged twice with the sequence (mean spacing: 16 days, SD: 8.5 days) to assess the repeatability of prostate R2* values. Local ethics approval was obtained for all volunteers. The repeatability was quantified using the ‘test-retest’ procedure [4] to calculate a percentage repeatability coefficient (%RC). Five patients with prostate cancer undergoing stereotactic ablative radiotherapy (SABR) over five treatment fractions on the MR Linac were imaged pre-treatment and immediately following fractions 1, 3 and 5. Ethical approval was obtained through the MOMENTUM study (ClinicalTrials.gov ID: NCT04075305). R2* maps were generated (Figure 1) and the median R2* values within the whole prostate were measured to assess changes over the treatment course. The pre-treatment and fraction 1 BOLD images were used to provide a measure of R2* repeatability in patients, assuming negligible treatment effects at this time point.

Figure 1: R2* maps of the prostate generated from each imaging session using the BOLD sequence.

Results:

The sequence was demonstrated to be capable of ordering the phantom’s 12 vials by the R2 values of the gels they contained and showed an excellent goodness of fit (R 2 = 0.99 for the mean of voxel-wise fits) of the observed R2* decay. The %RC using healthy volunteer imaging was calculated to be 11.9%. When imaging patients, the median whole prostate R2* values decreased by 12.9% (range: 6.2-20.6%, p<0.008 using a 2-tailed paired sample t-test), when