ESTRO 2024 - Abstract Book

S3918

Physics - Image acquisition and processing

ESTRO 2024

To obtain the shift in center of mass (CoM), the absolute difference between the CoM in T1W-image and different b value images and the ADC-map was computed for every marker, respectively (Hero imaging, Umeå Sweden).

The artefact extension caused by the markers in the images was obtained in every spatial direction by performing a statistical F-test. The null hypothesis was defined as no artefact being present in the voxel. A significance level of 1% was set.

Finally, a radial margin was determined for markers in dMRI and the ADC-map based on the artefact extension obtained from the F-test and the shift in CoM. The radial margin was depicted on in-vivo prostate images.

Results:

The radial margin along PE-direction was found to be 7.4 mm and 5.2 mm for Marker 1, 9.2 mm and 6.3 mm for Marker 2, and 10.8 mm and 7.2 mm for Marker 3 in dMRI and ADC-map, respectively. Along the FE-direction the radial margin was found to be 6.2 mm and 5.2 mm for Marker 1, 4.2 mm and 3.1 mm for Marker 2, and 4.1 mm and 3.6 mm for Marker 3 in dMRI and ADC-map, respectively. Along the SE-direction the radial margin was found to be 7.2 mm and 5.1 mm for Marker 1, 7.1 mm and 7.0 mm for Marker 2, and 7.3 mm and 5.3 mm for Marker 3 in dMRI and ADC map, respectively. The obtained values for artefact extension were similar for the different b-values, however smaller artefact extensions were obtained overall in the ADC-map. This can be due to the intensity shift from the artefact being relatively uniform across different b-values. Consequently, the slope of the signal intensity as a function of b-values, and thus the ADC values, were less adversely impacted.

The radial margins found for Marker 2 were shown as an example on in-vivo prostate images (Fig. 2), due to the positioning of Marker 2 being closest to a clinical scenario.