Encrypted login | home

Program Information

Fools Gold: An Evaluation of the Visibility and Detectability of Intraprostatic Gold Markers in MR Images for Treatment Planning

no image available
K Singhrao

K Singhrao*, D Ruan, Y Yang, G Chee, J Lewis, UCLA School of Medicine, Los Angeles, CA

Presentations

SU-H1-GePD-J(A)-2 (Sunday, July 30, 2017) 3:00 PM - 3:30 PM Room: Joint Imaging-Therapy ePoster Lounge - A


Purpose: Localization of fiducial markers is important for treatment planning and patient setup. This project aims to quantitatively compare the detectability of implanted fiducials in various pelvic MR images.

Methods: A retrospective study of 12 prostate cancer patients with same-day MR and CT images was conducted. MR images were acquired with different sequences on a 1.5T GE Signa and a 3.0T Siemens Trio and Skyra. An expert contoured the prostate and implanted markers in each image. The contrast to noise ratio (CNR) was evaluated between the marker and the local prostate background. Sharpness was characterized by the volume difference between 25% and 85% iso-intensity line with respect to the peak marker signal. Templates were derived from images of known markers from corresponding sequences and a template matching algorithm was performed to detect objects with a similar appearance to the fiducial. Specifically, a higher than 0.95 normalized correlation value was declared as a detected marker.

Results: For the GE patient cohort, the highest CNR at 2.80 was acquired on the LAVA sequence and lowest using the T2 FSE sequence at 2.58. For the Siemens patient cohort, the highest CNR at 3.91 was acquired using the GRE QFSS sequences and lowest was observed using T2 TSE images at 2.97. No false markers were detected on T1 (FSE/TSE) and T2 (FSE/TSE) images or on the T1 VIBE sequence. The GRE QFSS and LAVA sequences detected up to 2 objects inside the prostate that had a marker-like appearance.

Conclusion: We found that sequences with T1 weighting (such as the LAVA and T1 VIBE) or fat suppression sequences (such as GRE QFSS) had low false detection counts and high CNR rates, making them more suitable for intraprostatic marker detection.

Funding Support, Disclosures, and Conflict of Interest: Research grant support provided by Varian Medical Systems.


Contact Email: