Room: AAPM ePoster Library
To investigate the feasibility of Quantitative Susceptibility Mapping (QSM) for detection of implanted gold fiducial markers in prostate patients towards MRI-only SBRT treatment planning.
Seven prostate cancer patients each with three implanted gold fiducial markers participated in this study. All patients were scanned on a 3T MR scanner (Philips Achieva), on the same day or the day after implant procedure with a 3D double-echo gradient echo sequence (in addition to the T1/T2 weighted scans) with the following parameters: TE1/TE2/TR=2.3/4.6/5.6ms, FA=15?, Bandwidth=868Hz/pixel, FOV=250x250x100mm³, resolution=1.5mm³ isotropic, and scan time=6min. Gradient echo magnitude and phase images were processed to estimate susceptibility values (using QSM pipeline illustrated in Fig.1a). The estimated QSM was inverted for hyperintense depiction of gold markers. Markers were digitized manually on CT and QSM and were rigidly registered through Iterative Closest Point (ICP) algorithm. The differences in relative Euclidian distance between each pair of markers (pre-registration) and the distance between CT and MR-derived position of each marker (post-registration) were compared.
All implanted markers were visualized and detected successfully. Two patients had intraprostatic calcifications. Although calcification and gold are both diamagnetic and appear hyperintense on inverted QSM, the average estimated susceptibility values for gold markers and calcifications were significantly different thus were differentiable on QSM (Fig.1k/m). The estimated susceptibility/HU values on QSM/CT for gold and calcifications were 31.5±3ppm/1220±101HU and 14.6±0.9ppm/439±127HU, respectively.
The average difference between pair-wise Euclidean distances between markers on CT and QSM was 0.7±0.6mm. After rigid CT-QSM registration, for each marker the average distance between CT-derived and MR-derived positions was 0.7±0.5mm (p-value=0.7).
This work indicates the feasibility of QSM to replace CT for marker localization in SBRT treatment planning workflow. This is an ongoing study and future works includes more patient scans, automated fiducial localization, and dosimetric analysis incorporating atlas-based pseudo-CT for electron density estimation.