Room: AAPM ePoster Library
Purpose: In MRI-guided gynecological brachytherapy, CT imaging is still utilized for needle localization. MRI is only used for target and organs at risk delineation because needles provide no signal on standard protocols. Catheter lumen markers for MRI have been developed but are not clinically implemented yet. In this poultry phantom study, we evaluated and optimized 3D MR sequences for lumen marker detection.
Methods: A brachytherapy template with obturator and ProGuide Sharp Needles 6Fx294mm was introduced in the cavity of a cleaned turkey, partially filled with its neck and giblets. One Orion™ HDR MRI Lumen Marker was inserted in an interstitial needle and another in the obturator’s top needle. The phantom was imaged in a 3T SIEMENS Vida simulator using a Body 18 long and a spine matrix array coil. 3D T1-weighted MPRAGE and T2-weighted SPACE sequences were acquired in coronal orientation with parameters matching marker manufacturer recommendations as well as possible. Additional 3D sequences (PETRA shown to visualize empty needles and Dixon) were applied. Marker visibility was evaluated by assessing signal intensity of marker-filled catheters and ROIs in the turkey tissue and air cavity. Catheter-tissue and catheter-air intensity ratios (IR(ct), IR(ca), respectively) were calculated.
Results: Markers were detected in air but presented less signal than tissue on MPRAGE images (IR(ca):10.3±1.5, IR(ct):0.6±0.5). They were barely discerned from background and hypointense compared to tissue on SPACE images (IR(ca):8.4±2.1, IR(ct):0.4±1.0). Fat-only Dixon provided good marker distinction from both air and tissue (IR(ca):2.0±1.0, IR(ct):6.0±0.9). On PETRA, markers were identifiable in air only thanks to their darker border, and hyperintense to tissue similarly to empty needles (IR(ca):0.9±0.4, IR(ct):2.0±0.5).
Conclusion: Brachytherapy catheters with lumen markers were best visualized against tissue and air in a poultry phantom by optimized Dixon and PETRA sequences. Combination of such markers and imaging protocols would be useful for MR-only treatment planning.