ObjectivesTo implement and evaluate a dedicated receive array coil for simultaneous PET/MR in breast cancer.
MethodsThe 16 receiver channel coil design was optimized for simultaneous PET/MR. To assess MR performance, signal-to-noise ratio, parallel imaging capability and image quality was evaluated in phantoms, volunteers and patients and compared to clinical standard protocols. For PET evaluation, quantitative 18 F-FDG PET scans of phantoms and seven patients (14 lesions) were compared to scans without coil. In PET image reconstruction, a CT-based template of the coil was combined with the MR-acquired attenuation correction (AC) map of the phantom / patient.
ResultsMR image quality was comparable to clinical MR-only exams. PET evaluation in phantoms showed regionally varying SUV underestimation (mean 22%) due to attenuation caused by the coil. This was improved by implementing the CT-based coil template in the AC (< 2% SUV underestimation). Patient data showed that including the coil in the AC increased SUV values in lesions (21% ± 9 %).
ConclusionsUsing a dedicated PET/MR breast coil, state-of-the-art MRI was possible. In PET accurate quantification and image homogeneity could be achieved, if a CT-template of this coil was included in the attenuation correction for PET image reconstruction.
KeywordsMagnetic Resonance Imaging, Positron-Emission Tomography, Breast cancer, Bilateral breast imaging, RF coil array
Key Points• State-of-the-art breast MRI using a dedicated PET/MR breast coil is feasible.• A multi-channel design facilitates shorter MR acquisition times through parallel imaging.• The MR coil inside a simultaneous PET/MR system causes PET photon attenuation.• Including a coil CT-template in PET image reconstruction, accurate quantification is recovered. [2][3][4]. Although successfully applied in many oncologic applications, simultaneous PET/MR in breast cancer has so far been delayed due to the need for a dedicated breast coil, which enables prone positioning and achieves high image quality, as a precondition for state-of-the-art breast MRI [5]. However, the presence of MR related hardware, such as a breast coil, in the PET field-of-view (FOV) causes significant attenuation of the 511 keV annihilation photons, therefore hampering PET image quality [6]. Studies found, that the presence of MR head coils, which contain a substantial amount of plastic housing material, lead to 13-19% underestimation of PET activity concentration, if not accounted for during image reconstruction [7]. The attenuation effects for "lighter" or more "transparent" MR surface coils were only 4% overall, but up to 10-15% closer to the coil surface [8]. These results demonstrate that disregarding the presence of MR coils leads to substantial regional bias in PET quantification and illustrate the importance of accurate implementation of methodology for MR coil attenuation correction (AC) in simultaneous PET/MR applications. Even though MR coils are invisible in the conventional MR-acquired AC maps, it was shown that coil A...