Technical optimization of breast imaging on a combined PET/MR system

2012 
369 Objectives Breast imaging can benefit from PET/MRI. However, this attractive benefit comes with a price of technical challenges, particularly when the patient is prone positioned on a breast coil. Firstly the posterior of the patient may be truncated in MR and/or PET while centering the breasts in FOV. Secondly the existing scatter correction in PET may fail due to the low uptake and special geometry of breasts. This paper presents optimization techniques for the PET image quality in breast imaging on a combined PET/MR system. Methods Over 30 patients have been scanned on the Philips Ingenuity TF PET/MR system installed at University Hospitals of Geneva. The patients were prone positioned on a breast coil. The procedure started with an MR scan for the PET attenuation correction, followed by a PET and a diagnostic MR scan. In the PET image reconstruction a 3-segment MR-based attenuation correction approach was combined with the truncation compensation using emission data. The existing single-scatter simulation algorithm was optimized for the scatter correction. Results The PET Images after optimizations showed significant improvement in terms of overall uniformity and lesion contrast. The truncation artifact posterior to the patient disappeared after applying the compensation technique. The void region near the breast bases showed proper intensity distribution with the optimized scatter correction. When the PET image was fused with the MR image, one can clearly see the benefit of the combined PET/MR system in oncologic as well as other breast applications. Conclusions The PET image quality has been optimized for prone breast imaging on the PET/MR system. The derived PET images, particularly when fused with diagnostic MR images of superb soft tissue contrast, yielded excellent diagnostic value. Note that the same methodology can also be applied to the prone breast imaging on PET/CT systems as well
    • Correction
    • Cite
    • Save
    • Machine Reading By IdeaReader
    0
    References
    3
    Citations
    NaN
    KQI
    []