A dedicated breast PET/CT scanner that produces combined dual-modality images could improve the detection of breast diseases in challenging cases. A team at West Virginia University School of Medicine has been developing dedicated breast-PET imaging systems since 2000. Performance tests of their newest breast PET/CT prototype demonstrated that its performance was similar to that of current dedicated breast-CT and breast-PET systems (Med Phys. doi.org/10.1002/mp12780).
Women with dense, cystic, post-surgical and augmented breasts can be difficult to evaluate using current imaging techniques. PET is a good candidate for supplemental breast imaging of women with indeterminate mammograms, because tissue contrast is based on physiologic differences (characterized by radiotracer uptake) rather than tissue density.
Integrating CT with PET adds anatomical information regarding the size and shape of lesions identified with PET. CT data could also be used in corrections for physical processes that degrade PET images, such as Compton scattering and photon attenuation. Accurate quantitation of radiotracer uptake could also be improved by using CT images to calculate partial volume corrections for small structures identified in PET scans.
The PET/CT scanner
The PET/CT scanner developed by Raymond Raylman and co-authors consists of a PET scanner with two detector heads coupled to a 4×3 array of flat-panel positron-sensitive photomultipliers (PSPMTs). Specialized electronics reduce output signals from 64 to four channels for each PSPMT and produce an output signal whose amplitude represents the total amount of light detected by the PSPMT. The system design maximizes collection of scintillation light and light produced by elements at the edges of the scintillator array and reduces the processing load.
PET data are acquired by rotating the detectors in step-and-shoot mode, with the dwell time at each position selected by the user. The estimated imaging time is approximately six minutes per breast. The nominal field-of-view (FOV) is 20 cm (transaxial) and 15 cm (axial), and the nominal reconstructed voxel size is 1 mm3. The cone-beam CT scanner comprises a tungsten filament pulsed X-ray source and a flat-panel X-ray detector operating in portrait orientation. Its nominal FOV is 16 cm (transaxial) x 20 cm (axial).
Both components are mounted on a computer-controlled rotating gantry, along with a three-axis rotating arm to hold a biopsy gun – enabling the new system to perform image-guided biopsies of suspicious lesions. The scanner elements are mounted on linear slides so that their distance from the centre-of-rotation can be adjusted according to the desired mode of operation (PET or CT) and breast size. Specialized software registers the CT and PET images, harmonizing image voxel sizes and aligning the images. CT images are segmented for use with the Compton scatter and attenuation corrections utilized in PET image reconstruction.
Performance testing
The team reported that the PET/CT scanner performed as expected during testing with the NEMA NU2-2009 protocol. The system exhibited a spatial resolution of 2.2 mm (using filtered-backprojection reconstruction) 5 mm from the centre of the scanner. Imaging of a micro-hot-rod phantom illustrated the potential utility of the dual-modality images.
Raylman told medicalphysicsweb that the tests demonstrated the systems’ high performance, although some improvements are planned. The team is now working to reduce CT detector binning and implement an enhanced iterative reconstruction algorithm. These modifications should improve CT resolution and permit reduction of radiation dose to the breast by reducing X-ray flux, without degrading image quality. They have also developed a patient bed that incorporates shielding to block the PET detectors from annihilation photons produced by radiotracer in the patient’s organs and minimize scatter of X-rays into the patient’s torso.
When these improvements are made, the researchers will perform a phantom study to estimate lesion detectability with the PET and CT components, and measure radiation dose to the breast and torso. Raylman said that the team hopes to start pre-clinical testing within six months.
PET/CT advantages
“The addition of CT to dedicated breast-PET scanners makes it possible to supplement PET images by detecting lesions that may not preferentially uptake the PET radiotracer used in the study,” said Raylman. “Future CT scanners may be able to detect some micro-calcifications, which are not visible with PET. Finally, the combination of CT with PET enables accurate attenuation and scatter correction methods to be applied to PET images; potentially enhancing lesion detectability and facilitating accurate quantification of radiotracer uptake in the breast and breast lesions.”
Raylman predicts that radiation dose to the breast from the PET/CT scanner may be lower than that from a 3D tomosynthesis digital mammogram. Meanwhile, the anticipated exam time of 12 minutes should be much faster than the time required to perform an equivalent breast-MRI exam. The breast PET/CT exam should also be less expensive than a comparable MRI scan. Finally, breast PET/CT scanners could also be more accessible to patients than breast-MRI scanners, because they will be less expensive to purchase and far less challenging to install in smaller imaging centres and radiology departments.
“Being able to perform a PET-guided biopsy will also be advantageous to confirm diagnoses of suspicious lesions. Our system is currently the only dedicated breast-PET/CT scanner capable of performing image-guided biopsy,” Raylman added.
This system is one of two under development in the USA. Another initiative to develop a dedicated breast PET/CT scanner has been underway at UC Davis Medical Center since the mid-2000s.
