Clinical Potential of CZT-Based Photon Counting Detector CT for Lung Cancer Imaging
摘要
The clinical research CZT-based Photon Counting Detector CT (PCDCT) (TSX-501R, Canon Medical Systems, Otawara, Japan) was developed based on the Canon Aquilion Precision gantry (TSX-304A Canon Medical Systems, Otawara, Japan), which is an ultra-high-resolution CT system. The clinical research CZT-based PCD-CT array populates the full 500 mm field of view in the fan angle and covers up to 40 mm in the z-direction at the isocenter. The minimum achievable focal spot size is as small as 0.4 mm × 0.5 mm. The clinical research CZT-based PCD-CT has two different collimation (readout) modes: normal resolution (NR mode, 64 × 0.6 mm) and super high-resolution mode (SHR mode, 192 × 0.2 mm). The NR mode comprises 3 × 3 macro pixels, whereas the SHR mode uses 1 × 1 macro pixels to obtain images with higher spatial resolution than those obtained using the NR mode. In addition, PCD-CT has six bins of data as energy information. The counting mode generates images based on events with photon energies >30 keV, whereas the spectral mode generates images using five closed-energy bins with threshold settings of 30/45/55/65/80 keV, respectively. CT data obtained with PCD-CT can be reconstructed using a combination of noise-reduction techniques developed to date, such as hybrid iterative reconstruction (hybrid-IR) and deep learning reconstruction (DLR). We provide quantitative evaluation of the noise and resolution properties of the clinical research CZT-based PCD-CT system with physical phantoms. In particular, we will compare the NR and SHR modes of the PCD-CT system with hybrid-IR and DLR reconstruction methods. Phantom results show that PCD-CT SHR mode improves spatial resolution compared to NR mode, where the resolution is further improved with DLR than with hybrid-IR. We provide some examples illustrating the clinical benefits of Canon’s clinical research CZT-based PCD-CT system, focusing on lung cancer imaging applications. For clinical evaluation, the patient scans were acquired with iodine contrast injection on both energy-integrating detector CT (EID-CT) and PCD-CT scanners. The PCD-CT clinical images are compared with EID-CT images acquired from the same patient. Clinical images show that PCD-CT NR image reduces noise as compared with that of EID-CT image. PCD-CT SHR images with DLR show preserved or improved spatial resolution with significantly reduced image noise as compared with EID-CT images.