Powered with a new image chain, GE Healthcare’s Revolution Apex offers TrueFidelity for GSI, a first-of-its-kind Deep Learning Image Reconstruction engine to help transform image quality for dual-energy spectral CT. Utilizing a Deep Neural Network, the Deep Learning Image Reconstruction engine is trained to effectively differentiate and suppress noise in GSI projection data and produce dual energy TrueFidelity images with reduced image noise[1], preferred noise texture[2], enhanced contrast-noise-ratio[3] and low-contrast-detectability[4]. Revolution Apex represents the next generation of intelligent CT scanners in the GE Revolution family.
For more information, visit gehealthcare.com.
[1] Demonstrated in testing using the uniform section of the Catphan®600 with the CTP579 oval body annulus comparing pixel standard deviation in images reconstructed from the same raw data, at 0.625mm with DLIR-H and ASiR-V50%
[2] As demonstrated in a clinical evaluation consisting of 40 cases and 5 physicians, where each case was reconstructed with both DLIRfor GSI and ASiR-V and evaluated by 3of the physicians. In 88% of the reads, DLIR for GSI’s noise texture was rated better than ASiR-V’s.
[3] Demonstrated in testing using images of the CT ACR 464 Phantom (Gammex) and its 25 mm low contrast cylinderreconstructed from the same raw data with DLIR-L, DILR-M, and DLIR-H and ASiR-V 50%.
[4] Evaluated using the body MITA CT IQ Low Contrast Phantom (CCT189, the Phantom Laboratory) with the CTP579 oval body annulus and a model observer with images reconstructed from the same raw data with DLIR-H and ASiR-V 50%.