Evaluation of the quality of CT-like images obtained using a commercial flat panel detector system

PURPOSE: The development of flat panel detector technology has resulted in renewed interest in the possibility of generating CT-like images from rotational angiographic acquisitions. At least two commercial products now use cone beam reconstruction software in conjunction with flat panel detectors to produce such images. The purpose of the work presented here is to report on image quality obtained from one such system in objective and subjective terms and to compare it with the quality of images obtained from a modern multi-detector CT scanner. METHOD: The Image quality was assessed using a CATPHAN 500 model and an AAPM CT Performance Phantom model. Image noise, CT number accuracy, CT number consistency, Low Contrast Resolution, surface dose and Modulation Transfer Function were assessed for the flat panel detector and compared with results obtained from a 4 slice CT scanner. RESULTS: As expected image quality obtained from the CT scanner was much better than from the flat panel detector. Low contrast resolution was much worse and the surface dose was higher for the flat panel detector than the CT scanner. There was an inaccuracy in CT number determination and the noise was greater by a factor of two or three. Limiting resolution was better on images from the CT scanner. CONCLUSION: The poor low contrast resolution from flat panel detector was expected given the expected resolution of ±10 Hounsfield Units. These systems should not be considered as diagnostic CT scanners. However, the remaining performance figures indicate that the CT-like images obtained from this type of equipment are of sufficient quality for at least some clinical applications, such as detection of brain haemorrhages in the vascular suite.