Iterative reconstruction (IR) makes it possible to reduce radiation dosage without a reduction in imaging quality in low-contrast CT examinations, but its effectiveness varies with different doses and vendors, according to study results published online May 19 in the journal Radiology.
As one of the most frequently utilized imaging modalities, it is estimated that 70 million CT scans are performed in the United States each year. For this reason, significant time and resources have been dedicated to searching for ways of reducing patient exposure to radiation resulting from these examinations. One such effort is the increasing use of IR, according to lead author Ganesh Saiprasad, PhD, of the National Institute of Standards and Technology in Boulder, Colo., and colleagues. “The use of IR has been heavily promoted by vendors to reduce radiation exposure in CT imaging acquisition,” wrote the authors. “The main advantage of IR over filtered back projection (FBP) is the ability to incorporate attenuation corrections and reduction in image noise.”
Saiprasad and his team of researchers set out to evaluate how the benefits of reduced radiation dosage through the use of IR varied between multiple vendors and doses. To do so, they performed a series of randomized repeat examinations on an American College of Radiology CT accreditation phantom (module 2, low contrast) using multiple radiation exposures, vendors and vendor IR algorithms. Scanners used included the GE Discovery CT750 HD, Siemens Somatom Definition Flash, and Philips Brilliance iCT 256. They then asked 11 volunteers to interpret 900 images and analyzed the results using statistical graphics and analysis of variance.
Their results showed that IR resulted in a mean correct classification rate that was significantly higher than FBP for all three vendors tested (blinded as A, B, and C) as well as for all three exposure levels (87 percent IR and 81 percent FBP at 20 mGy; 70 percent IR and 64 percent FBP at 12 mGy; and 61 percent IR and 56 percent FBP at 7.2 mGy). Vendor B’s CT scanner outperformed the other two vendors across all exposure levels with a mean correct classification rate of 74 percent compared to 68 percent for vendors A and C. Overall, readers had a higher mean correct classification rate for images produced using IR (73 percent than those using FBP (67 percent.)
“The magnitude of the IR improvement, which is dependent on the particular vendor and dose combination, ranged from 1 percent to 13 percent correct classification,” the authors wrote. “The variability among vendors in FBP performance was expected, and the relative amount of improvement with IR was particularly interesting.”
The researchers believe the significance of their findings lies in the potential of iterative reconstruction to maintain image quality at reduced dosages, as well as the fact that different vendors and different scanners produce varying results. “CT radiation exposure to patients can be potentially reduced by using IR instead of [filtered back projection] without compromising low-contrast detectability,” wrote Saiprasad et al. “There is substantial variability among the major vendors not only in baseline image quality by using FBP but also in the relative benefit of the use of IR.”