Image reconstruction can dramatically reduce the required electrical current and drop chest CT-induced radiation to as low as 3.5 mGy while maintaining acceptable image quality, according to a study published in the May issue of Radiology.
As scores of studies examine methods for reducing patients’ soaring exposures to radiation, researchers have tended to focus on implementing more appropriate imaging or retrospectively comparing image quality at varying tube currents. In a recent study conducted at Massachusetts General Hospital (MGH) in Boston, MGH radiologist Sarabjeet Singh, MBBS, MMST, and colleagues prospectively imaged patients using varying chest CT protocols in an effort to determine the capacity of iterative reconstruction to sustain image quality as current and radiation levels were lowered.
Twenty-five patients undergoing contrast-enhanced chest CT exams, primarily for suspected, staging or restaging of malignant lesions, agreed to undergo an additional four sets of radiation-inducing images at 150, 110, 75 and 40 mAs. All images were acquired on a 64-row multidetector CT scanner with unchanged parameters (except for tube current).
Following acquisition, two thoracic radiologists reviewed all images, which were presented according to randomized sequences of current. All four image datasets were reconstructed using filtered back projection (FBP) and adaptive statistical iterative reconstruction (ASIR) methods. ASIR reconstructions were performed at 30 percent, 50 percent and 70 percent blending levels (ASIR 30, ASIR 50 and ASIR 70, respectively).
ASIR resulted in significantly reduced image noise at lower tube currents than FBP, with a significant majority of FBP image sets graded as subjectively unacceptable at 40 mAs. At the same time, no ASIR 50 and ASIR 70 image sets were labeled as unacceptable at 40 mAs, while four of 23 ASIR 30 image sets were considered to be of unacceptable quality at 40 mAs.
Although no lesions were missed on either FBP or ASIR sets, three of 23 FBP sets at 110 mAs and 5 of 23 FBP sets at 75 mAs were subjectively graded as unacceptable. These figures compared with only four of 207 sets graded as unacceptable for all blendings of ASIR at 110, 75 and 40 mAs.
The CT dose index volumes reported by the authors were 12.6, 8.3, 5.8 and 3.5 mGy for tube currents of 150, 110, 75 and 40 mAs, respectively. “We found that ASIR is associated with substantially reduced subjective and objective image noise at a lower radiation dose of 40 mAs, or 3.5 mGy, for chest CT,” Singh and colleagues wrote.
“Although FBP technique at 40 to 75 mAs maintained diagnostic confidence and lesion detectability, there was an unacceptable increase in image noise,” the authors continued. Objective noise was likewise measured to be lower at all blendings of ASIR when compared with FBP.
Singh and colleagues considered several limitations to their study, including the small sample size and the use of only one vendor’s ASIR, which they acknowledged to be “vendor specific.”
“Practical implications of our study include substantial radiation dose reduction with the ASIR technique down to 40 mAs for chest CT examination. Both ASIR and FBP techniques allow substantial reduction in radiation doses associated with chest CT,” the authors concluded.