Radiation dose can be reduced in pelvic computed tomography (CT) scans of total hip arthroplasty patients by using model-based iterative reconstruction (MBIR) algorithms and orthopedic metal artifact reduction software. However, such dose reduction is not appropriate for all diagnostic applications, according to a multi-specialty, multi-institutional team of researchers from The Netherlands.
The group evaluated the diagnostic quality of images acquired with up to 80% dose reduction and with different combinations of image processing software, and reported their findings in the November 2019 issue of Skeletal Radiology.
Led by Ruud H.H. Wellenberg, PhD, of the Department of Radiology at Isala Hospital in Zwolle, the team quantitatively and qualitatively assessed the quality of pelvic CT scans of patients with large-head, metal-on-metal total hip prostheses using Philips Healthcare’s MBIR and orthopedic metal artifact reduction (O-MAR) algorithms at five radiation dose levels. They compared these with images acquired using Philip’s image reconstruction algorithm iDose4 with O-MAR at full radiation dose.
Most CT vendors offer MBIR algorithms to help reduce CT radiation doses while improving image quality compared to standard filtered back-projection. These algorithms improve image reconstruction by incorporating system models and photon statistics. They can reduce the size of metal artifacts, helping to improve assessment of soft tissue surrounding implants and visibility of bone-metal interfaces. However, MBIR may eliminate small details and structures and add a smoothing appearance to images.
The additional use of orthopedic metal artifact reduction algorithms, also offered by CT scanner vendors, helps reduce severe metal artifacts caused by large metal implants. They post-process projection data and provide more regular attenuation profiles before image reconstruction. More regular attenuation profiles can improve performance of general iterative reconstruction algorithms.
The study focused on 76 patients undergoing routine 5- or 10-year follow-up pelvic scans during a 12-month period. After having a full-dose CT scan, they were randomly assigned to have a second CT with a radiation dose reduction of 20%, 40%, 57%, or 80%. All full-dose scans were reconstructed using the clinical practice level of noise suppression, iDose4’s Level 4. Reduced dose scans were reconstructed using both iDose4’s Level 4 and Philip’s MBIR algorithm at three levels. All images were also reconstructed with and without O-MAR.
CT numbers, noise or standard deviation, and contrast-to-noise ratios (CNRs) between muscle and fat were measured in the gluteus maximus muscle, fat, and bladder using a thin axial slice containing the largest diameter of the head of the metal-on-metal prosthesis with the most severe artifacts. Two experienced musculoskeletal radiologists independently evaluated seven aspects of each image set using a four-point Likert scale.
The CT numbers remained constant, noise decreased, and CNRs between muscle and fat increased in reduced-dose MBIR reconstructions compared with full dose iDose4 reconstructions for all three MBIR levels, regardless of which reduction level was used.
“By reducing metal artifacts, O-MAR reduced noise, improved CT number accuracy by bringing these values closer to expected CT numbers in the bladder, and improved CNR between muscle and fat,” they wrote.
However, the authors cautioned, “We recommend additionally evaluating conventional CT images without the use of metal artifact reduction software, because its algorithms may introduce secondary artifacts and could degrade the depiction of bone trabeculae and bone cortex.”
They also noted that the radiologists scored the quality of images reconstructed with iDose4 higher than the quality of reduced dose images reconstructed with MBIR. Images acquired with 80% dose reconstruction were not of diagnostic quality.
The authors stated that radiation dose levels could be reduced when evaluating the correct placement and status of prosthetic components in situ. “In high-contrast regions, a greater reduction of CT radiation dose may be applicable,” they wrote, adding that “most soft-tissue abnormalities can be assessed.”
But they recommend standard iterative reconstruction (iDose4) to evaluate delineation of bony structures and assessment of bone density, bone destruction, and sclerosis. “Care should be taken when using model-based iterative reconstruction algorithms in combination with reduced CT radiation dose in patients with total hip arthroplasties, especially in the evaluation of osseous structures and when focusing on small details and structures,” they concluded.