Objectives: The objectives of this study were to analyze the spatial resolution of different reconstruction kernels and acquisition protocols, including a prototypic high-resolution protocol in flat-panel (FP) and multidetector (MD) computed tomography (CT), and to evaluate contrast and artificial cartilage depiction quality of in vitro FPCT and MDCT arthrography.
Materials and methods: An image-quality cone beam phantom was used to compare resolution and different reconstruction kernels of the standard MDCT (120 and 80 kV) and the standard binned (2 × 2) and prototypic high-resolution unbinned (1 × 1) FPCT protocols (5- and 20-second runs each). With the resulting FPCT kernel best matching the standard MDCT kernel (U90u), artificial joint phantoms with differently sized groups of cartilage defects (2, 1, 0.5, and 0.3 mm in width) were then scanned using intra-articular iodinated contrast at 50 mgI/mL. In these joint phantoms, CT numbers and noise in the iodinated contrast and artificial cartilage tissue were measured and contrast-to-noise ratios (CNR) were calculated. Depiction quality of artificial cartilage defects was qualitatively rated by 2 independent radiologists.
Results: A sharp reconstruction kernel for all FPCT protocols suited best for matched resolution to the standard MDCT kernel. High-resolution 20-second 1 × 1 binning FPCT showed comparable resolution with MDCT in the range of 0.4 to 1.6 line pairs (lp) per millimeter with superior resolution in higher frequencies than 1.6 lp per millimeter (P < 0.001). Flat-panel computed tomographic 5-second runs were associated with higher image noise than the 20-second runs were. The CNR differed significantly among the protocols (P < 0.01) and was the highest in the 20-second FPCT, followed by the 5-second FPCT 2 × 2 and MDCT protocols. Interreader agreement for the depiction quality of artificial cartilage defects was substantial and high in the joint phantoms (0.74 and 0.81, respectively; P < 0.001). The best ratings of the artificial cartilage defect depiction quality were seen in the FPCT 20-second, followed by the FPCT 5-second and MDCT acquisitions. The depiction quality of smaller cartilage defects (1.0 and 1.67 lp per millimeter) was rated worst in the MDCT acquisitions.
Conclusions: In vitro FPCT arthrography offers superior CNR and artificial cartilage defect depiction quality to MDCT, and spatial resolution for small structures is higher when applying high-resolution acquisition protocols. Flat-panel computed tomography, thus, has the potential to improve workflow, and tailored high-resolution protocols may allow for advanced cartilage evaluation in CT arthrography.