Performance of high-resolution CT for detection and discrimination tasks related to stenotic lesions - A phantom study using model observers

Abstract

Background: Accurate detection and grading of atheromatous stenotic lesions within the cardiac, renal, and intracranial vasculature is imperative for early recognition of disease and guiding treatment strategies.

Purpose: In this work, a stenotic lesion phantom was used to compare high resolution and normal resolution modes on the same CT scanner in terms of detection and size discrimination performance.

Materials and methods: The phantom is comprised of three acrylic cylinders (each 15.0 cm in diameter and 1.3 cm thick) with a matching array of holes in each module. The outer two modules contain holes that are slightly larger than the corresponding hole in the central module to simulate stenotic narrowing in vasculature. The stack of modules was submerged in an iodine solution simulating contrast-enhanced stenotic lesions with a range of lumen diameters (1.32-10.08 mm) and stenosis severity (0%, 50%, 60%, 70%, and 80%). The phantom was imaged on the Canon Aquilion Precision high-resolution CT scanner in high-resolution (HR) mode (0.25 mm × 0.50 mm detector element size) and normal-resolution (NR) mode (0.50 mm × 0.50 mm) using 120 kV and two dose levels (14 and 21 mGy SSDE) with 30 repeat scans acquired for each combination_. Filtered back-projection (FBP) and a hybrid-iterative reconstruction (AIDR) were used with the FC18 kernel, as well as a deep learning algorithm (AiCE) which is only available for HR. A non-prewhitening model observer with an eye filter was implemented to quantify performance for detection and size discrimination tasks in the axial plane.

Results: Detection performance improved with increasing diameter, dose, and for AIDR in comparison to FBP for a fixed resolution mode. Performance in the HR mode was generally higher than NR for the smaller lumen diameters (1-5 mm) with decreasing differences as the diameter increased. Performance in NR mode surpassed HR mode for lumen diameters greater than ∼4 mm and ∼5 mm for 14 mGy and 21 mGy, respectively. AiCE provided consistently higher detection performance compared with AIDR-FC18 (48% higher for a 6 mm lumen diameter). Discrimination performance increased with increasing nominal diameter, dose, and for larger differences in stenosis severity. When comparing discrimination performance in HR to NR modes, the largest relative differences occur at the smallest nominal diameters and smallest differences in stenosis severity. The AiCE reconstruction algorithm produced the highest overall discrimination performance values, and these were significantly higher than AIDR-FC18 for nominal diameters of 7.14 and 10.08 mm.

Conclusions: HR mode outperforms NR for detection up to a specific diameter and the results improve with AiCE and for higher dose levels. For the task of size discrimination, HR mode consistently outperforms NR if AIDR-FC18 is used for dose levels of at least 21 mGy, and the results improve with AiCE and for the smallest differences in stenosis severity investigated (50% vs. 60%). High-resolution CT appears to be beneficial for detecting smaller simulated lumen diameters (<5 mm) and is generally advantageous for discrimination tasks related to stenotic lesions, which inherently contain information at higher frequencies, given the right reconstruction algorithm and dose level.

Figure 1.

Photographs of the three machined acrylic cylinders comprising the stenotic lesion phantom (a) and a schematic showing the indexing and relative locations of the array of holes at three different radial locations from the phantom center (R_A, R_B, and R_C). Table 1 details each hole diameter and simulated stenosis severity for the indexing showing in (b).

Figure 2.

Axial slices through the outer (top row) and central modules (central row) are shown along with a sagittal slice (bottom row) through the stenosis phantom. The images represent the ensemble average of 30 repeat scans (a) and a single scan (b) for FBP reconstruction with the FC18 kernel.

Figure 3.

Examples of the difference signal (a,c) and corresponding Fourier transform (b,d) are shown for the detection (a,b) and discrimination tasks (c,d). The lumen diameter is 7.14 mm for the detection task and the discrimination task represents the 50% stenosis (5.05 mm lumen diameter) and 70% stenosis (3.91 mm) for a nominal 7.14 mm diameter lumen.

Figure 4.

Detection performance ($d_{det}^{\prime }$) results are shown for FBP-FC18; 14 mGy SSDE (a) AIDR-FC18; 14 mGy (b), FBP-FC18; 21 mGy (c), and AIDR-FC18; 21 mGy (d). Estimations are plotted for NR mode (y-axis) and HR mode (x-axis) across all 84 lumen targets along with the line of identity (dashed line). The x and y error bars indicate the 95% confidence interval derived from bootstrapping across the 30 repeat scans.

Figure 5.

Detection performance ($d_{det}^{\prime }$) results shown as a ratio of HR to NR for lumens located at three different radial locations (R_A-C) from the central axis of the phantom. The linear fit was calculated for all data points except the 10.08 mm lumen diameter at the R_C radial location.

Figure 6.

Detection performance ${(d}_{det}^{\prime })$ plotted for HR mode at 21 mGy SSDE (a) and 14 mGy SSDE (b) for all 48 lumen targets (24 for outer disk and 24 for central disk) in the outer radial location (R_C).

Figure 7.

Detection performance ${(d}_{det}^{\prime })$ plotted for HR mode at 21 mGy SSDE for lumens located at the three different radial locations of 1.9 cm (R_A), 3.8 cm (R_B), and 6.4 cm (R_C) from the central axis of the phantom. The error bars indicate the 95% confidence interval derived from bootstrapping across the 30 repeat scans.

Figure 8.

Size discrimination ${(d}_{dis}^{\prime })$ results are plotted for HR mode (a-d) and NR mode (e-h) for FBP-FC18; 14 mGy SSDE (a,e) AIDR-FC18; 14 mGy (b,f), FBP-FC18; 21 mGy (c,g), and AIDR-FC18; 21 mGy (d,h) across all 24 lumen difference targets in the outer radial location (R_C). Resulted are plotted separately for 50% vs. 60% stenoses (50/60), 50/70, and 50/80%. The y-axis error bars indicate the 95% confidence interval derived from bootstrapping across the 30 repeat scans.

Figure 9.

Size discrimination ${(d}_{dis}^{\prime })$ results are plotted as a ratio of the fitted data (Figure 6) for HR to NR mode across all 24 lumen difference targets in the outer radial location (R_C).

Figure 10.

Size discrimination ${(d}_{dis}^{\prime })$ results are plotted for HR mode at 21 mGy SSDE (a,b) and 14 mGy SSDE (c,d) across various stenosis differences in the outer radial location (R_C). The error bars indicate the 95% confidence interval derived from bootstrapping across the 30 repeat scans.

Figure 11.

Discrimination performance ${(d}_{dis}^{\prime })$ plotted for HR mode at 21 mGy SSDE for lumens located at the three different radial locations of 1.9 cm (R_A), 3.8 cm (R_B), and 6.4 cm (R_C) from the central axis of the phantom. The error bars indicate the 95% confidence interval derived from bootstrapping across the 30 repeat scans.