Multi-Energy Low-Kiloelectron Volt versus Single-Energy Low-Kilovolt Images for Endoleak Detection at CT Angiography of the Aorta

Abstract

Purpose To compare image quality, diagnostic performance, and conspicuity between single-energy and multi-energy images for endoleak detection at CT angiography (CTA) after endovascular aortic repair (EVAR). Materials and Methods In this single-center prospective randomized controlled trial, individuals undergoing CTA after EVAR between August 2020 and May 2022 were allocated to imaging using either low-kilovolt single-energy images (SEI; 80 kV, group A) or low-kiloelectron volt virtual monoenergetic images (VMI) at 40 and 50 keV from multi-energy CT (80/Sn150 kV, group B). Scan protocols were dose matched (volume CT dose index: mean, 4.5 mGy ± 1.8 [SD] vs 4.7 mGy ± 1.3, P = .41). Contrast-to-noise ratio (CNR) was measured. Two expert radiologists established the reference standard for the presence of endoleaks. Detection and conspicuity of endoleaks and subjective image quality were assessed by two different blinded radiologists. Interreader agreement was calculated. Nonparametric statistical tests were used. Results A total of 125 participants (mean age, 76 years ± 8; 103 men) were allocated to groups A (n = 64) and B (n = 61). CNR was significantly lower for 40-keV VMI (mean, 19.1; P = .048) and 50-keV VMI (mean, 16.8; P < .001) as compared with SEI (mean, 22.2). In total, 45 endoleaks were present (A: 23 vs B: 22). Sensitivity for endoleak detection was higher for SEI (82.6%, 19 of 23; P = .88) and 50-keV VMI (81.8%, 18 of 22; P = .90) as compared with 40-keV VMI (77.3%, 17 of 22). Specificity was comparable among groups (SEI: 92.7%, 38 of 41; both VMI energies: 92.3%, 35 of 38; P = .99), with an interreader agreement of 1. Conspicuity of endoleaks was comparable between SEI (median, 2.99) and VMI (both energies: median, 2.87; P = .04). Overall subjective image quality was rated significantly higher for SEI (median, 4 [IQR, 4-4) as compared with 40 and 50 keV (both energies: median, 4 [IQR, 3-4]; P < .001). Conclusion SEI demonstrated higher image quality and comparable diagnostic accuracy as compared with 50-keV VMI for endoleak detection at CTA after EVAR. Keywords: Aneurysms, CT, CT Angiography, Vascular, Aorta, Technology Assessment, Multidetector CT, Abdominal Aortic Aneurysms, Endoleaks, Perigraft Leak Supplemental material is available for this article. © RSNA, 2024.

Keywords: Abdominal Aortic Aneurysms; Aneurysms; Aorta; CT; CT Angiography; Endoleaks; Multidetector CT; Perigraft Leak; Technology Assessment; Vascular.

Figure 1:

Flowchart depicts participant inclusion.

Figure 2:

Box plots depict objective image parameters for low-kilovolt single-energy images and low–kiloelectron volt virtual monoenergetic images at 40 keV and 50 keV, reconstructed from multi-energy CT. The line in each box depicts the median. The lower and upper hinges correspond to the first and third quartiles, respectively. The lower and upper whiskers extend to the smallest and largest values, respectively, no further than 1.5 times the IQR from the hinge. Points outside the whiskers depict outlier values, further than 1.5 times the IQR from the hinge. CNR = contrast-to-noise ratio, CTDI_vol = volume CT dose index.

Figure 3:

Violin plots depict subjective image parameters assessed by reader 1 (R1) and reader 2 (R2) for low-kilovolt single-energy images and low–kiloelectron volt virtual monoenergetic images at 40 keV and 50 keV, reconstructed from multi-energy CT. The point in each violin depicts the median. The lower and upper hinges extend to the smallest and highest values, respectively. The width of each violin refers to the frequency of the given value.

Figure 4:

Exemplary images of endoleaks after endovascular repair of an aortic aneurysm. (A) Single-energy CT angiographic image in an 84-year-old male patient at 80 kV in the arterial phase shows endoleak type I (arrow). (B, C) Multi-energy CT angiographic images in a 72-year-old male patient in the arterial phase demonstrate endoleak type I (arrow) at (B) 40 keV and at (C) 50 keV. (D) Single-energy CT angiographic image at 80 kV in an 80-year-old male patient in the arterial phase shows endoleak type II (arrow). (E, F) Multi-energy CT angiographic images in a 75-year-old male patient in the arterial phase demonstrate endoleak type II (arrow) at (E) 40 keV and at (F) 50 keV.

Figure 5:

Exemplary images of endoleaks after endovascular repair of an aortic aneurysm. (A) Single-energy CT angiographic image in a 77-year-old male patient at 80 kV in the arterial phase shows endoleak type II (arrow). (B, C) Multi-energy CT angiographic images in a 77-year-old male patient in the arterial phase demonstrate endoleak type II (arrows) at (B) 40 keV and at (C) 50 keV. Please note the highest CT attenuation but also the highest image noise at 40 keV, resulting in lower subjective image quality. Noise is given as the SD of CT attenuation in the psoas muscle. Orange circular outline indicates the measured region of interest. CNR = contrast-to-noise ratio.