Advantages of 70-kV CT Angiography for the Visualization of the Adamkiewicz Artery: Comparison with 120-kV Imaging

Ethics

Our institutional review board of Kobe University Hospital approved this retrospective study; the requirement for written, informed consent was waived because of its design. All patient records and information were anonymized before analysis.

Study Population

A total of 185 consecutive subjects who underwent CTA to detect the AKA between January 2014 and September 2016 were enrolled in this study. Subjects who had a history of aortic repair, spinal arteriovenous fistula, or vertebral or spinal tumor were excluded. Ultimately, we included 120 patients (median age, 69 years; range, 28–85 years; 25.8% women) who underwent CTA to detect the AKA before aortic repair. Because we changed scan mode in January 2015 when a new CT scanner was installed, CTA was performed with a 120-kV scanner in the first 60 subjects and a 70-kV scanner in next 60 subjects.

Data Acquisition

Seventy-kilovolt CTA was performed on a 192-section dual-source CT scanner (Somatom Force; Siemens, Erlangen, Germany). Dual-power scan mode was applied with the following parameters: 128 × 0.6 mm detector configuration, 0.5 seconds per rotation, 0.45 pitch, 1170-mA tube current, and 1300-mAs effective tube current–time product. In dual-power scan mode, the total radiation is divided equally between each x-ray tube, and the data from each detector are summed.¹⁷ Thus, the maximum photon flux is expected to double, and because of the reduction in load on the x-ray tube, we achieved high tube current with a long scan range. One hundred twenty-kilovolt CTA was performed on a 64- or a 320-section multidetector CT scanner (Aquilion 64 or Aquilion ONE; Toshiba Medical Systems, Tokyo, Japan). The scanning parameters used were like those in previous reports^8,12,18 and were as follows: 64-section helical scan mode, 64 × 0.5 mm detector configuration, 0.5 seconds per rotation, 0.64 pitch, 400-mA tube current, and 312.5-mAs effective tube current–time product. The scan range extended from the thoracic inlet to the lesser trochanter of the hip.

Injection Protocol and Scan Timing

Iopamidol 370 mg I/mL (Iopamiron 370; Bayer Yakuhin, Osaka, Japan) was administered through a 20-ga intravenous catheter placed in the right antecubital vein at 5 mL/s, followed by a 30-mL physiologic saline flush at the same rate using a dual-head power injector (Dual Shot GX 7; Nemoto Kyorindo, Tokyo, Japan). Scan timing was set as the peak timing in the true lumen of the descending aorta, which was determined by the timing-bolus method in 70-kV CTA, and by the bolus-tracking method in 120-kV CTA. This technical limitation meant that the ROI could not be moved simultaneously during acquisition of the time-intensity curve in the ROI in the 192-section dual-source CT scanner. For the timing-bolus method, 10 mL of contrast material followed by a 20-mL physiologic saline flush was injected at 5 mL/s, respectively. Timing-bolus scanning was started 10 seconds after the start of injection. The peak time of the CT number in the ROI placed in the true lumen of the descending aorta at the level of the tenth thoracic vertebra was obtained with console software (DynEva, syngo VA50A; Siemens). The 70-kV scan was then started 10 seconds after the peak time. In the bolus-tracking method, the ROI was placed in the descending aorta at the level of the sixth thoracic vertebra, and the trigger threshold was set at 200 HU. The 120-kV scan was automatically started 5 seconds after the trigger. The volume of contrast material used for CTA in the 70- and 120-kV scans was 90 and 100 mL, respectively. The scans were repeated to obtain images in the early and late arterial phases. The scan of the late arterial phase was obtained approximately 10 seconds after the end of the first phase, to evaluate the washout of contrast material in the vessels to distinguish the AKA from the similar hairpin-shaped radiculomedullary vein,⁷ because vessels that are more prominent in the late phase are likely to be veins.

Image Reconstruction

Seventy-kilovolt CTA image data were reconstructed in the axial plane with 0.6-mm thickness, 0.6-mm interval, 200-mm FOV, and a vascular convolution kernel (Bv44) with Advanced Modeled Iterative Reconstruction (Siemens) with strength level 3. One hundred twenty-kilovolt CTA image data were reconstructed in the axial plane with 0.5-mm thickness, 0.5-mm reconstruction interval, and 200-mm FOV using a soft-tissue convolution kernel (FC13) with the filter back-projection method. The assignable values of both protocols were set as the parameters of reconstruction. Image data were then transferred to a workstation (Ziostation 2, Version 2.4.2.3; Ziosoft, Tokyo, Japan) for further processing and analysis. Curved coronal multiplanar reformatted images (1.0-mm thickness at 1.0-mm intervals) fitted to the curvature of the spinal cord in the early and late arterial phases were obtained.¹⁹

Radiation Dose Metrics

The volume CT dose index (CTDI_vol) and the dose-length product reported by the scanner after imaging of a 32-cm phantom were recorded. The size-specific dose estimate (SSDE) was calculated by multiplying the CTDI_vol by a conversion factor based on the effective diameter,²⁰ which was measured using the lateral diameter of the upper abdomen on the standard posterior-anterior CT radiograph.

Qualitative Image Analysis

To assess performance for identifying the AKA, we initially evaluated the image datasets with the 2 scan modes in a consensus reading by 2 board-certified diagnostic radiologists (T.N. with 10 years of experience and A.K.K. with 15 years of experience), who were unaware of the identity of subjects and CTA protocols. The AKA was identified with a stepwise approach and a 6-point scoring system (Fig 1), modified from a previously published report.⁸ First, we sought the “hairpin vessel” and assessed it as detectable or undetectable (score 1 or 0, respectively). Second, we assessed continuity to the aorta as definite, conceivable, equivocal, or undetectable (score 3, 2, 1, or 0, respectively). We carefully evaluated the presence of the collateral pathway from the aorta. Third, we evaluated the washout of the attenuation of the vessel from the early-to-late arterial phase (score 1 or 0). Finally, we summed the scores and defined positive identification of the vessel as the AKA when the total score was ≥3. Furthermore, the artery connecting the aorta to the AKA was defined as the CSA. The origin level and the side of the AKA and the CSA were recorded.

• Download figure
• Open in new tab
• Download powerpoint

Fig 1.

The AKA was identified according to the following stepwise approach: 1) detect the hairpin vessel; 2) assess the continuity of the vessel to the aorta mainly around the pedicle of the vertebral arch (white arrows) as definite (a), conceivable (b), equivocal (c), or undetectable; 3) check the washout from the early-to-late arterial phase (white arrowhead), because vessels that are more prominent in the late arterial phase are likely to be veins (black arrowhead); and 4) sum the scores. We defined positive identification of the AKA as a total score of ≥3.

The results of the consensus image interpretation by 2 board-certified diagnostic radiologists were determined to be the reference standard for following analyses. To assess the reproducibility of CSA identification in both imaging protocols, 2 radiologists who were unaware of the identity of subjects and CTA protocols evaluated images independently (observer A, T.N., ≥2 months after the first consensus reading, which was judged to be acceptable to minimize the influence of the first review; observer B, S.S., with 6 years of experience).

Quantitative Image Analysis

For the quantitative image analyses, a cardiovascular radiologist (S.S., observer B) unaware of the scan protocols evaluated the following measurements: Mean CT number and SD of the aorta were calculated from the measurement with the circular ROIs placed at the fifth, eighth, and twelfth thoracic vertebrae and the third lumbar vertebra.¹⁸ The ROIs were drawn to include as much vessel lumina as possible, while avoiding areas of artifacts and calcification. Furthermore, for subjects in whom the AKA had been identified with the previous consensus reading, we performed the following measurement: Because it was difficult to measure the CT number of the AKA accurately due to its small diameter and tortuous route, the CT number of the anterior spinal artery (ASA) was used as an indicator.²¹ The mean CT number of the ASA was obtained from the mean of the measured CT number on a sagittal and a coronal multiplanar reformation adjusted to the ASA course along the length of 1 vertebral body (Fig 2). Furthermore, the mean CT number and SD of the spinal cord were calculated from measurements at 3 positions in the spinal cord in the axial plane (top, middle, and bottom level of the previously set ROI for the ASA, Fig 2). The mean CT number of the bone cortex of the vertebral body was calculated from the measurement of the cortex around the right and left intervertebral foramina at the AKA running level (Fig 2). The SNR and CNRs were calculated with the following equations:

• Download figure
• Open in new tab
• Download powerpoint

Fig 2.

The ROIs placed for measuring CT numbers are shown in a coronal image of the ASA (A), a sagittal image of the ASA (B), an axial image of the spinal cord (C), and a sagittal image of the vertebral bone cortex around the vertebral foramen at the AKA (white arrow) running level (D).

Statistical Analysis

Continuous variables are expressed as the mean ± SD or median and interquartile range (IQR). Categoric variables are expressed as proportions. To evaluate differences between 70- and 120-kV CTA, we used the Welch t test for body mass index, mean CT number of the aorta, SNR_Aorta, mean CT number of the ASA, SD of the spinal cord, CNR_ASA-Cord, and CNR_ASA-Bone. The Wilcoxon test was used to compare age, CTDI_vol, SSDE, and dose-length product. The χ² test was used to evaluate differences in AKA detection rates, CSA location, presence of the collateral pathway, aortic disease (aneurysm versus dissection), and sex. The Cochran-Armitage test was used to evaluate differences between the scores.

Diagnostic accuracy and the Cohen κ for agreement were evaluated by comparing the side and level of the CSA interpreted by 2 independent observers against the reference standard. We also assessed the effect of the presence of the collateral pathway to the AKA on diagnostic accuracy.

For statistical analysis, commercially available (JMP 13.0; SAS Institute, Cary, North Carolina) and open-source (R statistical and computing software; http://www.r-project.org/) programs were used. For all analyses, P < .05 indicated statistical significance.