# Dual-layer spectral detector CT for improved assessment of right adrenal vein in patients with primary aldosteronism

## TL;DR

This study shows that using a new CT imaging technique improves the visibility of the right adrenal vein and accessory hepatic vein in patients with primary aldosteronism.

## Contribution

The study demonstrates that dual-layer spectral CT virtual-monoenergetic images can replace traditional multi-phase CT for adrenal vein assessment.

## Key findings

- Virtual-monoenergetic images at 40 keV provided the highest contrast-to-noise ratio for adrenal vein assessment.
- Using VMI40 in the portal venous phase achieved equivalent adrenal vein detection rates to late arterial phase polyenergetic images.
- VMI40 improved accessory hepatic vein assessability compared to conventional portal venous phase imaging.

## Abstract

To investigate whether virtual-monoenergetic images (VMI) from dual-layer spectral-detector computed tomography (DLCT) improve right adrenal vein (RAV) assessment before adrenal venous sampling (AVS), and portal-venous-phase (PVP) images alone can replace conventional multiphase CT combining late-arterial phase (LAP) and PVP.

Sixty-three patients with primary aldosteronism who underwent LAP and PVP DLCT before AVS were retrospectively analyzed. Conventional polyenergetic-images (PEI) and VMI at 40–70 keV (VMI40–70) were reconstructed. Image noise and contrast-to-noise ratio (CNR) of RAV were quantified. VMI with the highest CNR was predefined as the best-VMI. Two radiologists evaluated four image series (PEI and best-VMI at LAP and PVP) separately for RAV visibility, accessory hepatic vein (AHV) assessability, and image noise using five-point scales (1 = worst, 5 = best). RAV detection and AHV assessability rates were compared among image series.

In both enhancement phases, VMI40–70 showed lower image noise and VMI40–60 higher CNR than PEI (all p < 0.05), with VMI40 providing the highest CNR and being designated as the best VMI. Compared to PEI at LAP, CNR was lower with PEI at PVP but higher with VMI40 at PVP (both, p < 0.001). In subjective analysis, VMI40 showed better scores than PEI for image noise (LAP: 4.0 ± 0.5 vs. 2.4 ± 0.6; PVP: 4.0 ± 0.6 vs. 2.5 ± 0.5) and RAV visibility (LAP: 4.5 ± 0.8 vs. 3.5 ± 0.8; PVP: 3.7 ± 1.0 vs. 2.6 ± 0.9) (all p < 0.001). With PEI, the RAV detection rate was lower at PVP than at LAP (54.8% vs. 89.7%, p < 0.001). Use of VMI40 at PVP achieved the same detection rate as PEI at LAP (89.7%, p = 1.0) and higher AHV assessability rate than PEI at PVP (96.8% vs. 80.2%, p < 0.001).

DLCT-VMI improved RAV assessments before AVS. Using VMI40 for PVP provided RAV depiction equivalent to PEI at LAP and superior AHV assessability to PEI at PVP, suggesting its potential as a surrogate for dedicated multiphase CT.