Imaging of Cardiovascular Implantable Electronic Device Leads With Photon-Counting Detector Computed Tomography

Computed tomography (CT) imaging of cardiovascular implantable electronic device (CIED) leads is currently hampered by large metal artifacts. Recently, photon-counting detector CT (PCD-CT) has been clinically introduced, offering high-resolution imaging with thin slice thicknesses and improved contrast-to-noise ratios. Suspected complications of CIED such as perforation, infection and venous obstruction could potentially be imaged with PCD-CT if metal artifacts were effectively reduced through adapted scan protocols and iterative metal artifact reduction algorithms (IMAR). The study evaluated the impact of various scan and reconstruction parameters, including different IMAR settings, on CIED lead visualization with PCD-CT in order to develop an optimized scan and reconstruction protocol for imaging leads.

Five different CIED leads were evaluated in a nonbeating heart phantom using a dual-source PCD-CT with electrocardiography-gated spectral standard resolution (collimation 144 × 0.4 mm) and nonspectral ultra-high resolution (UHR) mode (collimation 120 × 0.2 mm) spiral data acquisition. One scan was performed for each lead and each acquisition mode. Images were reconstructed with different slice thicknesses (0.2 mm, 0.4 mm, 0.6 mm), convolution kernels (Bv40, Bv44, Bv48, and Bv56), virtual monoenergetic energy levels (60-140 keV in steps of 10 keV), without and with different IMAR settings. The extent of metal artifacts was objectively evaluated using 4 different parameters. Additionally, 3 observers subjectively assessed image quality using a 5-point scale.

Metal artifacts increased with sharper kernels and higher keV levels in virtual monoenergetic reconstructions. The artifacts were not dependent on slice thickness. No significant differences in metal artifacts were observed between UHR and standard-resolution scans when using similar reconstruction parameters. IMAR effectively reduced artifacts across all kernels, slice thicknesses, and keV levels, with the "neuro coils" setting showing the best performance. Subjective analysis of image quality revealed that thinnest slices and sharpest kernels (0.2 mm, Bv56) allowed for better delineation of fine structures, such as the shape of helices, while reconstructions with thicker slices and softer kernels (0.6 mm, Bv40) were preferred for visualizing general lead appearance and adjacent anatomical structures.

Ultra-high and standard resolution PCD-CT with IMAR enables good-quality imaging of CIED leads, showing even small details without compromising the visibility of nearby structures. A dedicated acquisition and reconstruction protocol comprising an UHR scan with 2 reconstructions (0.2 mm/Bv56 and 0.6 mm/Bv40, using IMAR) appears optimal for PCD-CT imaging of CIED leads.