Comparison of intravascular ultrasound- and centerline computed tomography-determined aortic diameters during thoracic endovascular aortic repair

Abstract

Background: Accurate sizing of stent grafts during thoracic endovascular aortic repair (TEVAR) is critical for a successful outcome. Centerline measurements using three-dimensional reconstruction of preoperative computed tomography angiography (CTA) is the current standard for stent graft sizing. However, this technique is predicated on an idealized straightened aorta and does not account for the variability in the aortic diameter during the cardiac cycle or the overall status of the patient's volume. Intravascular ultrasound (IVUS) offers real-time cross-sectional imaging of the aorta orthogonal to the support wire, thereby providing an adjunctive method for aortic diameter determination at the time of TEVAR.

Methods: A retrospective review was performed on all patients who underwent TEVAR for nontraumatic aortic pathology from July 2015 to December 2015. Preoperative CTA images were reconstructed on a dedicated three-dimensional workstation. CTA centerline aortic diameter measurements were performed in major and minor axes at 1-cm intervals from the left subclavian origin to 20 cm distally. The IVUS images were acquired intraoperatively through 1-cm stepwise pullback along the aorta from the left subclavian origin to 20 cm. IVUS aortic diameters were measured at the maximum phase during the cardiac cycle. The average values of major and minor axes diameters from both modalities were calculated at each location for comparison. Linear regression analysis was performed to evaluate correlation, and Bland-Altman plots assessed agreement between different imaging modalities.

Results: During the study period, 26 patients underwent TEVAR. Of these, 20 patients had adequate CTA and IVUS images, providing 355 paired measurements. There was a high correlation between CTA- and IVUS-determined aortic diameters (R = 0.62; P < .001). However, Bland-Altman analysis showed that, compared with CT, IVUS resulted in larger aortic diameters, with the mean difference of 3.09 mm. There was a significant variability between IVUS and CTA with the standard deviation of difference (SD diff) of 4.56 mm. When stratified by the aortic position, a high degree of agreement was observed at the base of the left subclavian (position 0), with a mean difference of -2.69 mm and an SD diff of 4 mm. The agreement was the lowest at the angulated aortic segments (2 cm to 7 cm distal to the subclavian origin) with a mean difference up to 7.96 mm and an SD diff up to 8.27 mm.

Conclusions: IVUS imaging and centerline CTA may provide significantly different aortic diameter measurements, particularly in angulated aortic segments. Caution must be taken when sizing a stent graft using CTA alone, particularly in an angulated proximal landing zone.