Computed Tomography-Applied Software-Guided Early Detection of Junctional Separation Between Aortic Endograft Components

Late junctional separation between fenestrated aortic endograft components resulting in type IIIa endoleak and sac repressurization is an uncommon but serious complication. Despite close surveillance imaging, the early signs of component separation can be easily missed. We sought to develop an automated method of detecting component separation after endovascular aortic repair (EVAR). We identified a case of complete fenestrated endograft component separation resulting in type IIIa endoleaks and subsequent aortic sac enlargement. We retrospectively reviewed the surveillance images using a dedicated workstation with segmentation software to generate a centerline from the supraceliac aorta to the bilateral iliac limbs of the endograft. For junctional separation detection, the distance from the lowest renal fenestration to the flow divider of the bifurcated distal body endograft was measured at each surveillance interval (Fig 1). Differences between the baseline and subsequent computed tomography scans were suggestive of component separation. This manual process was then implemented into a software algorithm to automate the measurement. The software generated measurements (Fig 2) were then compared with the manual measurements (Table). The manually measured distances between the lowest renal fenestration to the distal body flow divider (LRF-DFD) had increased continuously throughout the surveillance period (Table). In contrast, sac shrinkage was observed with no evidence of endoleak until complete separation of the components had occurred (Fig 1). The manual measurement scheme for the LRF-DFD distance was then coded into segmentation software (Nurea, Bordeaux, France) for automated detection of component separation. Analysis using the automated scheme correlated well with manual assessments (Table). Main body component separation in complex EVARs can be predicted by the LRF-DFD distance. Automation of this measurement can be accomplished quickly with dedicated segmentation software, providing a simple and standardized method for evaluating modular components at risk of junctional separation. Ongoing development for an automated evaluation of other issues occurring after complex EVAR is underway.Fig 2Automated software measurements.View Large Image Figure ViewerDownload Hi-res image Download (PPT)TableManual measurements vs automated software measurementsVariableSurveillance year2013201420152019Measurement methodManualSoftware basedManualSoftware basedManualSoftware basedManualSoftware basedDistance, mm84.278.6991.299.7111.1113.35167.4142.31Sac maximal diameter, mm63.961.363.258.657.355.891.786.7Sac volume, mm3NA290,560NA269,480NA252,090NA641,657NA, Not applicable. Open table in a new tab