12 P. Perini et al.a bFig. 2.1 3D-VR (a) and MIP (b) reconstructions of a descending thoracic aortic aneurysmFig. 2.2 Practical steps for the generation of a “stretched” aorta usingcentreline technique. A centreline of ﬂow (green line; left image) isgenerated by the workstation. Before a stretched reconstruction isobtained (right image), the centreline can be modiﬁed by adding,retrieving, or moving the numerous dots of the line (middle image)
132 Advanced Computed Tomography Imaging, Workstations, and Planning ToolsNowthat3Dworkstationsareintuitiveand“user-friendly,”they are accessible to cardiovascular surgeons and not onlyexperienced radiologists. Reliable default VR and MIP tem-plates and quick access to advanced segmentation algorithmsthat automatically edit and grow vessel territories are essen-tial (Fig. 2.3).Vessel caliber, patency, tortuosity, and burden of calciumand thrombus are important vascular features to assesspreoperatively (Fig. 2.4).Diameters, lengths, and angles are often necessary dimen-sions to measure (Fig. 2.5).Although much of this imaging information can be visu-alized on conventional axial images, 3D-VR, multiplanar,and curved planar reconstructions provide quick and clearvisualization of the complex relationships of anatomy andpathology (Fig. 2.6).The combined use of the various visualization techniquesis critical in surgical planning.The advantage of VR is the accurate spatial perceptionthrough a complete 3D angiographic overview. Care, how-ever, must be taken in interpreting these reformatted images.For example, a critical stenosis may appear like a completeFig. 2.3 3D-VR of a juxtarenal abdominal aortic aneurysm treatedwith a fenestrated endograftFig.2.4 CPR of a renal artery used for planning a fenestrated endograftdepicting a severe stenosis at the origin of the vesselFig. 2.5 Sagittal MPR showing the superior mesenteric artery and itsangle with the aorta. The catheterization of this vessel via a femoralapproach during fenestrated endovascular aneurysm repair is antici-pated to be challenging (angle of the target vessel to the aorticwall<60°)
14 P. Perini et al.occlusion. It is important to correlate 3D ﬁndings with corre-sponding 2D images to avoid such pitfalls. With MPR, a 2Danalysis through the original dataset is performed in axial,coronal, sagittal, or oblique orientations. Analysis of the ves-sel wall and the ﬂow lumen with accurate display of stenosis,occlusions, and calciﬁcation can be performed. The only dis-advantage is the limited spatial display. MIP is also a 2Danalysis option for an angiographic overview, butsemiautomated or complete manual editing is required toremove structural overlay. It can be useful to depict smallcaliber vessels and poorly enhanced vessels. Its accuracy is,however, limited in calciﬁed vessels. Conﬁrmation of stenosisand vessel caliber measurements should always be done withorthogonal MPR. As vessels curve in and out of the planes,standard MPRs cannot display an entire vascular territory andﬂow lumen in one image. To obtain a complete longitudinalvessel display, the solution is to generate a longitudinal cross-section using either 2D or rotating CPR techniques.Themeasurementsrequiredforaccurateplanningofbranchedand fenestrated endografts are complex and beyond the scope ofwhat can be achieved accurately with standard 2D axial imagesand table positions to measure aortic lengths and the relativepositions of visceral arteries. Indeed, there is signiﬁcant potentialfor error when trying to measure aortic lengths using a combina-tion of coronal and sagittal images of the angulated aorta. Theevolutionofmodernworkstationshasconsignedthesedifﬁcultiesto history with rapid generation of accurate 3D images now fea-sible in real-time (Figs. 2.7 and 2.8).It is likely that some of the ongoing improvements in clini-cal outcomes that are continuously being reported in the endo-vascular literature are in part attributable to more accurate graftdesign, with consequent beneﬁts in terms of improved targetvessel perfusion rates, less graft migration/endoleak, andshorter procedure times. Clearly, in striving to improve clinicaloutcomes, it is incumbent on all endovascular surgeons tobecome comfortable with this remarkable technology.a bFig. 2.6 3D-VR (a) andstretched CPR (b) of aninfrarenal aortic aneurysmassociated with a right commoniliac aneurysm. The combinationof the various visualizationtechniques is vital to properlysize and plan the endograft
152 Advanced Computed Tomography Imaging, Workstations, and Planning ToolsFig. 2.7 Various phases of the planning of a fenestrated endograft usinga workstation. On the upper left, the 3D-VR is generated. The centreline(in green) is used to generate the stretched CPR (right). The 2D image(lower left) is the reconstruction perpendicular to the centre-lumen line.It is used to precisely assess the diameter of the aorta. AAA abdominalaortic aneurysm, CIA common iliac artery, RA renal artery, SMA superiormesenteric artery
16 P. Perini et al.Fig. 2.8 Postoperative 3D-VR of a type II thoracoabdominal aneurysmtreated with a four-branch endograft (right). A meticulous analysis ofthe preoperative CT scan on the workstation was mandatory to designan endograft that perfectly matched the aortic anatomy (left). The lengthand diameter of the sealing zone in each target vessel has also beenevaluated. LRA left renal artery, RRA right renal artery, SMA superiormesenteric artery CT celiac trunk?