Abstract
To the Editors:
We read with great interest the recent article by Spanos et al 1 investigating the anatomical differences between intact and ruptured abdominal aortic aneurysms (AAA). In this retrospective study, 31 patients with elective AAA repair were compared with 31 patients matched for age and smoking who had a ruptured aneurysm. Preoperative computed tomography angiography scans were analyzed to identify the prognostic value of anatomical characteristics as predictors of rupture. Interestingly, the authors highlighted significant differences between intact and ruptured aneurysms regarding the infrarenal aortic neck diameters, the neck length and calcification, the common iliac artery (CIA) lengths, the left CIA and external iliac artery diameters, as well as the total and true lumen aneurysm volumes. Receiver operating characteristics curve analysis revealed that left CIA length as well as total aneurysm volume were predictors of rupture [area under the curve 0.70 (p<0.035) and 0.68 (p<0.042), respectively]. Surprisingly, maximum diameter was not a predictor of rupture (0.62, p=0.151). Although the maximum aortic diameter is a well-established indication for AAA surgical treatment, 2 this study highlights the real need to develop new tools to better evaluate the risk of rupture in patients with AAA.
Imaging is a key step in the diagnosis and the evaluation of the prognosis of patients. We acknowledge the work of the authors to provide a detailed anatomical characterization of the AAA using semiautomatic software. Such an approach requires the training and expertise of the operators, is tedious and time-consuming, and is subjected to interobserver reproducibility. The authors limited this bias by having 2 investigators perform the measurements, using a third observer to resolve any disagreement. In addition, they checked the interobserver variability according to the Bland and Altman method.
The automated segmentation method would be of interest to standardize and facilitate AAA morphological characterization. With this aim, we recently developed a fully automated pipeline on computed tomography images that enables automatic analysis of AAA characteristics, including measurements of the diameters, lengths, and volumes of the aneurysm and the vessels, as well as the detection and quantification of calcifications and intraluminal thrombus. 3 Such an approach may have potential applications in both clinical practice and medical research by facilitating image analysis and quantitative measurements.
Many advances in imaging technology are to be expected within the next few years, and as Domanin and Trimarchi 4 nicely commented, this will help vascular surgeons improve preoperative planning and sizing, evaluate the prognosis of patients and the stability of aortic devices, and anticipate postoperative outcomes.
Artificial intelligence has brought new insights in medicine, with a wide range of potential applications for the diagnosis, prognosis, or treatment of patients. 5 Machine learning algorithms are currently being developed to better assess the prognosis of patients with AAA, including the risk of AAA growth or rupture.6,7 Further research should be oriented toward improving imaging analysis. Combining a detailed anatomical characterization of the AAA with the clinical and biological characteristics of patients would allow development of multivariable scores to better evaluate the risk of rupture in patients with AAA. This is a very active area, and we truly believe that major advances are to be expected within the next decade.
