Assessing abdominal aortic aneurysm growth using radiomic features of perivascular adipose tissue after endovascular repair

作者：Rui Lv, Ge Hu, Shenbo Zhang, Zhe Zhang, Jin Chen, Kefei Wang, Zhiwei Wang, Zhengyu Jin · 发表于：Insights into Imaging · 年份：2024 · DOI：10.1186/s13244-024-01804-7 · 被引用次数：11 · 研究领域：Aortic aneurysm repair treatments、Cardiovascular Disease and Adiposity、Renal and Vascular Pathologies

OBJECTIVES: The study aimed to investigate the relationship between the radiomic features of perivascular adipose tissue (PVAT) and abdominal aortic aneurysm (AAA) growth after endovascular aneurysm repair (EVAR). METHODS: Patients with sub-renal AAA who underwent regular follow-up after EVAR between March 2014 and March 2024 were retrospectively collected. Two radiologists segmented aneurysms and PVAT. Patients were categorised into growing and non-growing groups based on volumetric changes observed in two follow-up computed tomography examinations. One hundred seven radiomic features were automatically extracted from the PVAT region. Univariable and multivariable logistic regression was performed to analyse radiomic features and clinical characteristics. Furthermore, the performance of the integrated clinico-radiological model was compared with models using only radiomic features or clinical characteristics separately. RESULTS: A total of 79 patients (68 ± 9 years, 89% men) were enroled in this study, 19 of whom had a growing aneurysm. Compared to the non-growing group, PVAT of growing AAA showed a higher surface area to volume ratio (non-growing vs growing, 0.63 vs 0.70, p = 0.04), and a trend of low dependence and high dispersion manifested by texture features (p < 0.05). The area under the curve of the integrated clinico-radiological model was 0.78 (95% confidence intervals 0.65-0.91), with a specificity of 87%. The integrated model outperformed models using only radiomi...