van Praagh Gijs D, Vos Francine, Legtenberg Stijn, Wouthuyzen-Bakker Marjan, Kouijzer Ilse J E, Aarntzen Erik H J G, de Vries Jean-Paul P M, Slart Riemer H J A, Alic Lejla, Sinha Bhanu, Saleem Ben R
Department of Nuclear Medicine & Molecular Imaging, Medical Imaging Center, University Medical Center Groningen, University of Groningen, 9713 GZ Groningen, The Netherlands.
Department of Surgery, Division of Vascular Surgery, University Medical Center Groningen, University of Groningen, 9713 GZ Groningen, The Netherlands.
Diagnostics (Basel). 2025 Aug 2;15(15):1944. doi: 10.3390/diagnostics15151944.
We evaluated the feasibility of a machine-learning (ML) model based on clinical features and radiomics from [F]FDG PET/CT images to differentiate between infected and non-infected intracavitary vascular grafts and endografts (iVGEI). Three ML models were developed: one based on pre-treatment criteria to diagnose a vascular graft infection (" features"), another using radiomics features from diagnostic [F]FDG-PET scans, and a third combining both datasets. The training set included 92 patients (72 iVGEI-positive, 20 iVGEI-negative), and the external test set included 20 iVGEI-positive and 12 iVGEI-negative patients. The abdominal aorta and iliac arteries in the PET/CT scans were automatically segmented using SEQUOIA and TotalSegmentator and manually adjusted, extracting 96 radiomics features. The best-performing models for the features and features were selected from 343 unique models. Most relevant features were combined to test three final models using ROC analysis, accuracy, sensitivity, and specificity. The combined model achieved the highest AUC in the test set (mean ± SD: 0.91 ± 0.02) compared with the -only model (0.85 ± 0.06) and the model (0.73 ± 0.03). The combined model also achieved a higher accuracy (0.91 vs. 0.82) than the diagnosis based on all the MAGIC criteria and a comparable sensitivity and specificity (0.70 and 1.00 vs. 0.76 and 0.92, respectively) while providing diagnostic information at the initial presentation. The AUC for the combined model was significantly higher than the model ( = 0.02 in the bootstrap test), while other comparisons were not statistically significant. This study demonstrated the potential of ML models in supporting diagnostic decision making for iVGEI. A combined model using pre-treatment clinical features and PET-radiomics features showed high diagnostic performance and specificity, potentially reducing overtreatment and enhancing patient outcomes.
我们评估了一种基于临床特征和[F]FDG PET/CT图像的放射组学的机器学习（ML）模型区分感染性和非感染性腔内血管移植物及内置物（iVGEI）的可行性。开发了三种ML模型：一种基于诊断血管移植物感染的治疗前标准（“特征”），另一种使用诊断性[F]FDG-PET扫描的放射组学特征，第三种结合了两个数据集。训练集包括92例患者（72例iVGEI阳性，20例iVGEI阴性），外部测试集包括20例iVGEI阳性和12例iVGEI阴性患者。使用SEQUOIA和TotalSegmentator对PET/CT扫描中的腹主动脉和髂动脉进行自动分割并手动调整，提取96个放射组学特征。从343个独特模型中选择了表现最佳的“特征”模型和“特征”模型。结合最相关的特征，使用ROC分析、准确性、敏感性和特异性测试三个最终模型。与仅使用“特征”的模型（0.85±0.06）和“特征”模型（0.73±0.03）相比，组合模型在测试集中的AUC最高（平均值±标准差：0.91±0.02）。组合模型的准确性（0.91对0.82）也高于基于所有MAGIC标准的诊断，并且在初始就诊时提供诊断信息的同时，敏感性和特异性相当（分别为0.70和1.00对0.76和0.92）。组合模型的AUC显著高于“特征”模型（自助法测试中P=0.02），而其他比较无统计学意义。本研究证明了ML模型在支持iVGEI诊断决策方面的潜力。使用治疗前临床特征和PET放射组学特征的组合模型显示出高诊断性能和特异性，可能减少过度治疗并改善患者预后。
