Predictive model for interventional efficacy in lower extremity arteriosclerosis obliterans
10.3760/cma.j.cn115396-20240521-00153
- VernacularTitle:下肢动脉硬化闭塞症介入疗效的预测模型
- Author:
Zhenwei YANG
1
;
Qingrui WU
;
Wenjie MA
;
Ye TIAN
Author Information
1. 新疆医科大学第一附属医院血管甲状腺外科，乌鲁木齐 830054
- Keywords:
Arteriosclerosis obliterans;
Lower extremity;
Logistic models;
Nomograms;
GLASS
- From:
International Journal of Surgery
2024;51(7):446-454
- CountryChina
- Language:Chinese
-
Abstract:
Objective:To develop a predictive model for the intervention efficacy of lower extremity atherosclerotic occlusive disease (LEASO) and evaluate its performance to predict the outcomes of intervention therapy for patients with lower extremity atherosclerotic occlusive disease.Methods:This study retrospectively analyzed data from 238 patients with lower extremity atherosclerotic occlusive disease (LEASO), including 188 males and 50 females, aged between 35 and 88 years with a mean age of 68 years. These patients were randomly divided in a 7∶3 ratio into a training set ( n=166) and a testing set ( n=72) based on adverse outcomes, both training and test sets were divided into MALEs and non-MALEs groups. The training set had 67 MALEs and 99 non-MALEs, while the test set had 26 MALEs and 46 non-MALEs. Important variables related to outcome events were selected using LASSO regression in the training set and incorporated into a multifactorial logistic regression model to construct a predictive model. The model was visualized using forest plots and its performance was evaluated using data from both the training and testing sets. Results:Through LASSO regression, SIIRI(Systemic immune inflammatory response index, SIIRI), Rutherford >4, IP(Infrapopliteal, IP)>1, and P(Pedal, P)≥1 were selected as predictive indicators for the model. The area under the curve, sensitivity, and specificity of the model in the training set and testing set were 0.813, 80.6%, 72.7%, and 0.764, 65.4%, 80.4%. The calibration curve was consistent with expectations. The decision curves of the model had the highest accuracy, net benefit rate for clinical application of the model when the threshold probabilities of the training set and test set were in the range of 0~0.79 and 0~0.66.Conclusions:The predictive model built using preoperative Rutherford classification, IP classification, P classification, and SIIRI can identify high-risk individuals for early detection of MALEs and provide targeted intensified treatment. This model has practical significance in improving the prognosis of such patients and can be applied in clinical practice.
