Development and validation of a multi-slice CTA-based prediction model for poor outcomes in isolated superior mesenteric artery dissection

TL;DR

This paper develops a prediction model using CT scans and clinical data to identify high-risk patients with superior mesenteric artery dissection.

Contribution

A novel multi-slice CTA-based prediction model using machine learning to predict poor outcomes in ISMAD patients.

Findings

The random forest model outperformed other models with an AUC of 0.849.

Key predictors included blood lactate levels, SMA stenosis, and intestinal wall thickening.

Abstract

A prediction model for poor outcomes in patients with isolated superior mesenteric artery dissection (ISMAD) was constructed and validated based on multi-slice spiral CT angiography (MSCTA) imaging features and clinical indicators, aiming to provide a basis for early clinical identification of high-risk patients and formulation of individualized treatment strategies.

A total of 360 patients with ISMAD who were admitted to our hospital from January 2021 to December 2024 were retrospectively included. They were randomly divided into a training set (n = 252) and a validation set (n = 108) at a ratio of 7:3. The demographic characteristics, clinical symptoms and signs, laboratory test results, and MSCTA imaging features of the patients were collected. In the training set, indicators associated with poor outcomes were screened by univariate analysis, least absolute shrinkage and selection operator (LASSO) regression, and multivariate logistic regression analysis. Random forest, support vector machine, and gradient boosting models were constructed. The efficacy of the models was evaluated by the area under the receiver operating characteristic curve (AUC), the optimal model was selected, and the importance of key prediction indicators was analyzed.

There was no significant difference in the baseline data between the training set and the validation set (P > 0.05). Multivariate logistic regression analysis indicated that the visual analog scale (VAS) for abdominal pain, blood lactate levels, minimum diameter of the true lumen of the superior mesenteric artery (SMA), degree of stenosis of the SMA trunk, degree of intestinal wall thickening, and range of false lumen thrombosis formation were independent risk factors for poor outcomes (P < 0.05). The AUC of the random forest model (0.849) was significantly higher than that of the support vector machine (0.828) and gradient boosting models (0.818), making it the optimal model.

A random forest model constructed based on MSCTA imaging features and clinical indicators can effectively predict poor outcomes in patients with ISMAD. Blood lactate levels, VAS score for abdominal pain, minimum true lumen diameter, degree of SMA trunk stenosis, intestinal wall thickening, and extent of false lumen thrombosis were identified as key predictors.

Figures