Objectives To develop and validate a model that integrates dual-energy computed tomography (DECT) and clinical features for the prediction of the risk of frequent gout flares (FrGF). Materials and methods This retrospective cohort study included 1204 patients with gout who were randomly divided into training and validation cohorts and followed for 12 months after DECT. The patients were categorized into an FrGF group (≥ 2 flares, n = 606) and an infrequent gout flares (InGF) group (< 2 flares, n = 598) based on flare frequency during follow-up. Clinicoradiological features that were significant in univariate analysis were dimensionally reduced via least absolute shrinkage and selection operator regression, followed by multivariate logistic regression to identify independent risk factors for FrGF. An integrated clinicoradiological model was constructed using a nomogram. Model performance was evaluated using the receiver-operating characteristic (ROC) curve and clinical utility by decision curve analysis. The integrated model was compared with the clinical model using the DeLong test to evaluate the incremental value of radiological features. Results Independent risk factors for FrGF were chronic arthritis, recurrent gout at baseline, intermittent medication, non-adherence to a low-purine diet, high monosodium urate burden, elevated Sharp/van der Heijde score, and tarsal bone erosion. The integrated model had an area under the ROC curve (AUC) of 0.95 (95% CI 0.92–0.97) in the validation cohort, outperforming the clinical model (AUC 0.83, 95% CI 0.79–0.87), demonstrating the incremental diagnostic value of DECT. Conclusion The integrated model demonstrated excellent discriminative ability, calibration, and clinical value, enabling accurate prediction of the risk of FrGF. Key Points <Emphasis Type="BoldItalic">Question</Emphasis> Frequent gout and infrequent gout differ in treatment strategies, but how can clinicians prospectively and accurately identify frequent gout? <Emphasis Type="BoldItalic">Findings</Emphasis> Based on clinical and DECT features, this study identified 7 independent risk factors for frequent gout and developed and validated a predictive model. <Emphasis Type="BoldItalic">Clinical relevance</Emphasis> This study’s predictive model exhibits favorable predictive performance, and its web-based calculator helps clinicians assess the risk of frequent gout flares and predict clinical intervention efficacy, thus providing valuable evidence for formulating individualized clinical treatment plans. Graphical Abstract

Combination of dual-energy CT findings and clinical features predicts risk of frequent flares of gout

期刊论文
发表日期：

2025年11月1日

查看全文

Yaoning Wei,
Kai Wang,
Jinling Liu,
Meihan Chen,
Zhitao Yang,
Changgui Li,
Lin Han,
Ying Chen,
Pei Nie,
Wenjian Xu

摘要

Objectives

To develop and validate a model that integrates dual-energy computed tomography (DECT) and clinical features for the prediction of the risk of frequent gout flares (FrGF).

Materials and methods

This retrospective cohort study included 1204 patients with gout who were randomly divided into training and validation cohorts and followed for 12 months after DECT. The patients were categorized into an FrGF group (≥ 2 flares, n = 606) and an infrequent gout flares (InGF) group (< 2 flares, n = 598) based on flare frequency during follow-up. Clinicoradiological features that were significant in univariate analysis were dimensionally reduced via least absolute shrinkage and selection operator regression, followed by multivariate logistic regression to identify independent risk factors for FrGF. An integrated clinicoradiological model was constructed using a nomogram. Model performance was evaluated using the receiver-operating characteristic (ROC) curve and clinical utility by decision curve analysis. The integrated model was compared with the clinical model using the DeLong test to evaluate the incremental value of radiological features.

Results

Independent risk factors for FrGF were chronic arthritis, recurrent gout at baseline, intermittent medication, non-adherence to a low-purine diet, high monosodium urate burden, elevated Sharp/van der Heijde score, and tarsal bone erosion. The integrated model had an area under the ROC curve (AUC) of 0.95 (95% CI 0.92–0.97) in the validation cohort, outperforming the clinical model (AUC 0.83, 95% CI 0.79–0.87), demonstrating the incremental diagnostic value of DECT.

Conclusion

The integrated model demonstrated excellent discriminative ability, calibration, and clinical value, enabling accurate prediction of the risk of FrGF.

Key Points

Question Frequent gout and infrequent gout differ in treatment strategies, but how can clinicians prospectively and accurately identify frequent gout?

Findings Based on clinical and DECT features, this study identified 7 independent risk factors for frequent gout and developed and validated a predictive model.

Clinical relevance This study’s predictive model exhibits favorable predictive performance, and its web-based calculator helps clinicians assess the risk of frequent gout flares and predict clinical intervention efficacy, thus providing valuable evidence for formulating individualized clinical treatment plans.

Graphical Abstract