CECT-based simplified scoring model for differentiating low-risk thymomas from hyper-attenuating thymic cysts: A STARD compliant article.

This retrospective study aims to evaluate the effectiveness of a simplified scoring model utilizing contrast-enhanced computed tomography (CECT) in distinguishing low-risk thymomas (LRTs) from thymic cysts in patients with anterior mediastinal hyper-attenuating nodules. A total of 32 patients of LRTs and 40 patients of hyper-attenuating thymic cysts who underwent chest biphasic CECT preoperatively from January 2015 to December 2022 were included. The traditional CT imaging features and clinical features of each patient were analyzed. A predictive model was built by multivariable logistic regression, and subsequently, a simplified scoring model was developed according to the regression coefficients of each risk factor of LRTs. The performance of risk factors and models were assessed by receiver operating characteristic curve, decision curve analysis, and DeLong test. Compared to hyper-attenuating thymic cysts, LRTs tended to be located off-midline, higher CT values in the venous phase, and moderate to severe enhancement (all P < .001). Based on the above risk factors of LTRs, the predictive model achieved an area under the receiver operating characteristic curve of 0.938. While the simplified scoring model demonstrated comparable diagnostic ability (area under the receiver operating characteristic curve = 0.936, P = .42), with ideal sensitivity (0.719), accuracy (0.861), and specificity (0.975). Decision curve analysis indicated this scoring model provided a higher clinical net benefit. Biphasic CECT had a strong diagnostic capability in differentiating LRTs from hyper-attenuating thymic cysts. The diagnostic scoring model is straightforward and convenient, making it easy to popularize.