ABSTRACT
Background: The use of an artificial intelligence (AI)-based diagnostic system can significantly aid in analyzing the histogram of pulmonary nodules. The aim of our study was to evaluate the value of computed tomography (CT) histogram indicators analyzed by AI in predicting the tumor invasiveness of ground-glass nodules (GGNs) and to determine the added value of contrast-enhanced CT (CECT) compared with nonenhanced CT (NECT) in this prediction. Methods: This study enrolled patients with persistent GGNs who underwent preoperative NECT and CECT scanning. AI-based histogram analysis was performed for pathologically confirmed GGNs, which was followed by screening invasiveness-related factors via univariable analysis. Multivariable logistic models were developed based on candidate CT histogram indicators measured on either NECT or CECT. Receiver operating characteristic (ROC) curve and precision-recall (PR) curve were used to evaluate the models' performance. Results: A total of 116 patients comprising 121 GGNs were included and divided into the precancerous lesion and adenocarcinoma groups based on invasiveness. In the AI-based histogram analysis, the mean CT value [NECT: odds ratio (OR) =1.009; 95% confidence interval (CI): 1.004-1.013; P<0.001] and solid component volume (NECT: OR =1.005; 95% CI: 1.000-1.010; P=0.032) were associated with the adenocarcinoma and used for multivariable logistic modeling. The area under ROC curve (AUC) and PR curve (AUPR) were not significantly different between the NECT model (AUC =0.765, 95% CI: 0.679-0.837; AUPR =0.907, 95% CI: 0.825-0.953) and the optimal CECT model (delayed phase: AUC =0.772, 95% CI: 0.687-0.843; AUPR =0.895, 95% CI: 0.812-0.944). No significantly different metrics were observed between the NECT and CECT models (precision: 0.707 vs. 0.742; P=0.616). Conclusions: The AI diagnostic system can help in the diagnosis of GGNs. The system displayed decent performance in GGN detection and alert to malignancy. Mean CT value and solid component volume were independent predictors of tumor invasiveness. CECT provided no additional improvement in diagnostic performance as compared with NECT.
ABSTRACT
Objective: To assess the feasibility and diagnostic performances of synthetic magnetic resonance imaging (SyMRI) combined with diffusion-weighted imaging (DWI) and differential subsampling with cartesian ordering (DISCO) in breast imaging reporting and data system (BI-RADS) 4 lesions. Methods: A total of 98 BI-RADS 4 patients, including 68 cases assigned to a malignant group and 33 cases assigned to a benign group, were prospectively enrolled, and their MRI and clinical information were collected. Two physicians jointly analyzed the characteristics of conventional MRI. T1, T2, proton density (PD), and ADC values were obtained from three different regions of interest (ROIs). Logistic regression analyses were used to select features and build models, and a nomogram was constructed with the best model. Results: Using the ROI delineation method at the most obvious enhancement to measure the ADC value revealed the best diagnostic performance in diagnosing BI-RADS type 4 mass lesions. The diagnostic efficiency of the maximum level drawing method of the quantitative relaxation model was better than that of the whole drawing method and the most obvious enhancement method. The best relaxation model (model A) was composed of two parameters: T2stand and ΔT1%stand (AUC=0.887), and the BI-RADS model (model B) was constructed by two MRI features of edge and TIC curve (AUC=0.793). Using the quantitative parameters of SyMRI and DWI of the best ROC method combined with DISCO enhanced MRI features to establish a joint diagnostic model (model C: edge, TIC curve type, ADClocal, T2stand, ΔT1%stand) showed the best diagnostic efficiency (AUC=0.953). The nomogram also had calibration curves with good overlap. Conclusions: The combined diagnosis model of SyMRI and DWI quantitative parameters combined with DISCO can improve the diagnostic efficiency of BI-RADS 4 types of mass lesions. Also, the line diagram based on this model can be used as an auxiliary diagnostic tool.
