基于MRI的临床-影像-影像组学列线图模型鉴别良恶性实性肺结节或肿块
该回顾性研究纳入145例实性肺结节或肿块患者,随机分为训练组102例和验证组43例,基于临床特征、影像特征及多序列MRI影像组学构建列线图模型预测恶性病变。多序列MRI影像组学模型(T1WI+T2WI+ADC)在训练组和验证组的AUC分别为0.858和0.774;联合临床、影像及影像组学的模型预测效能显著优于临床模型、影像模型和影像组学模型(均P<0.05)。
要点
- 多序列MRI影像组学模型(T1WI+T2WI+ADC)训练组AUC为0.858,验证组AUC为0.774。
- 年龄、DWI值、LSR值、ADC值为恶性病变的独立预测因素。
- 临床-影像-影像组学联合模型预测效能显著优于单一模型(均P<0.05)。
核心结论
- 基于MRI的临床-影像-影像组学模型有助于鉴别良恶性实性肺结节或肿块,可能有助于肺部疾病的精准医学。
研究设计与样本
- 回顾性分析145例实性肺结节或肿块患者(76例男性;平均年龄58.4岁±13.7[SD]),随机分为训练组102例和验证组43例,基于临床特征、影像特征及多序列MRI影像组学构建列线图模型。
诊断标准与关键数值
- 多序列MRI影像组学模型(T1WI+T2WI+ADC)训练组AUC为0.858(95%CI:0.775,0.919),验证组AUC为0.774(95%CI:0.621,0.887)。
- 临床-影像-影像组学联合模型预测效能显著优于临床模型、影像模型和影像组学模型(均P<0.05)。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
RATIONALE AND OBJECTIVES: Pulmonary nodules or masses are highly prevalent worldwide, and differential diagnosis of benign and malignant lesions remains difficult. Magnetic resonance imaging (MRI) can provide functional and metabolic information of pulmonary lesions. This study aimed to establish a nomogram model based on clinical features, imaging features, and multi-sequence MRI radiomics to identify benign and malignant solid pulmonary nodules or masses.
MATERIALS AND METHODS: A total of 145 eligible patients (76 male; mean age, 58.4 years ± 13.7 [SD]) with solid pulmonary nodules or masses were retrospectively analyzed. The patients were randomized into two groups (training cohort, n = 102; validation cohort, n = 43). The nomogram was used for predicting malignant pulmonary lesions. The diagnostic performance of different models was evaluated by receiver operating characteristic (ROC) curve analysis.
RESULTS: Of these patients, 95 patients were diagnosed with benign lesions and 50 with malignant lesions. Multivariate analysis showed that age, DWI value, LSR value, and ADC value were independent predictors of malignant lesions. Among the radiomics models, the multi-sequence MRI-based model (T1WI+T2WI+ADC) achieved the best diagnosis performance with AUCs of 0.858 (95%CI: 0.775, 0.919) and 0.774 (95%CI: 0.621, 0.887) for the training and validation cohorts, respectively. Combining multi-sequence radiomics, clinical and imaging features, the predictive efficacy of the clinical-imaging-radiomics model was significantly better than the clinical model, imaging model and radiomics model (all P < 0.05).
CONCLUSION: The MRI-based clinical-imaging-radiomics model is helpful to differentiate benign and malignant solid pulmonary nodules or masses, and may be useful for precision medicine of pulmonary diseases.
引用
- Xie K, Cui C, Li X, et al. MRI-Based Clinical-Imaging-Radiomics Nomogram Model for Discriminating Between Benign and Malignant Solid Pulmonary Nodules or Masses. Acad Radiol. 2024. PMID: 38644089. doi:10.1016/j.acra.2024.03.042
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。