| 曾显荣,胡佑威,刘庆玲,党鸽,马捷.多层螺旋CT特征联合模型列线图鉴别0期与ⅠA1期肺腺癌[J].中国医学影像技术,2022,38(10):1503~1508 |
| 多层螺旋CT特征联合模型列线图鉴别0期与ⅠA1期肺腺癌 |
| Combined model nomogram based on multi-slice spiral CT characteristics for differentiating stage 0 and ⅠA1 lung adenocarcinoma |
| 投稿时间：2022-01-24 修订日期：2022-06-20 |
| DOI：10.13929/j.issn.1003-3289.2022.10.014 |
| 中文关键词: 肺肿瘤 肿瘤分期 诊断,鉴别 体层摄影术,X线计算机 列线图 |
| 英文关键词:lung neoplasms neoplasm staging diagnosis, differential tomography, X-ray computed nomogram |
| 基金项目: |
| 作者 | 单位 | E-mail | | 曾显荣 | 深圳市人民医院(暨南大学第二临床医学院, 南方科技大学第一附属医院)放射科, 广东 深圳 518020 | | | 胡佑威 | 深圳市人民医院(暨南大学第二临床医学院, 南方科技大学第一附属医院)放射科, 广东 深圳 518020 | | | 刘庆玲 | 深圳市人民医院(暨南大学第二临床医学院, 南方科技大学第一附属医院)放射科, 广东 深圳 518020 | | | 党鸽 | 深圳市人民医院(暨南大学第二临床医学院, 南方科技大学第一附属医院)神经内科, 广东 深圳 518020 | | | 马捷 | 深圳市人民医院(暨南大学第二临床医学院, 南方科技大学第一附属医院)放射科, 广东 深圳 518020 | cjr.majie@vip.163.com |
|
| 摘要点击次数: 1962 |
| 全文下载次数: 495 |
| 中文摘要: |
| 目的 评估多层螺旋CT （MSCT）特征联合模型列线图鉴别0期与ⅠA1期肺腺癌的价值。方法 回顾性分析经手术病理证实的230例肺腺癌患者，其中单发病灶210例，17例存在2个、3例存在3个病灶。根据病理分期将患者分为0期组（n=83）和IA1期组（n=147，含112例微浸润性腺癌和35例浸润性腺癌）。采用单因素分析和二元logistic回归分析筛选肺腺癌分期（0期与ⅠA1期）的独立影响因子，并以之构建联合模型，绘制列线图将其可视化。绘制受试者工作特征（ROC）曲线，计算曲线下面积（AUC），评估各独立影响因子及联合模型鉴别0期与ⅠA1期肺腺癌的效能；绘制校准曲线，行决策曲线分析（DAC），验证模型的效能。结果 分叶征（OR=4.28，P=0.02）、微血管征（OR=2.55，P=0.04）、脐凹征（OR=7.02，P=0.04）、最小密度（OR=1.01，P<0.01）及实性成分占比（OR=1.15，P=0.03）是0期与ⅠA1期肺腺癌的独立影响因子。联合模型鉴别0期与ⅠA1期肺腺癌的AUC为0.88，高于各单一独立影响因子（P均<0.05）。联合模型预测结果与实际情况较吻合，其准确率为0.80，一致性指数为0.88，平均绝对误差为0.01。阈值为0.2~0.9时，联合模型的临床净效益均高于各独立影响因子。结论 MSCT特征联合模型列线图可有效鉴别0期与ⅠA1期肺腺癌。 |
| 英文摘要: |
| Objective To explore the value of the combined model nomogram based on multi-slice spiral CT (MSCT) characteristics for differentiating stage 0 and ⅠA1 lung adenocarcinoma. Methods Data of 230 patients with lung adenocarcinoma confirmed by pathology after surgical resection were retrospectively analyzed, including 210 cases with single lesion, 17 with 2 and 3 cases with 3 lesions, respectively. According to pathological stage, the patients were divided into stage 0 group (n=83) and stage ⅠA1 group (n=147, including 112 microinvasive adenocarcinomas and 35 invasive adenocarcinomas). Univariate analysis and binary logistic regression analysis were used to screen the independent impact factors of lung adenocarcinoma stage (stage 0 and stage ⅠA1) and to construct a combined model, then the nomogram was drawn for visualization. Receiver operating characteristic (ROC) curve was obtained, and the area under the curve (AUC) was calculated to evaluate the efficacy of single impact factors and combined model for differentiating stage 0 and ⅠA1 lung adenocarcinoma. Then the calibration curve and decision curve analysis (DAC) were drawn to verify the efficacy of the model. Results Lobulation sign (OR=4.28, P=0.02), microvascular sign (OR=2.55, P=0.04), notch sign (OR=7.02, P=0.04), minimum density (OR=1.01, P<0.01) and consolidation tumor ratio (OR=1.15, P=0.03) were the independent impact factors of lung adenocarcinoma stage (stage 0 and stage ⅠA1). AUC of the combined model in distinguishing stage 0 and ⅠA1 lung adenocarcinoma was 0.88, higher than that of single independent factors (all P<0.05). The predicted performance of the combined model was consistent with the actual situation, with accuracy of 0.80, C-index of 0.88 and the mean absolute error of 0.01. When the threshold was 0.2 to 0.9, the net clinical benefit of the combined model was higher than that of single independent impact factor. Conclusion Combined model nomogram based on MSCT features could effectively differentiate stage 0 and ⅠA1 lung adenocarcinoma. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
