国际医药卫生导报 ›› 2025, Vol. 31 ›› Issue (3): 458-462.DOI: 10.3760/cma.j.cn441417-20240111-03022

• 临床研究 • 上一篇    下一篇

CT纹理分析及形态学特征对肺磨玻璃结节良恶性的诊断价值

马旭阳1  蒋小锋1  苏美俊1  郑琦1  李政和1  黄燕1  杨宁丽2  唐敏丽1  杨全新1   

  1. 1西安交通大学,西安 710000;2延安大学咸阳医院CT/MRI室,咸阳 712000

  • 收稿日期:2024-01-11 出版日期:2025-02-01 发布日期:2025-02-21
  • 通讯作者: 杨全新,Email:quanxin1962@163.com
  • 基金资助:

    咸阳市秦都区科技计划(QK2022sf02)

Diagnostic value of CT texture analysis and morphological features in benign and malignant pulmonary ground glass nodules

Ma Xuyang1, Jiang Xiaofeng1, Su Meijun1, Zheng Qi1, Li Zhenghe1, Huang Yan1, Yang Ningli2, Tang Minli1, Yang Quanxin1   

  1. 1 Xi'an Jiaotong University, Xi'an 710000, China; 2 CT/MRI Room, Yan'an University Xianyang Hospital, Xianyang 712000, China

  • Received:2024-01-11 Online:2025-02-01 Published:2025-02-21
  • Contact: Yang Quanxin, Email: quanxin1962@163.com
  • Supported by:

    Xianyang City Qindu District Science and Technology Plan (QK2022sf02)

摘要:

目的 探讨电子计算机断层扫描(CT)纹理分析及形态学特征对肺磨玻璃结节良恶性的诊断价值。方法 选取2021年1月至2023年1月延安大学咸阳医院收治的80例肺磨玻璃结节患者进行回顾性研究,所有患者均接受手术切除治疗,并取得病理结果,依据病理检查结果将患者分为良性组(34例)和恶性组(46例)。良性组:男20例,女14例,年龄35~83(58.32±11.25)岁。恶性组:男26例,女20例,年龄34~85(58.93±12.48)岁。收集所有患者术前CT图像资料,对比分析两组患者肺结节形态学特征及CT纹理参数。绘制受试者操作特征曲线(ROC),计算曲线下面积(AUC),分析CT纹理参数及形态学特征对肺良、恶性磨玻璃结节的诊断效能。统计学方法采用t检验、H检验、χ2检验。结果 良性组均和、熵、熵和、熵差均低于恶性组(63.93±5.32比68.93±12.36、1.53±0.28比1.94±0.37、1.03±0.22比1.33±0.58、0.90±0.08比1.23±0.11),差异均有统计学意义(t=2.209、5.413、2.863、14.823,均P<0.05);恶性组长径、短径、体积、实性部分占位、平均CT值均大于良性组,分叶征、毛刺征、空泡征、支气管充气征、胸膜凹陷、血管集束、支气管截断占比均高于良性组,差异均有统计学意义(均P<0.05)。多因素logistic回归分析显示,均和、熵、熵和、熵差对肺磨玻璃良、恶性结节的检出率差异均有统计学意义(均P<0.05)。经ROC分析,均和、熵、熵和、熵差诊断肺磨玻璃结节良、恶性的AUC分别为0.757、0.690、0.734、0.763,特异度分别为88.24%、79.38%、67.29%、76.37%,灵敏度分别为58.25%、59.34%、68.27%、79.36%,CT纹理结合形态学特征对肺磨玻璃结节良恶性的诊断具有更高的特异度和灵敏度,分别为81.34%、95.68%。结论 CT纹理及形态学特征对鉴别诊断肺磨玻璃结节良、恶性具有重要价值,能够为肺磨玻璃结节癌前病变及临床治疗方案的制定提供一定的指导和参考。

关键词:

肺磨玻璃结节, 良恶性, 形态学, 电子计算机断层扫描, 纹理分析

Abstract:

Objective To explore the diagnostic value of computerized tomography (CT) texture analysis and morphological features in benign and malignant pulmonary ground glass nodules. Methods A retrospective study was conducted on the clinical data of 80 patients with pulmonary ground glass nodules admitted to Yan'an University Xianyang Hospital from January 2021 to January 2023. All the patients received surgical resection and pathological findings were obtained, according to the pathological findings, all patients were divided into a benign group (34 cases) and a malignant group (46 cases). In the benign group, there were 20 males and 14 females, aged 35-83 (58.32±11.25) years. In the malignant group, there were 26 males and 20 females, aged 34-85 (58.93±12.48) years. Preoperative CT image data of all the patients were collected, and the CT texture parameters and morphological characteristics of pulmonary nodules in the two groups were compared and analyzed. The receiver operating characteristic curve (ROC) was drawn, and the area under the curve (AUC) was calculated to analyze the diagnostic efficacy of CT texture parameters and morphological characteristics in the diagnosis of benign and malignant pulmonary ground glass nodules. The statistical methods used were t test, H test, and χ2 test. Results The mean sum, entropy, entropy sum, and entropy difference of the benign group were lower than those of the malignant group (63.93±5.32 vs. 68.93±12.36, 1.53±0.28 vs. 1.94±0.37, 1.03±0.22 vs. 1.33±0.58, 0.90±0.08 vs. 1.23±0.11), with statistically significant differences (t=2.209, 5.413, 2.863, and 14.823, all P<0.05). The long diameter, short diameter, volume, solid partial space, average CT value, lobular sign, burr sign, vacuolar sign, air bronchial sign, pleural depression, vascular bunching, and bronchial truncation in the malignant group were higher than those in the benign group, with statistically significant differences (all P<0.05). Multivariate logistic regression analysis showed that the mean sum, entropy, entropy sum, and entropy difference had statistical significance in the detection of benign and malignant pulmonary ground glass nodules (all P<0.05). According to ROC analysis, the AUC for benign and malignant pulmonary ground glass nodules were 0.757, 0.690, 0.734, and 0.763, the specificities were 88.24%, 79.38%, 67.29%, and 76.37%, and the sensitivities were 58.25%, 59.34%, 68.27%, and 79.36%; CT texture combined with morphological features had higher specificity and sensitivity for the diagnosis of benign and malignant pulmonary ground glass nodules, which were 81.34% and 95.68%, respectively. Conclusion CT texture and morphological features are of great value in the differential diagnosis of benign and malignant pulmonary ground glass nodules, and can provide certain guidance and reference for the formulation of precancerous lesions and clinical treatment of pulmonary ground glass nodules.

Key words:

Pulmonary ground glass nodules, Benign and malignant, Morphology, Computed tomography, Texture analysis