| 叶颖剑,安鹏.临床-CT环境影像组学鉴别慢性阻塞性肺疾病背景下周围型肺癌与肺炎性肿块[J].中国医学影像技术,2024,40(7):1030~1035 |
| 临床-CT环境影像组学鉴别慢性阻塞性肺疾病背景下周围型肺癌与肺炎性肿块 |
| Clinical-CT environmental-radiomics for distinguishing peripheral lung cancer and inflammatory mass under background of chronic obstructive pulmonary disease |
| 投稿时间：2023-12-24 修订日期：2024-03-03 |
| DOI：10.13929/j.issn.1003-3289.2024.07.015 |
| 中文关键词: 肺肿瘤 肺疾病,慢性阻塞性 影像组学 体层摄影术,X线计算机 |
| 英文关键词:lung neoplasms pulmonary disease, chronic obstructive radiomics tomography, X-ray computed |
| 基金项目: |
|
| 摘要点击次数: 629 |
| 全文下载次数: 246 |
| 中文摘要: |
| 目的 观察临床-CT环境影像组学鉴别慢性阻塞性肺疾病(COPD)背景下周围型肺癌(PLC)与肺炎性肿块的价值。方法 回顾性分析经病理证实PLC的86例COPD (PLC组)及155例肺炎性肿块COPD(炎性肿块组),按7 ∶ 3比例将其分为训练集(n=170)及测试集(n=71);基于增强CT勾画病灶ROI1(不均匀强化区)、ROI2(均匀强化区)、ROI3(肿瘤周围带),生成相应Radscore 1、2、3。比较组间临床资料、常规CT及环境影像组学资料,行logistic回归分析,建立临床模型、CT环境影像组学模型及临床-CT环境影像组学模型,评估其鉴别PLC与肺炎性肿块的效能。结果 病灶形态、强化方式、Radscore 2及3均为鉴别COPD背景下PLC与肺炎性肿块的因素(P均＜0.05)。基于此建立的临床模型、CT环境影像组学模型及临床-CT环境影像组学模型鉴别训练集COPD背景下PLC与肺炎性肿块的曲线下面积(AUC)分别为0.763、0.859及0.892,在测试集分别为0.729、0.843及0.882;临床-CT影像组学模型AUC最高(P均＜0.05),其准确率为83.53%、敏感度为81.97%、特异度为84.40%。结论 临床-CT环境影像组学有助于鉴别COPD背景下PLC与肺炎性肿块。 |
| 英文摘要: |
| Objective To observe the value of clinical-CT environmental-radiomics for distinguishing peripheral lung cancer (PLC) and inflammatory mass under the background of chronic obstructive pulmonary disease (COPD). Methods Data of 86 COPD patients with pathologically confirmed PLC (PLC group) and 155 COPD with inflammatory masses (inflammatory mass group) were retrospectively analyzed. The patients were divided into training set (n=170) and test set (n=71) at the ratio of 7∶3. Based on enhanced CT, the lesion ROI1 (uneven enhancement area), ROI2 (uniform enhancement area) and ROI3 (tumor surrounding zone) were delineated, and the corresponding Radscore 1, 2 and 3 were generated. Clinical, routine CT and environmental-radiomics data were compared between groups. Logistic regression analysis was performed, then clinical model, CT environmental-radiomics model and clinical-CT environmental-radiomics model were established, and their efficacy for distinguishing PLC and inflammatory mass were analyzed. Results Lesion morphology, enhancement mode, Radscore 2 and 3 were all impact factors for distinguishing PLC and pneumonia mass under background of COPD (all P＜0.05). The area under the curve (AUC) of the established clinical model, CT environmental-radiomics model and clinical-CT environmental-radiomics model for distinguishing PLC and pneumonia mass in COPD was 0.763, 0.859 and 0.892 in training set, and was 0.729, 0.843 and 0.882 in test set, respectively. AUC of clinical-CT environmental-radiomics model had the highest AUC (all P＜0.05), with the accuracy of 83.53%, sensitivity of 81.97% and specificity of 84.40%. Conclusion Clinical-CT environmental-radiomics was helpful for distinguishing PLC and pneumonia mass under the background of COPD. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
