目的：探讨肺结节（≤2 cm）及其周围组织的影像组学特征在其良恶性鉴别中的价值。方法：回顾性分析了206例肺结节患者的薄层CT轴位平扫图像，包括恶性106例，良性100例，由两名放射科医师（分别具有3年和10年的胸部CT影像诊断经验）在不知病理结果的情况下分别对206例结节进行良恶性评估；同时对206例结节进行影像组学分析，通过3D Slicer 勾画三维容积感兴趣区（VOI）、Analysis-Kinetics分析软件提取影像组学特征，使用Lasso-logistic回归分析进行特征筛选，并分别建立肺结节模型、联合肺结节及其周围5 mm、10 mm、15 mm组织模型。通过受试者操作特征曲线（ROC）分析评价模型的诊断效能，使用DeLong检验比较不同模型的效能，并与放射科医师的诊断结果进行比较。结果：两名放射科医师的AUC值分别为0.81、0.69。在验证组中，肺结节模型的AUC值为0.82，联合结节及其周围5 mm、10 mm、15 mm组织模型的AUC值分别为0.88、0.76、0.82，除联合结节及其周围10 mm组织模型外，其余模型效能均高于医师组，但各模型之间的效能差异无统计学意义（DeLong检验，P>0.05）。结论：针对≤2 cm的肺结节，基于结节及其周围组织的影像组学模型可提高鉴别结节良恶性的能力。
Abstract
Objective: To investigate the value of intranodular and perinodular radiomic features to distinguish malignant from benign pulmonary nodules(≤2 cm). Methods: A retrospective analysis of 206 patients with pulmonary nodules with noncontrast CT images, including 106 malignant patients and 100 benign patients, was performed by two radiologists with 3 and 10 years of chest CT imaging experience, respectively, to evaluate without pathological results. At the same time, radiomic analysis was performed on 206 cases of nodules, and VOIs were delineated by 3D slicer, radiomic features were extracted by Analysis-Kinetics(A.K.) analysis software. The Lasso-logistic regression was performed to select features, and to establish lung nodule models, models of combining intranodular with perinodular 5 mm, 10 mm, and 15 mm tissue, respectively. ROC curve analysis and Delong test was used to compare the diagnostic efficacy among models and radiologists. Results: The AUC values of the two radiologists were 0.81 and 0.69. In the validation group, the AUC value of the lung nodule model was 0.82, and the AUC values of combining intranodular with perinodular 5 mm, 10 mm, and 15 mm tissue models were 0.88, 0.76, and 0.82. Except for the combining intranodular with perinodular 10 mm tissue model, the efficacy of the other models was higher than that of the physician group, but there was no significant difference in efficacy between the models(DeLong test, P>0.05). Conclusion: For lung nodules ≤2 cm, the models based on intranodular and perinodular radiomic features can improve the ability to distinguish benign and malignant nodules.
关键词
硬币病变 /
肺 /
体层摄影术 /
螺旋计算机
Key words
Coin lesion, pulmonary /
Tomography, spiral computed
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
[1]Siegel RMJ, Zou Z, Jemal A. Cancer statistics[J]. Cancer J Clin, 2014, 64(1): 9-29.
[2]Rami-Porta R, Bolejack V, Crowley J, et al. The IASLC lung cancer staging project: proposals for the revisions of the T descriptors in the forth coming eighth edition of the TNM classification for lung cancer[J]. J Thorac Oncol, 2015, 10(7): 990-1003.
[3]McWilliams A, Tammemagi MC, Mayo JR, et al. Probability of cancer in pulmonary nodules detected on first screening CT[J]. N Engl J Med, 2013, 369(10): 910-919.
[4]Siegelman SS, Zerhouni EA, Leo FP, et al. CT of the solitary pulmonary nodule[J]. AJR, 1980, 135(1): 1-13.
[5]彭涛，肖建明，张仕慧，等. 基于多参数MRI及影像组学建立机器学习模型诊断临床显著性前列腺癌[J]. 中国医学影像技术，2019，35（10）：1526-1530.
[6]范丽，方梦捷，董迪，等. 影像组学对磨玻璃结节型肺腺癌病理亚型的预测效能[J]. 中华放射学杂志，2017，51（12）：912-917.
[7]谢元亮，杜丹，谢伟，等. DCE-MRI纹理分析鉴别宫颈鳞癌与腺癌及预测分级的价值[J]. 放射学实践，2019，34（8）：835-840.
[8]孙丽，李绍东，吴雪元，等. 磁共振纹理分析评价直肠癌新辅助放化疗疗效的价值[J]. 中国临床医学影像杂志，2019，30（8）：581-585；598.
（下转485页）
（上接481页）
[9]Prasanna P, Patel J, Partovi S, et al. Radiomic features from the peritumoral brain parenchyma on treatment-naive multi-parametric MR imaging predict long versus short-term survival in glioblastoma multiforme: preliminary findings[J]. Eur Radiology, 2017, 27(10): 4188-4197.
[10]Wang X, Zhao XY, Li Q, et al. Can peritumoral radiomics increase the efficiency of the prediction for lymph node metastasis in clinical stage T1 lung adenocarcinoma[J]. Eur Radiol, 2019, 29(11): 6049-6058.
[11]Braman NM, Etesami M, Prasanna P, et al. Intratumoral and peritumoral radiomics for the pretreatment prediction of pathological complete response to neoadjuvant chemotherapy based on breast DCE-MRI[J]. Breast Cancer Res, 2017, 19(1): 57.
[12]Mohiuddin K, Haneuse S, Sofer T, et al. Relationship between margin distance and local recurrence among patients undergoing wedge resection for small(≤2 cm) non-small cell lung cancer[J]. J Thorac Cardiovasc Surg, 2014, 147(4): 1169-1175.
[13]Lambin P, Rios-Velazquez E, Leijenaar R, et al. Radiomics: extracting more information from medical images using advanced feature analysis[J]. Eur J Cancer, 2012, 48(4): 441-446.
[14]Way TW, Hadjiiski LM, Sahiner B, et al. Computer-aided diagnosis of pulmonary nodules on CT scans: segmentation and classification using 3D active contours[J]. Med Phys, 2006, 33(7): 2323-2337.
[15]Way TW, Sahiner B, Chan HP, et al. Computer-aided diagnosis of pulmonary nodules on CT scans: improvement of classification performance with nodule surface features[J]. Med Physics, 2009, 36(7): 3086-3098.
[16]Liu Y, Balagurunathan Y, Atwater T, et al. Radiological image traits predictive of cancer status in pulmonary nodules[J]. Clin Cancer Res, 2017, 23(6): 1442-1449.
[17]Beig N, Khorrami M, Alilou M, et al. Perinodular and intranodular radiomic features on lung CT images distinguish adenocarcinomas from granulomas[J]. Radiology, 2019, 290(3): 783-792.
基金
本课题受山西省自然科学基金资金资助（项目编号201801D121200；201701D121151）。
