切换至 "中华医学电子期刊资源库"

中华结直肠疾病电子杂志 ›› 2025, Vol. 14 ›› Issue (05) : 457 -467. doi: 10.3877/cma.j.issn.2095-3224.2025.05.010

论著

基于增强CT的二维、三维影像组学和联合模型对术前预测结直肠癌脉管侵犯价值研究
张娴1, 王彬瞻2, 王馨媛2, 罗再1, 王庆国3, 程云章2, 黄陈1,()   
  1. 1200080 上海交通大学医学院附属第一人民医院胃肠外科
    2200093 上海理工大学健康科学与工程学院
    3200080 上海交通大学医学院附属第一人民医院放射科
  • 收稿日期:2025-02-25 出版日期:2025-10-25
  • 通信作者: 黄陈

Preoperative prediction of lymphovascular invasion in colorectal cancer: a comparative study of 2D, 3D radiomics models, and integrated clinical-radiological models based on enhanced CT

Xian Zhang1, Binzhan Wang2, Xinyuan Wang2, Zai Luo1, Qingguo Wang3, Yunzhang Cheng2, Chen Huang1,()   

  1. 1Department of Gastrointestinal Surgery, Shanghai General Hospital, Shanghai 200080, China
    2School of Healith Science and Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China
    3Department of Radiology, Shanghai General Hospital, Shanghai 200080, China
  • Received:2025-02-25 Published:2025-10-25
  • Corresponding author: Chen Huang
引用本文:

张娴, 王彬瞻, 王馨媛, 罗再, 王庆国, 程云章, 黄陈. 基于增强CT的二维、三维影像组学和联合模型对术前预测结直肠癌脉管侵犯价值研究[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(05): 457-467.

Xian Zhang, Binzhan Wang, Xinyuan Wang, Zai Luo, Qingguo Wang, Yunzhang Cheng, Chen Huang. Preoperative prediction of lymphovascular invasion in colorectal cancer: a comparative study of 2D, 3D radiomics models, and integrated clinical-radiological models based on enhanced CT[J/OL]. Chinese Journal of Colorectal Diseases(Electronic Edition), 2025, 14(05): 457-467.

目的

探讨基于增强CT的二维(2D)、三维(3D)影像组学模型和结合临床特征的联合模型在术前预测结直肠癌脉管侵犯(LVI)中的价值。

方法

回顾性收集2015年1月至2019年12月在上海交通大学医学院附属第一人民医院接受结直肠癌根治术的303例患者的临床病理和增强CT静脉期影像资料,根据LVI状态分为阳性(n=122)和阴性(n=181)。按7:3比例随机分为训练集(n=212)和测试集(n=91),其中训练集LVI阳性90例,阴性122例;测试集LVI阳性32例,阴性59例。通过卡方检验、Mann-Whitney U检验及多因素Logistic回归分析筛选与结直肠癌LVI相关的临床病理指标。在静脉期增强CT图像上由2名放射科医师分别手动勾画肿瘤的2D感兴趣区(ROI)和3D感兴趣区(VOI),通过Mann-Whitney U检验、组间相关系数(ICC)及最小绝对收缩和选择算子(LASSO)筛选影像组学特征。基于筛选的特征,采用支持向量机(SVM)、Logistic回归(LR)、随机森林(RF)和极端梯度提升树(XGBoost)机器学习算法分别构建2D、3D影像组学模型和联合模型。使用受试者工作特征曲线(ROC)、曲线下面积(AUC)和Delong检验评估模型性能,并使用校准曲线和决策曲线评估模型的诊断效能和临床获益情况。

结果

N分期、神经侵犯和癌结节是结直肠癌LVI的临床独立风险因素,基于LR分类器构建的各模型表现较佳。2D影像组学模型在训练集和测试集的AUC分别为0.734(95%CI:0.660~0.798)和0.683(95%CI:0.562~0.805);3D影像组学模型在训练集和测试集的AUC分别为0.784(95%CI:0.720~0.843)和0.726(95%CI:0.602~0.830);结合N分期、神经侵犯和癌结节三个临床特征的2D联合模型在训练集和测试集的AUC分别为0.843(95%CI:0.786~0.899)和0.819(95%CI:0.723~0.905);3D联合模型在训练集和测试集的AUC分别为0.907(95%CI:0.865~0.943)和0.874(95%CI:0.790~0.941)。校准曲线显示2D和3D模型具有相当的预测准确性。决策曲线显示3D联合模型具有更大的临床获益。Delong检验结果显示,各分类器在开发2D和3D预测模型上的性能差异并不显著,而3D联合模型与2D影像组学模型差异有统计学意义(P<0.05)。

结论

2D和3D勾画方法各有优劣,基于LR构建的3D联合模型预测性能较好。基于两种勾画方法构建的2D、3D模型均能在术前有效预测结直肠癌患者的LVI状态,为临床灵活决策提供重要依据。

Objective

To evaluate the efficacy of 2D and 3D radiomics models based on contrast-enhanced CT and a combined model incorporating clinical features for preoperative assessment of lymphovascular invasion(LVI) in colorectal cancer.

Methods

A retrospective study included 303 patients who underwent radical surgery for colorectal cancer at Shanghai General Hospital from January 2015 to December 2019. Patients were categorized into LVI-positive (n=122) and LVI-negative (n=181) groups. The cohort was randomly divided into a training set (n=212; 90 LVI-positive, 122 LVI-negative) and a test set (n=91; 32 LVI-positive, 59 LVI-negative) in a 7:3 ratio. Clinical predictors of LVI were identified using chi-square tests, Mann-Whitney U tests, and multivariate logistic regression. Two radiologists manually delineated 2D regions of interest (ROI) and 3D volumes of interest (VOI) on venous-phase contrast-enhanced CT images. Radiomic features were selected via Mann-Whitney U test, intraclass correlation coefficient(ICC) and least absolute shrinkage and selection operator(LASSO) regression. Based on selected features, support vector machine(SVM), logistic regression(LR), random forest(RF), and extreme gradient boosting(XGBoost) algorithms were employed to construct 2D, 3D radiomics models, and combined models. Model performance was assessed using receiver operating characteristic(ROC) curves, area under the curve(AUC), and the DeLong test. Calibration curves and decision curve analysis were used to evaluate diagnostic accuracy and clinical utility.

Results

N stage, perineural invasion, and tumor deposits were identified as independent clinical risk factors for LVI. Models built with the LR classifier demonstrated superior performance. The 2D radiomics model achieved AUC of 0.734 (95%CI: 0.660~0.798) and 0.683 (95%CI: 0.562~0.805) in the training and test sets, respectively. The 3D radiomics model yielded AUC of 0.784 (95%CI: 0.720~0.843) and 0.726 (95%CI: 0.602~0.830). The 2D combined model (integrating N stage, perineural invasion, and tumor deposits) achieved AUC of 0.843 (95%CI: 0.786~0.899) and 0.819 (95%CI: 0.723~0.905), while the 3D combined model showed AUC of 0.907 (95%CI: 0.865~0.943) and 0.874 (95%CI: 0.790~0.941). Calibration curves indicated comparable predictive accuracy between 2D and 3D models. Decision curve analysis revealed greater clinical net benefit for the 3D combined model. DeLong tests showed no significant performance differences between classifiers for 2D/3D models, but the 3D combined model outperformed the 2D radiomics model (P<0.05).

Conclusion

While 2D and 3D segmentation methods exhibit distinct advantages and limitations, the 3D combined model constructed with LR demonstrats superior predictive performance. Both 2D and 3D radiomics models enable effective preoperative prediction of LVI in colorectal cancer, offering valuable guidance for clinical decision-making.

图1 结直肠癌患者增强CT静脉期肿瘤感兴趣区域(绿色区域)勾画。1A:基于患者增强CT静脉期水平面的勾画,1B:ROI勾画,1C:VOI勾画
图2 LASSO影像组学特征筛选过程。2A、2B:通过10折交叉验证确定最佳λ值(2A:基于ROI勾画,2B:基于VOI勾画);2C、2D:LASSO特征筛选的收敛过程(2C:基于ROI勾画,2D:基于VOI勾画)
表1 训练组和测试组结直肠癌患者临床及病理特征比较[例(%)]
表2 LVI阳性和阴性组临床及病理特征比较[例(%)]
表3 结直肠癌脉管侵犯与临床及病理特征多因素Logistic回归分析
图3 筛选的影像组学特征。3A:基于ROI勾画,3B:基于VOI勾画
表4 不同机器学习算法的预测性能
图4 基于ROI和VOI构建LVI预测模型的ROC曲线。4A、4D:2D影像组学模型的ROC曲线(4A:训练集,4D:测试集);4B、4E:3D影像组学的ROC曲线(4B:训练集,4E:测试集);4C、4F:LR开发的预测模型的ROC曲线(4C:训练集,4F:测试集)
表5 LR算法开发的模型的Delong检验比较
图5 LR开发的LVI预测模型的校准曲线和决策曲线。5A:联合模型的校准曲线,5B:训练集决策曲线,5C:测试集决策曲线
[1]
中国医师协会结直肠肿瘤专业委员会. 结直肠癌腹膜转移诊治专家共识(2025版)[J]. 中华胃肠外科杂志, 2025, 28(5): 441-449.
[2]
Rönnow CF, Arthursson V, Toth E, et al. Lymphovascular infiltration, not depth of invasion, is the critical risk factor of metastases in early colorectal cancer: retrospective population-based cohort study on prospectively collected data, including validation[J]. Ann Surg, 2022, 275(1): e148-e154.
[3]
邱佳辉,蔡谦谦,杨彦,等. 神经脉管浸润联合肿瘤间质比对结直肠癌预后的预测价值[J]. 上海交通大学学报(医学版), 2022, 42(8): 1070-1080.
[4]
Zwager LW, Bastiaansen BAJ, Montazeri NSM, et al. Deep submucosal invasion is not an independent risk factor for lymph node metastasis in T1 colorectal cancer: a meta-analysis[J]. Gastroenterology, 2022, 163(1): 174-189.
[5]
Baxter NN, Kennedy EB, Bergsland E, et al. Adjuvant therapy for stage II colon cancer: ASCO guideline update[J]. J Clin Oncol, 2022, 40(8): 892-910.
[6]
Fan L, Mac MT, Frishberg DP, et al. Interobserver and intraobserver variability in evaluating vascular invasion in hepatocellular carcinoma[J]. J Gastroenterol Hepatol, 2010, 25(9): 1556-1561.
[7]
中华人民共和国国家卫生健康委员会. 中国结直肠癌诊疗规范(2023版)[J]. 中华消化外科杂志, 2023, 22(6): 667-698.
[8]
Urbaniec-Stompór J, Michalak M, Godlewski J. Correlating ultrastructural changes in the invasion area of colorectal cancer with CT and MRI imaging[J]. Int J Mol Sci, 2024, 25(18): 9905.
[9]
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.
[10]
Xu M, Zhou Y, Jin C, et al. Expectation maximisation pseudo labels[J]. Med Image Anal, 2024, 94: 103125.
[11]
Bimurzayeva A, Kim MJ, Ahn JS, et al. Three-dimensional body composition parameters using automatic volumetric segmentation allow accurate prediction of colorectal cancer outcomes[J]. J Cachexia Sarcopenia Muscle, 2024, 15(1): 281-291.
[12]
Zhu Y, Wei Y, Chen Z, et al. Different radiomics annotation methods comparison in rectal cancer characterisation and prognosis prediction: a two-centre study[J]. Insights Imaging, 2024, 15(1): 211.
[13]
赵颖燕,李伟伟,陶玲玲,等. 超声影像组学预测甲状腺乳头状癌颈部淋巴结跳跃性转移[J]. 中国超声医学杂志, 2025, 41(1): 1-3.
[14]
沈洁,张晔,金晓梅,等. 瘤内与瘤周CT影像组学模型结合临床及常规CT特征鉴别肺原位腺癌与微浸润性腺癌[J]. 中国医学影像技术, 2024, 40(6): 869-873.
[15]
Huang P, Li S, Deng Z, et al. Machine-learning based computed tomography radiomics nomgram for predicting perineural invasion in gastric cancer[J]. Curr Med Imaging, 2025, 21: e15734056323323.
[16]
Guiot J, Vaidyanathan A, Deprez L, et al. A review in radiomics: making personalized medicine a reality via routine imaging[J]. Med Res Rev, 2022, 42(1): 426-440.
[17]
Kocak B, Akinci D'Antonoli T, Mercaldo N, et al. METhodological RadiomICs Score (METRICS): a quality scoring tool for radiomics research endorsed by EuSoMII[J]. Insights Imaging, 2024, 15(1): 8.
[18]
Kocak B, Baessler B, Bakas S, et al. Check List for Evaluation of radiomics research (CLEAR): a step-by-step reporting guideline for authors and reviewers endorsed by ESR and EuSoMII[J]. Insights Imaging, 2023, 14(1): 75.
[19]
Li Y, Gu X, Yang L, et al. Prediction of lymphovascular invasion in esophageal squamous cell carcinoma by computed tomography-based radiomics analysis: 2D or 3D?[J]. Cancer Imaging, 2024, 24(1): 141.
[20]
Li Y, Liu Y, Liang Y, et al. Radiomics can differentiate high-grade glioma from brain metastasis: a systematic review and meta-analysis[J]. Eur Radiol, 2022, 32(11): 8039-8051.
[21]
Watzenboeck ML, Heidinger BH, Rainer J, et al. Reproducibility of 2D versus 3D radiomics for quantitative assessment of fetal lung development: a retrospective fetal MRI study[J]. Insights Imaging, 2023, 14(1): 31.
[22]
Zhang Y, Xu Z, Wu S, et al. Construction of 3D and 2D contrast-enhanced CT radiomics for prediction of CGB3 expression level and clinical prognosis in bladder cancer[J]. Heliyon, 2023, 9(9): e20335.
[23]
史宏伟,任继亮,袁瑛,等. 基于2D与3D分割的T2WI直方图分析在腮腺肿瘤鉴别中的对比研究[J]. 中国临床医学影像杂志, 2024, 35(6): 391-395.
[24]
苏亚英. 瘤内和瘤周动态对比增强MRI影像组学模型预测宫颈鳞癌同步放化疗反应的研究[D]. 张家口: 河北北方学院, 2023.
[25]
Zhang L, Diao B, Fan Z, et al. Radiomics for differentiating pancreatic mucinous cystic neoplasm from serous cystic neoplasm: systematic review and meta-analysis[J]. Acad Radiol, 2025, 32(5): 2679-2688.
[26]
LaValley MP. Logistic regression[J]. Circulation, 2008, 117(18): 2395-2399.
[27]
van Griethuysen JJM, Fedorov A, Parmar C, et al. Computational radiomics system to decode the radiographic phenotype[J]. Cancer Res, 2017, 77(21): e104-e107.
[1] 严征远, 张恒, 曹能琦, 方兴超, 陈大敏. 单孔+1腹腔镜结直肠癌根治切除术的有效性及安全性临床观察[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(06): 615-618.
[2] 杨春燕, 周晓苹. 机器人辅助技术在腹腔镜结直肠癌根治术中的研究进展[J/OL]. 中华普外科手术学杂志(电子版), 2025, 19(05): 584-588.
[3] 郑哲宇, 张磊, 张大伟, 潘卫东, 黄晓明. 全腹腔镜下ALPPS治疗结直肠癌肝转移的安全性和疗效[J/OL]. 中华肝脏外科手术学电子杂志, 2025, 14(05): 748-753.
[4] 鲁旭, 李华. 结直肠癌肝转移肝移植治疗研究进展[J/OL]. 中华肝脏外科手术学电子杂志, 2025, 14(04): 508-514.
[5] 关旭, 杨明. 中国微创手术的光辉历程与经自然腔道取标本手术的革新之路[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(05): 385-388.
[6] 中国医师协会外科医师分会, 中华医学会外科分会胃肠外科学组, 中华医学会外科分会结直肠外科学组, 中国抗癌协会大肠癌专业委员会, 中国医师协会结直肠肿瘤专业委员会, 中国临床肿瘤学会结直肠癌专家委员会, 中国医师协会外科医师分会结直肠外科医师委员会, 中国医师协会肛肠医师分会肿瘤转移委员会, 中华医学会肿瘤学分会结直肠肿瘤学组, 中国医疗保健国际交流促进会转移肿瘤治疗学分会, 中国医疗保健国际交流促进会结直肠病分会. 结直肠癌肝转移诊断和综合治疗指南(V 2025)[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(05): 398-411.
[7] 中国抗癌协会NOSES专业委员会, 中国抗癌协会大肠癌专业委员会. 肿瘤整合诊治技术指南·NOSES技术(结直肠癌部分)[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(05): 412-416.
[8] 赵南, 张明凯, Bhargava Divija, 赵世光, 张大明. 结直肠癌脑转移的临床特征与治疗策略进展[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(05): 427-435.
[9] 卢新泉, 刁德昌, 林佳鑫, 李洪明, 万进. "Henle干孤立化解剖"在腹腔镜右半结肠癌根治术中的应用[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(04): 375-379.
[10] 赵发宽, 向岩帕保, 杨家成, 梁伟纲, 张金源. 单孔+1腹腔镜下左半结肠癌根治术一例(附视频)[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(04): 380-384.
[11] 中国医师协会结直肠肿瘤专业委员会免疫学组. 结直肠癌免疫治疗专家共识(2025版)[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(04): 289-298.
[12] 中国医师协会结直肠肿瘤专业委员会术中放疗协作组. 结直肠癌术中放疗专家共识(2025版)[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(04): 299-306.
[13] 中国医师协会结直肠肿瘤专业委员会. 结直肠癌卵巢转移诊疗专家共识(2025版)[J/OL]. 中华结直肠疾病电子杂志, 2025, 14(04): 307-318.
[14] 鲁旭娟, 孔振兴. miR-146a靶向PTP1B对结直肠癌细胞增殖、迁移及侵袭的影响[J/OL]. 中华消化病与影像杂志(电子版), 2025, 15(04): 294-299.
[15] 李艳, 呼霞, 赵娟. 超声引导下腰方肌阻滞与竖脊肌平面阻滞对腹腔镜结直肠癌根治术后应激指标、炎症因子、T淋巴细胞亚群的影响[J/OL]. 中华消化病与影像杂志(电子版), 2025, 15(04): 393-397.
阅读次数
全文


摘要


AI


AI小编
你好!我是《中华医学电子期刊资源库》AI小编,有什么可以帮您的吗?