The clinical-MRI nomogram model can effectively predict lymphatic vascular infiltration of cervical cancer
-
摘要:
目的 基于临床-MRI影像组学的列线图模型在预测宫颈癌淋巴脉管浸润中的价值。 方法 回顾性分析2019年1月~2023年11月于蚌埠医科大学第一附属医院术前行MRI检查且术后病理证实为宫颈癌的患者168例。收集患者的临床和影像资料，按照7：3的比例随机分为训练集（n=112）和验证集（n=56）。通过单-多因素Logistic回归分析筛选与宫颈癌淋巴脉管浸润相关的临床独立危险因素；分别于T2WI和T1WI增强序列矢状位手动勾画感兴趣区，提取瘤内、瘤周及瘤内+瘤周影像组学特征，通过对影像特征降维并筛选最优特征构建影像组学模型；结合临床预测因子与影像组学评分构建列线图模型。采用ROC曲线下面积、校准曲线、决策曲线分析评价模型的预测效能。 结果 两组中性粒细胞与淋巴细胞计数比值及淋巴结是否转移的差异有统计学意义（P＜0.05），线图模型的预测效能最佳，其曲线下面积在训练集和验证集分别为0.932（95% CI：0.862~0.984）、0.896（95% CI：0.803~0.990）显著高于瘤内、瘤周影像组学模型和临床模型。 结论 本研究构建的列线图模型在预测宫颈癌淋巴脉管浸润方面具有较高的诊断性能，可以术前为临床决策提供重要指导。 Abstract:Objective To evaluate the value of clinically-MRI nomogram model in predicting lymphatic vasculature infiltration of cervical cancer. Methods A retrospective analysis was performed on 168 patients who underwent preoperative MRI examination and were pathologically confirmed as cervical cancer in the First Affiliated Hospital of Bengbu Medical University from January 2019 to November 2023. Clinical and imaging data of patients were collected and randomly divided into two groups, the training set (n=112) and the validation set (n=56), according to a ratio of 7:3. The clinical independent risk factors associated with lymphatic vascular infiltration of cervical cancer were screened by uni-multivariate Logistic regression analysis. The regions of interest were manually delineated in sagittal position of enhanced sequences on T2WI and T1WI respectively, and the intratumoral, peritumoral and intratumoral + peritumoral imaging features were extracted. The imaging model was constructed by dimensionality reduction of the image features and selection of the optimal features. A nomogram model was constructed by combining clinical predictors and imaging omics scores. The area under ROC curve, calibration curve and decision curve were used to analyze and evaluate the prediction efficiency of the model. Results The ratio of neutrophil to lymphocyte count and lymph node metastasis between the two groups were statistically significant (P<0.05). The graph model had the best prediction performance, and the area under the curve was 0.932 (95% CI: 0.862-0.984) and 0.896 (95% CI: 0.803-0.990) respectively, which was significantly higher than that of the tumor. Conclusion The Nomogram model established in this study has high diagnostic performance in predicting lymphatic vascular infiltration of cervical cancer, and can provide important guidance for clinical decision-making before surgery. -
Key words:
- cervical cancer /
- lymphatic vessel infiltration /
- intratumoral /
- peritumor /
- imaging omics /
- magnetic resonance /
- Nomogram
-
图 1 瘤内及瘤周ROI
Figure 1. ROI of intratumoural and peritumoural. The green filled area was the intratumoural ROI and the surrounding red ring-like area was the peritumoural ROI.
图 2 利用LASSO算法筛选出的12个最优影像学特征
Figure 2. The 12 optimal imaging features selected by LASSO algorithm.
图 3 不同模型在训练组和验证组中的ROC曲线
Figure 3. ROC of different models in training group and test group. A: Training group; B: Test group.
图 5 训练集（A）及测试集（B）的校准曲线
Figure 5. Calibration curves of training set (A) and test set (B).
表 1 宫颈癌患者基本临床病理资料比较
Table 1. Comparison of basic clinicopathological data of patients with cervical cancer
Variable Training group(n=112) Test group(n=56) LVSI(+)(n=34) LVSI(-)(n=78) P LVSI(+)(n=18) LVSI(-)(n=38) P Age [year, M(P25, P75)] 54.50(47.75, 62.75) 52.00(44.75, 58.25) 0.382 53.50(43.75, 59.25) 54.00(50.75, 60.00) 0.450 Menstrual status [n(%)] 0.163 0.460 Menstruation 13(38.24) 37(47.44) 8(44.44) 13(34.21) Menopause 21(61.76) 41(52.56) 10(55.56) 25(65.79) SCC-AG (μg/L, Mean±SD) 7.70±12.19 7.17±12.43 0.834 8.21±12.79 3.54±4.32 0.147 NLR (Mean±SD) 2.95±1.76 2.09±0.94 0.010 3.15±1.78 2.16±1.00 0.037 PLR [M(P25, P75)] 9.02(6.70, 14.55) 7.33(6.01, 9.57) 0.018 12.13(8.30, 14.61) 7.49(5.51, 9.62) 0.003 N [×109/L, M(P25, P75)] 68.35(57.95, 73.70) 3.20(2.67, 4.21) 0.002 61.45(51.27, 68.33) 59.30(54.10, 66.63) 0.623 L [×109/L, M(P25, P75)] 26.90(18.40, 38.33) 30.70(25.70, 36.90) 0.083 22.50(18.18, 27, 18) 30.60(25.05, 36.18) 0.002 Hb [×109/L, M(P25, P75)] 123.50(111.75, 136.00) 120.00(105.75, 130.25) 0.288 121.50(109.50, 131.25) 126.00(116.50，130.25) 0.380 LY [×109/L, M(P25, P75)] 1.54(1.22, 2.01) 1.74(1.41, 2.07) 0.120 1.62(1.23, 2.02) 1.56(1.32，1.87) 0.916 MO [×109L, Mean±SD] 0.48±0.25 0.39±013 0.670 0.51±0.25 1.22±5.13 0.559 ALB [g/L, M(P25, P75)] 43.25(41.25, 44.40) 43.10(40.90, 45.25) 0.967 42.85(38.45, 45.20) 126.00(116.50, 130, 25) 0.236 PLT [L, M(P25, P75)] 238.50(222.75, 277, 75) 240.50(187.00, 279.25) 0.046 251.00(223.00, 278.00) 232.00(189.00, 275.50) 0.343 Depth of invasion [n(%)] 0.011 0.928 ＜1/2 2(5.88) 21(26.92) 5(27.78) 11(28.95) ≥1/2 32(94.12) 57(73.08) 13(72.22) 27(71.05) Pathological classification [n(%)] 0.732 0.220 Squamous cell carcinoma 32(94.12) 72(92.31) 18(100.00) 35(92.11) Adenocarcinoma 2(5.88) 6(7.69) 0(0.00) 3(7.89) LNM [n(%)] < 0.001 0.080 Yes 18(52.94) 15(19.23) 9(50.00) 28(73.68) No 16(47.06) 63(80.77) 9(50.00) 10(26.32) SCC-AG: Squamous cell carcinoma antigen; NLR: Neutrophil-to-lymphocyte ratio; PLR: Platelet-to-lymphocyte ratio; N: Neutrophil count; L: Lymphocyte; Hb: Hemoglobin; LY: lyLmphocyte count; MO: Monocyte count; ALB: Albumin; PLT: Blood platelet; LNM: Lymphatic metastasis. 
下载: 导出CSV
表 2 单-多因素Logistic回归分析筛选LVSI临床危险因素
Table 2. Univariate and multivariate Logistic regression analysis for screening clinical risk factors of LVSI.
Variable Univariate Logistic regression Multivariate Logistic regression B OR (95% CI) P B OR (95% CI) P Age 0.001 1.001(0.972, 1.031) 0.972 Menstrual status -0.251 0.778(0.401, 1.510) 0.458 SCC-AG 0.014 1.014(0.987, 1.043) 0.315 NLR 0.461 1.585(1.214, 2.069) 0.001 0.894 2.444(1.008, 5.925) 0.048 PLR 0.112 1.119(1.045, 1.198) 0.001 -0.041 0.959(0.830, 1.109) 0.575 N 0.047 1.048(1.013, 1.083) 0.006 -0.097 0.908(0.804, 1.024) 0.117 L -0.058 0.944(0.909, 0.980) 0.003 0.029 0.971(0.856, 1.102) 0.649 HB -0.052 0.949(0.868, 1.038) 0.256 LY -0.345 0.708(0.366, 1.369) 0.305 MO -0.040 0.961(0.790, 1.168) 0.687 AlB 0.003 1.003(0.983, 1.023) 0.769 PlT 0.002 1.002(0.997, 1.007) 0.547 Depth of invasion -0.896 0.408(0.167, 0.999) 0.050 Pathological classification 0.743 2.103(0.438, 10.093) 0.353 LNM -1.369 0.254(0.126, 0.513) < 0.001 -1.143 0.319(0.150, 0.677) 0.003
下载: 导出CSV
表 3 不同模型预测效能比较
Table 3. Comparison of prediction performance among different models
Models Training group (n=112) Test group (n=56) AUC (95% CI) Sensitivity Specificity AUC (95% CI) Sensitivity Specificity Clinical characteristic 0.762(0.665, 0.860) 0.735 0.679 0.726(0.594, 0.858) 0.944 0.500 Tumor 0.795(0.692, 0.898) 0.676 0.923 0.756(0.606, 0.906) 0.667 0.824 Peritumor 0.729(0.619, 0.839) 0.618 0.808 0.672(0.531, 0.813) 0.556 0.842 Tumor+3 mm 0.827(0.733, 0.920) 0.824 0.821 0.804(0.692, 0.916) 0.879 0.684 Nomogram 0.932(0.862, 0.984) 0.912 0.897 0.896(0.803, 0.990) 0.722 0.947
下载: 导出CSV
-
[1] Haldorsen IS, Lura N, Blaakær J, et al. What is the role of imaging at primary diagnostic work-up in uterine cervical cancer?[J]. Curr Oncol Rep, 2019, 21(9): 77. doi: 10.1007/s11912-019-0824-0 [2] Wang SX, Liu XW, Wu Y, et al. Habitat-based radiomics enhances the ability to predict lymphovascular space invasion in cervical cancer: a multi-center study[J]. Front Oncol, 2023, 13: 1252074. doi: 10.3389/fonc.2023.1252074 [3] Weyl A, Illac C, Lusque A, et al. Prognostic value of lymphovascular space invasion in early-stage cervical cancer[J]. Int J Gynecol Cancer, 2020, 30(10): 1493-9. doi: 10.1136/ijgc-2020-001274 [4] Li KX, Sun HZ, Lu ZM, et al. Value of[18F]FDG PET radiomic features and VEGF expression in predicting pelvic lymphatic metastasis and their potential relationship in early-stage cervical squamous cell carcinoma[J]. Eur J Radiol, 2018, 106: 160-6. doi: 10.1016/j.ejrad.2018.07.024 [5] 张婧, 温宏武. 早期宫颈腺癌淋巴脉管间隙浸润相关临床病理因素及生存分析[J]. 现代妇产科进展, 2020, 29(11): 819-24, 829. [6] Wang Y, Chen X, Pu H, et al. Roles of DWI and T2-weighted MRI volumetry in the evaluation of lymph node metastasis and lymphovascular invasion of stage IB-ⅡA cervical cancer[J]. Clin Radiol, 2022, 77(3): 224-30. doi: 10.1016/j.crad.2021.12.011 [7] Zhang ZH, Wan XJ, Lei XY, et al. Intra-and peri-tumoral MRI radiomics features for preoperative lymph node metastasis prediction in early-stage cervical cancer[J]. Insights Imaging, 2023, 14(1): 65. doi: 10.1186/s13244-023-01405-w [8] 崔雅琼, 黄刚, 王莉莉, 等. 基于临床-多参数磁共振影像组学特征预测宫颈癌脉管浸润和预后的研究[J]. 磁共振成像, 2023, 14(2): 73-82. [9] 郭丽美, 王峻, 武文奇, 等. 合成MRI定量参数评估宫颈鳞癌淋巴脉管间隙浸润的初步研究[J]. 磁共振成像, 2023, 14(11): 103-7, 120. [10] 江瑞信, 叶浩翊, 郭栋华, 等. MR DWI在宫颈癌病理分型及分级中的应用研究[J]. 现代医用影像学, 2020, 29(5): 820-3. [11] 王露萍, 李丽, 李刚, 等. 宫颈鳞癌DCE-MRI定量参数与临床病理分级的相关性[J]. 中国计划生育学杂志, 2023, 31(10): 2511-4, 2520-1. [12] 吴佩琪, 刘于宝, 陈祉妍, 等. 基于MRI的瘤周影像组学在肿瘤研究中的应用进展[J]. 分子影像学杂志, 2023, 46(1): 164-9. doi: 10.12122/j.issn.1674-4500.2023.01.32 [13] Wang ZL, Zhang S, Ma YF, et al. A nomogram prediction model for lymph node metastasis in endometrial cancer patients[J]. BMC Cancer, 2021, 21(1): 748. doi: 10.1186/s12885-021-08466-4 [14] Lei HF, Xu SX, Mao XD, et al. Systemic immune-inflammatory index as a predictor of lymph node metastasis in endometrial cancer[J]. J Inflamm Res, 2021, 14: 7131-42. doi: 10.2147/JIR.S345790 [15] Li ZC, Li HL, Wang SY, et al. MR-based radiomics nomogram of cervical cancer in prediction of the lymph-vascular space invasion preoperatively[J]. J Magn Reson Imaging, 2019, 49(5): 1420-6. doi: 10.1002/jmri.26531 [16] 冯永志. 术前血清炎症指标及肿瘤影像特征与肝癌病理分级的相关性研究[J]. 基层医学论坛, 2023, 27(13): 82-5. [17] 李伟兰, 刘芳. 炎症生化指标与宫颈癌临床病理特征的关系探讨[J]. 华北理工大学学报: 医学版, 2023, 25(4): 260-4. [18] 于秋宇. 术前外周血炎症和营养指标与ⅠB-ⅡA期宫颈癌临床病理特征和预后的关系[D]. 长春: 吉林大学, 2023. [19] Chen SB, Zhang L, Yan GY, et al. Neutrophil-to-lymphocyte ratio is a potential prognostic biomarker in patients with ovarian cancer: a meta-analysis[J]. Biomed Res Int, 2017, 2017: 7943467. [20] Du JQ, Zhang F, Wang CQ, et al. Effects of peripheral blood neutrophil/lymphocyte ratio levels and their changes on the prognosis of patients with early cervical cancer[J]. Front Oncol, 2023, 13: 1139809. doi: 10.3389/fonc.2023.1139809 [21] Guo Q, Gao YF, Lin YY, et al. A nomogram of preoperative indicators predicting lymph vascular space invasion in cervical cancer[J]. Arch Gynecol Obstet, 2024, 309(5): 2079-87. doi: 10.1007/s00404-024-07385-6 [22] Wenzel HHB, van Kol KGG, Nijman HW, et al. Cervical cancer with ≤5 mm depth of invasion and > 7 mm horizontal spread-Is lymph node assessment only required in patients with LVSI?[J]. Gynecol Oncol, 2020, 158(2): 282-6. doi: 10.1016/j.ygyno.2020.04.705 [23] 陈刚, 李娜. 淋巴脉管间隙浸润与早期宫颈癌患者术后生存率和淋巴结转移的相关性分析[J]. 临床医学研究与实践, 2022, 7(35): 17-20. [24] Pérez-Morales J, Tunali I, Stringfield O, et al. Peritumoral and intratumoral radiomic features predict survival outcomes among patients diagnosed in lung cancer screening[J]. Sci Rep, 2020, 10(1): 10528. doi: 10.1038/s41598-020-67378-8 [25] Cui LP, Yu T, Kan YY, et al. Multi- parametric MRI-based peritumoral radiomics on prediction of lymph-vascular space invasion in early-stage cervical cancer[J]. Diagn Interv Radiol, 2022, 28(4): 312-21. doi: 10.5152/dir.2022.20657 [26] Hu F, Zhang YH, Li M, et al. Preoperative prediction of microvascular invasion risk grades in hepatocellular carcinoma based on tumor and peritumor dual-region radiomics signatures[J]. Front Oncol, 2022, 12: 853336. doi: 10.3389/fonc.2022.853336 [27] Zeng SY, Yang P, Xiao SM, et al. Development and validation of prognostic nomographs for patients with cervical cancer: seer-based Asian population study[J]. Sci Rep, 2024, 14(1): 7681. doi: 10.1038/s41598-024-57609-7 [28] Rafiee A, Mohammadizadeh F. Association of lymphovascular space invasion (LVSI) with histological tumor grade and myometrial invasion in endometrial carcinoma: a review study[J]. Adv Biomed Res, 2023, 12: 159. [29] 程思佳, 翟晓阳, 周仕豪, 等. 基于术前MRI特征及定量参数列线图预测乳腺癌患者淋巴血管侵犯[J]. 磁共振成像, 2024, 15(5): 111-8. [30] 马晓亮, 沈敏花, 马凤华, 等. 多参数MRI组学列线图模型预测子宫内膜样腺癌淋巴血管间隙侵犯[J]. 复旦学报: 医学版, 2024, 51(3): 306-14, 322. doi: 10.3969/j.issn.1672-8467.2024.03.003 [31] Chen JY, Wang XR, Lv HY, et al. Development and external validation of a clinical-radiomics nomogram for preoperative prediction of LVSI status in patients with endometrial carcinoma[J]. J Cancer Res Clin Oncol, 2023, 149(15): 13943-53. doi: 10.1007/s00432-023-05044-y -
点击查看大图
图(6) / 表(3)
计量
- 文章访问数: 64
- HTML全文浏览量: 27
- PDF下载量: 11
- 被引次数: 0
