基于双模态超声的乳腺癌患者术前淋巴结转移负荷模型构建分析

杨骏宇; 沈吉; 沈思平; 李丹; 吴万波

doi:10.16766/j.cnki.issn.1674-4152.003471

基于双模态超声的乳腺癌患者术前淋巴结转移负荷模型构建分析

doi: 10.16766/j.cnki.issn.1674-4152.003471

1.
湖州市中心医院超声科, 浙江湖州 313000
2.
湖州市中心医院乳腺外科

基金项目:

浙江省医药卫生科技计划项目 2020ZH041

详细信息

通讯作者:
李丹，E-mail：175907922@qq.com

中图分类号: R737.9 R730.41
计量
- 文章访问数: 52
- HTML全文浏览量: 26
- PDF下载量: 0
- 被引次数: 0
出版历程
- 收稿日期: 2023-10-14
- 网络出版日期: 2024-05-29

Construction and analysis of preoperative lymph node metastasis load model of breast cancer patients based on dual-mode ultrasound

YANG Junyu¹,
SHEN Ji¹,
SHEN Siping¹,
LI Dan^{1
, ,},
WU Wanbo²

1.
Department of Ultrasound, Huzhou Central Hospital, Huzhou, Zhejiang 313000, China

摘要

摘要: 目的构建基于双模态超声的乳腺癌患者术前淋巴结转移负荷风险模型，以识别淋巴结高转移负荷人群，并为手术方案的制定提供参考。方法选取2021年5月—2023年5月湖州市中心医院接受手术治疗的乳腺癌患者168例作为研究对象。术前行双模态超声检查并根据淋巴结转移状态分为高转移负荷组62例(≥3个转移淋巴结)和低转移负荷组106例(≤2个转移淋巴结)。通过LASSO-logistic回归进一步筛选变量构建基于双模超声的乳腺癌患者术前淋巴结高转移负荷列线图预测模型并验证模型效能。结果经过LASSO回归分析10倍交叉验证最终筛选出脉管侵犯、导管扩张、Adler血流分级、弹性应变率4个代表性特征。多因素logistic分析显示：存在脉管侵犯(OR=2.250，95% CI: 1.012~5.002)、Adler血流分级为Ⅱ/Ⅲ级(OR=2.929，95% CI: 1.256~6.827)、导管扩张(OR=2.548，95% CI: 1.066~6.093)、弹性应变率高(OR=4.167，95% CI: 2.486~6.982)均为乳腺癌患者术前淋巴结转移高负荷危险因素(P < 0.05)。列线图显示预测乳腺癌患者术前淋巴结高转移负荷相关因素的C-index为0.834(95% CI: 0.771~0.898)且校正曲线显示实测值与预测值基本相符。结论基于常规二维超声、剪切波弹性成像超声图像构建的模型结合了多种模态的特点，使其信息互补，有助于提高乳腺癌术前淋巴结转移负荷定性及定位分析的准确性。
- 乳腺癌 /
- 双模态超声 /
- 淋巴结 /
- 转移负荷 /
- 预测模型
Abstract: Objective To construct a risk model using dual-mode ultrasound to identify breast cancer patients with a high lymph node metastasis load. This model provides references for the formulation of surgical protocols. Methods The study included 168 breast cancer patients who underwent surgical treatment at Huzhou Central Hospital between May 2021 and May 2023. Preoperative dual-mode ultrasonography was performed and the patients were divided into two groups based on the status of lymph node metastasis: high metastatic load group (62 cases, ≥3 metastatic lymph nodes) and low metastatic load group (106 cases, ≤2 metastatic lymph nodes). LASSO-logistic regression was used to screen variables and construct a prediction model for preoperative high metastatic load of lymph nodes in breast cancer patients based on dual-mode ultrasound. The effectiveness of the model was also verified. Results Four representative characteristics, namely vascular invasion, catheter dilation, Adler blood flow grading and elastic strain rate were selected through LASSO regression analysis with 10-fold cross-validation. Multivariate logistic analysis revealed that vascular invasion (OR=2.250, 95% CI 1.012-5.002), Adler blood flow grade Ⅱ/Ⅲ (OR=2.929, 95% CI: 1.256-6.827), catheter dilation (OR=2.548, 95% CI: 1.066-6.093) and high elastic strain rate (OR=4.167, 95% CI: 2.486-6.982) were identified as risk factors for preoperative high burden of lymph node metastasis in breast cancer patients (P < 0.05). The C-index for factors predicting preoperative high metastatic load of lymph nodes in breast cancer patients was 0.834 (95% CI: 0.771-0.898), and the calibration curve demonstrated good agreement between measured and predicted values. Conclusion The model constructed using conventional two-dimensional ultrasound and shear-wave elastography ultrasound images combines the characteristics of multiple modes to make their information complementary. This is helpful in improving the accuracy of qualitative and localization analysis of lymph node metastasis load before breast cancer.
- Breast cancer /
- Two-mode ultrasound /
- Lymph nodes /
- Transfer load /
- Prediction model

HTML全文

图 1 LASSO逻辑回归进行双模态超声特征筛选结果

注：A根据最小准则的1个标准误差，通过10倍交叉验证选择LASSO模型中的调优参数(λ)，以最小二项平均偏差的λ值用来选择特征。并在最小准则和1-SE准则在最优值处绘制虚线；B为10个超声特征的LASSO系数图谱，使用10倍交叉验证选择的值处绘制垂直线，选取4个非零系数为最优结果。

Figure 1. Results of dual-mode ultrasonic feature screening by LASSO logistic regression

下载: 全尺寸图片幻灯片

图 2 乳腺癌患者术前淋巴结高转移负荷列线图模型

Figure 2. Preoperative nomogram model of high metastatic load of lymph nodes in breast cancer patients

下载: 全尺寸图片幻灯片

图 3 乳腺癌患者术前淋巴结高转移负荷预测模型校正曲线

Figure 3. Correction curve of prediction model of high metastatic load of lymph nodes in breast cancer patients

下载: 全尺寸图片幻灯片

表 1 2组乳腺癌患者一般资料比较

Table 1. Comparison of general data of 2 groups of breast cancer patients

项目	高转移负荷组 (n=62)	低转移负荷组 (n=106)	统计量	P值
年龄(x±s, 岁)	48.26±8.95	47.65±9.02	0.425^a	0.672
家族史[例(%)]	12(19.35)	20(18.87)	0.006^b	0.938
绝经[例(%)]	21(33.87)	30(28.30)	0.574^b	0.449
分子分型[例(%)]			1.129^b	0.771
Luminal A型	20(32.26)	35(33.02)
Luminal B型	15(24.19)	32(30.19)
HER2阳性型	16(25.81)	25(23.58)
三阴性型	11(17.74)	14(13.21)
肿瘤部位[例(%)]			0.258^b	0.612
左侧	37(59.68)	59(55.66)
右侧	25(40.32)	47(44.34)
外上象限[例(%)]	48(77.42)	86(81.13)	0.334^b	0.563
组织学分级[例(%)]			0.216^b	0.898
Ⅰ级	21(33.87)	35(33.02)
Ⅱ级	26(41.94)	48(45.28)
Ⅲ级	15(24.19)	23(21.70)
注：^a为t值，^b为χ²值。

下载: 导出CSV

表 2 2组乳腺癌患者双模态超声参数比较

Table 2. Comparison of dual-mode ultrasound parameters in 2 groups of breast cancer patients

特征	高转移负荷组 (n=62)	低转移负荷组 (n=106)	统计量	P值
肿瘤最大径(x±s，cm)	3.38±1.06	3.03±0.97	2.130^a	0.035
回声[例(%)]			1.273^b	0.259
均匀	36(58.06)	52(49.06)
不均匀	26(41.94)	54(50.94)
形态[例(%)]			0.736^b	0.391
规则	16(25.81)	34(32.08)
不规则	46(74.19)	72(67.92)
边界[例(%)]			4.000^b	0.046
清晰	30(48.38)	68(64.15)
不清晰	32(51.62)	38(35.85)
纵横比[例(%)]			0.353^b	0.553
< 1	52(83.87)	85(80.19)
≥ 1	10(16.13)	21(19.81)
微钙化[例(%)]			5.349^b	0.021
有	36(58.06)	42(39.62)
无	26(41.94)	64(60.38)
微分叶[例(%)]			6.553^b	0.010
有	27(43.55)	26(24.53)
无	35(56.45)	80(75.47)
脉管侵犯[例(%)]			5.861^b	0.015
有	27(43.55)	27(25.47)
无	35(56.45)	79(74.53)
Adler血流分级[例(%)]			4.012^b	0.045
0/I级	24(38.71)	58(54.72)
Ⅱ/Ⅲ级	38(61.29)	48(45.28)
长短径比值[例(%)]			4.331^b	0.037
≥2	18(29.03)	48(45.28)
< 2	44(70.97)	58(54.72)
导管扩张[例(%)]			4.665^b	0.031
有	23(37.10)	23(21.70)
无	39(62.90)	83(78.30)
弹性应变率(x±s，%)	2.49±0.34	2.11±0.29	7.371^a	< 0.001
超声弹性评分(x±s，分)	2.17±0.38	2.90±0.30	12.949^a	< 0.001
注：^a为t值，^b为χ²值。

下载: 导出CSV

表 3 乳腺癌患者术前淋巴结转移负荷风险多因素logistic回归分析

Table 3. Multivariate Logistic regression analysis of risk of preoperative lymph node metastatic load in breast cancer patients

变量	B	SE	Waldχ²	P值	OR值	95% CI
Adler血流分级	1.075	0.432	6.191	0.013	2.929	1.256~6.827
脉管侵犯	0.811	0.408	3.954	0.047	2.250	1.012~5.002
导管扩张	0.935	0.445	4.422	0.035	2.548	1.066~6.093
弹性应变率	1.427	0.263	29.353	< 0.001	4.167	2.486~6.982

下载: 导出CSV

参考文献(17)

[1]	KATSURA C, OGUNMWONYI I, KANKM H K, et al. Breast cancer: presentation, investigation and management[J]. Br J Hosp Med(Lond), 2022, 83(2): 1-7.
[2]	MARINO M A, AVENDNO D, ZAPATA P, et al. Lymph node imaging in patients with primary breast cancer: concurrent diagnostic tools[J]. Oncologist, 2020, 25(2): e231-e242. doi: 10.1634/theoncologist.2019-0427
[3]	GU J, TONG T, XU D, et al. Deep learning radiomics of ultrasonography for comprehensively predicting tumor and axillary lymph node status after neoadjuvant chemotherapy in breast cancer patients: a multicenter study[J]. Cancer, 2023, 129(3): 356-366. doi: 10.1002/cncr.34540
[4]	SONG B I. A machine learning-based radiomics model for the prediction of axillary lymph-node metastasis in breast cancer[J]. Breast Cancer, 2021, 28(3): 664-671. doi: 10.1007/s12282-020-01202-z
[5]	GRADISHAR W J, MORAN M S, ABRAHAM J, et al. Breast cancer, version 3.2022, NCCN clinical practice guidelines in oncology[J]. J Natl Compr Canc Netw, 2022, 20(6): 691-722. doi: 10.6004/jnccn.2022.0030
[6]	中国抗癌协会乳腺癌专业委员会. 中国抗癌协会乳腺癌诊治指南与规范(2019年版)[J]. 中国癌症杂志, 2019, 29(8): 609-679. https://www.cnki.com.cn/Article/CJFDTOTAL-ZGAZ202312004.htm
[7]	American Institute of Ultrasound in Medicine, American Society of Breast Surgeons. AIUM practice guideline for the performance of a breast ultrasound examination[J]. J Ultrasound Med, 2009, 28(1): 105-109. doi: 10.7863/jum.2009.28.1.105
[8]	LAOHAWIRIYAKAMOL S, MHTTNOBON S, PUTTAWIBUL P. Intraoperative molecular analysis of total tumor load in sentinel lymph node: a predictor of axillary status in early breast cancer[J]. Asian Pac J Cancer Prev, 2022, 23(1): 349-354. doi: 10.31557/APJCP.2022.23.1.349
[9]	SHI J Y, WANG X, DING G Y, et al. Exploring prognostic indicators in the pathological images of hepatocellular carcinoma based on deep learning[J]. Gut, 2021, 70(5): 951-961. doi: 10.1136/gutjnl-2020-320930
[10]	ZHU W, XIE L, HAN J, et al. The application of deep learning in cancer prognosis prediction[J]. Cancers (Basel), 2020, 12(3): 603. doi: 10.3390/cancers12030603
[11]	王莹, 崔颖颖, 王鑫慧, 等. 基于磁共振超短回波时间序列的影像组学模型预测非小细胞肺癌淋巴结转移[J]. 磁共振成像, 2023, 14(3): 17-20, 41. https://www.cnki.com.cn/Article/CJFDTOTAL-CGZC202303004.htm WANG Y, CUI Y Y, WANG X H, et al. 3D-ultrashort echo time MRI-based radiomics model facilitates the assessment of lymph node metastasis in non-small cell lung cancer[J]. Chinese Journal of Magnetic Resonance Imaging, 2023, 14(3): 17-20, 41. https://www.cnki.com.cn/Article/CJFDTOTAL-CGZC202303004.htm
[12]	胡文娟, 宋彬, 谢晓利, 等. 多参数磁共振组学预测甲状腺乳头状癌淋巴结转移的价值[J]. 放射学实践, 2023, 38(7): 863-867. https://www.cnki.com.cn/Article/CJFDTOTAL-FSXS202307009.htm HU W J, SING B, XIE X L, et al. Radiomics based on multiparametric MRI for preoperative prediction of lymph node metastasis in papillary thyroid carcinoma[J]. Radiologic Practice, 2023, 38(7): 863-867. https://www.cnki.com.cn/Article/CJFDTOTAL-FSXS202307009.htm
[13]	姚晓倩, 洪敏萍, 蔡宏杰, 等. 基于超声影像组学术前预测浸润性乳腺癌患者腋窝淋巴结状态[J]. 现代实用医学, 2023, 35(1): 116-119. https://www.cnki.com.cn/Article/CJFDTOTAL-NBYX202301037.htm YAO X Q, HONG M P, CAI H J, et al. Prediction of axillary lymph node status in patients with invasive breast cancer based on pre-academic ultrasound imaging group[J]. Modern Practical Medicine, 2023, 35(1): 116-119. https://www.cnki.com.cn/Article/CJFDTOTAL-NBYX202301037.htm
[14]	刘晗, 徐楠, 吴杰, 等. 基于灰阶超声联合剪切波弹性成像的影像组学模型诊断乳腺癌腋窝淋巴结转移的临床价值[J]. 临床超声医学杂志, 2023, 25(4): 277-283. https://www.cnki.com.cn/Article/CJFDTOTAL-LCCY202304007.htm LIU H, XU N, WU J, et al. Clinical value of bimodal radiomics model based on gray-scale ultrasound combined with shear wave elastography in the diagnosis of axillary lymph node metastasis of breast cancer[J]. Journal of Clinical Ultrasound in Medicine, 2023, 25(4): 277-283. https://www.cnki.com.cn/Article/CJFDTOTAL-LCCY202304007.htm
[15]	ZHU Y, ZHUO W, JIA X H, et al. Preoperative axillary ultrasound in the selection of patients with a heavy axillary tumor burden in early-stage breast cancer: what leads to false-positive results?[J]. J Ultrasound Med, 2018, 37(6): 1357-1365. doi: 10.1002/jum.14545
[16]	PATEL B K, SMREEN N, ZHUO Y, et al. MR elastography of the breast: evolution of technique, case examples, and future directions[J]. Clin Breast Cancer, 2021, 21(1): e102-e111. doi: 10.1016/j.clbc.2020.08.005
[17]	陈佳佳, 孟利伟, 李星云. 乳腺癌超声特征和ER、PR、CerbB-2、Ki-67阳性表达的相关性研究[J]. 中华全科医学, 2021, 19(10): 1721-1724. doi: 10.16766/j.cnki.issn.1674-4152.002151 CHEN J J, MENG L W, LI X Y. Relationship between ultrasound features and expression of ER, PR, CerbB-2 and Ki-67 in breast cancer[J]. Chinese Journal of General Practice, 2021, 19(10): 1721-1724. doi: 10.16766/j.cnki.issn.1674-4152.002151