The correlation between mammographic density and the clinical and pathological features of breast cancer based on deep learning techniques

LIN Xiaohui; XU Mei; MA Jie

doi:10.19732/j.cnki.2096-6210.2020.03.008

您当前的位置：

首页 >

文章列表页 >

The correlation between mammographic density and the clinical and pathological features of breast cancer based on deep learning techniques

更新时间：2025-12-15

- The correlation between mammographic density and the clinical and pathological features of breast cancer based on deep learning techniques
- Vol. 29, Issue 3, Pages: 250-257(2020)
- 作者机构：
- 作者简介：
- 基金信息：
- DOI：10.19732/j.cnki.2096-6210.2020.03.008
  CLC：
- Published Online：28 June 2020，
  
  Published：28 June 2020
- 稿件说明：
移动端阅览
林小慧,许梅,马捷. 基于深度学习技术探讨乳腺X线密度与乳腺癌临床及病理的相关性[J]. 肿瘤影像学, 2020, 29(3): 250-257 https://doi.

org/10.19732/j.cnki.2096-6210.2020.03.008
林小慧,许梅,马捷. 基于深度学习技术探讨乳腺X线密度与乳腺癌临床及病理的相关性[J]. 肿瘤影像学, 2020, 29(3): 250-257 https://doi. DOI： 10.19732/j.cnki.2096-6210.2020.03.008.

org/10.19732/j.cnki.2096-6210.2020.03.008 DOI：

摘要

目的：

基于深度学习技术探讨乳腺癌患者乳腺X线密度（mammographic density，MD）与临床及病理学指标的相关性，以期为个体化治疗及预后评估提供影像学依据。

方法：

回顾并分析2019年412月于深圳市人民医院经病理学检查证实的205例乳腺癌患者的临床、病理及术前乳腺X线摄影检查资料。根据第5版乳腺影像报告和数据系统（Breast Imaging Reporting and Data System，BI-RADS）中乳腺构成分类的标准，采用深度学习技术辅助放射科医师对MD进行评估，将a型（脂肪型）和b型（散在纤维腺体型）乳腺归为非致密组，c型（不均匀致密型）和d型（极度致密型）乳腺归为致密组，采用

检验比较两组间临床（年龄、绝经状态、家族史）及病理学（肿瘤大小、组织学分级、淋巴结状态、神经或脉管侵犯情况、病理学类型及各免疫组织化学指标）资料的差异性。采用Spearman等级相关分析MD与年龄组、孕激素受体（progesterone receptor，PR）表达状态的相关性。

结果：

205例患者中，非致密组56例，致密组149例。两组间年龄、绝经状态的差异有统计学意义（

＜0.001），年龄组小、绝经前女性患者MD高于年龄组大和绝经后，而家族史的差异无统计学意义（

＞0.05）。两组间PR表达状态差异有统计学意义（

＜0.05），致密组PR阳性表达率（77.85%，116/149）高于非致密组（60.71%，34/56），而肿瘤大小、组织学分级、淋巴结状态、是否神经或脉管侵犯、病理学类型、雌激素受体（estrogen receptor，ER）、人表皮生长因子受体2（human epidermal growth factor receptor 2，HER2）及Ki-67表达情况的差异均无统计学意义（

均＞0.05）；Spearman等级相关分析显示，MD与年龄呈显著负相关（

=-0.481，

＜0.001），与PR表达呈正相关（

=0.196，

＜0.05）。

结论：

MD与年龄、绝经状态及PR表达有一定的相关性，对乳腺癌预后具有潜在的预测作用。

Abstract

Objective:

To explore the correlation between mammographic density (MD) and the clinical and pathological metrics in breast cancer patients based on deep learning technology

in order to provide imaging evidence for individualized treatment and prognosis evaluation.

Methods:

The clinical

pathological and reoperative mammographic data of 205 cases of breast cancer confirmed by pathology from Apr. 2019 to Dec. 2019 were retrospectively analyzed in Shenzhen Peoples Hospital. According to the criteria for categories of breast composition in the 5th edition of Breast Imaging Reporting and Data System (BI-RADS)

the D was evaluated by radiologist assisted by deep learning system. Type a (almost entirely fat) and type b (scattered fibroglandular dense) were included in non-dense group

type c (heterogeneously dense) and type d (extremely dense) were included in dense group.

test was used to compare the differences in clinical (age

menopausal status

family history) and pathological (tumor size

histological grade

lymph node status

nerve or vascular invasion

pathological type and immunohistochemical parameters) metrics between the two groups. Spearman test was used to analyze the correlation between MD

age and progesterone eceptor (PR) state.

Results:

Among the 205 patients

56 were in the non-dense group and 149 in the dense group. The differences in age and menopausal status between the two groups were statistically significant (

＜0.001)

the MD of the younger age group and premenopausal women was higher than that it in the older age group and post-menopausal women

while the differences in family history were not statistically significant (

＞0.05).There were tatistical differences in progesterone receptor (PR) status between the two groups (

＜0.05)

PR positive expression rate of dense group (77.85%

116/149) was higher than the non-dense group (60.71%

34/56)

and the tumor size

histological grade

lymp

h node status

neural or vascular invasion status

pathological ype

estrogen receptor (ER) and human epidermal growth factor receptor 2 (HER2) and Ki-67 showed no significant differences (

＞0.05); Spearman test howed MD was significantly negatively correlated with age (

=-0.481

＜0.001) and positively correlated with PR expression (

=0.196

＜0.05).

Conclusion:

The MD is correlated with age

menopausal status and PR status

and has a potential role in predicting the prognosis of breast cancer.

关键词

Keywords

references

Views

1696

下载量

CNKI被引量

Alert me when the article has been cited

提交

Tools

Publicity Resources

The diagnostic value of MRI in solid papillary carcinoma

Radiomics in Breast Cancer

Analysis of the diagnostic value of conventional ultrasound and mammography in different types of breast masses

Clinical value of calcification afterneoadjuvant chemotherapy in breast cancer

Retrospective differential diagnosis of suspicious calcification on mammography of breast fibroadenoma and invasive breast cancer　

Related Author

LIU Shijie

WANG Lijun

LUO Ran

GUAN Wenbin

WANG Dengbin 　nbsp

XIE Tianwen

PENG Weijun

CHEN Jing

Related Institution

Department of Radiology, Xinhua Hospital, Shanghai Jiao Tong University School of Medicine

Department of Pathology, Xinhua Hospital, Shanghai Jiao Tong University School of Medicine

Department of Ultrasound, Shanghai General Hospital, Shanghai Jiao Tong University

Department of Radiology, Shanghai General Hospital, Shanghai Jiao Tong University

Department of Radiology, Liaocheng People&rsquo

AI问答

⁰