|
| 论文编号: |
|
| 第一作者所在部门: |
|
| 中文论文题目: |
Spectral CT based radiomics signature: A potential biomarker for preoperative prediction of lymph node metastasis in breast cancer
|
| 英文论文题目: |
Spectral CT based radiomics signature: A potential biomarker for preoperative prediction of lymph node metastasis in breast cancer
|
| 论文题目英文: |
|
| 作者: |
Dong, D.
|
| 论文出处: |
|
| 刊物名称: |
2016 San Antonio Breast Cancer Symposium
|
| 年: |
2017
|
| 卷: |
77 |
| 期: |
- |
| 页: |
- |
| 联系作者: |
|
| 收录类别: |
|
| 影响因子: |
|
| 摘要: |
Purpose:
To investigate the usefulness of radiomics signature based on computed tomographic (CT) spectral imaging, during the late arterial phase (AP) and portal venous phase (PVP), in preoperative predicting the lymph node (LN) metastasis in breast cancer (BC).
patients and methods:
This retrospective study was institutional review board approved, and written informed consent was obtained from all patients. We examined 60 female patients (LN metastasis positivity was 50%) with CT spectral imaging during the AP and the PVP and data was gathered from 2014 to 2016. Excised lymph nodes were located and labeled during surgery according to location on preoperative CT images and were evaluated histopathologically. For each patient, two 3D Hounsfeld Unit (HU) gradient maps which revealed the HU change of each voxel were built by quadratic fitting the spectral HU curves during the two phases respectively. Then the radiomics features were then extracted from the regions of BC and a suspicious LN judged manually in these maps. The potential association of the four groups of radiomics features with LN status was assessed by using a Mann-Whitney U test. The area under curves (AUC) of the receiver operating characteristic curves (ROC) were compared with data obtained from the conventional CT image.
results:
The 3D HU gradient map showed a great power of distinguishing among different components and was considered as a more effective tool for revealing the intratumour heterogeneity than the conventional CT image since the slope of spectral HU curves were significantly higher in malignant tumor. More than 500 radiomics features extracted from the regions of LN during the AP and the PVP exhibited significant differences (P <0.05). Moreover, the numbers of this kind of features extracted from the regions of BC were more than 200. The highest AUC of single feature was 0.70, which was higher than those from the conventional CT image.
Conclusion:
Quantitative radiomics features based on 3D HU gradient maps have the potential to be exploited as an effective biomarker for preoperative prediction of lymph node metastasis in breast cancer.
|
| 英文摘要: |
|
| 外单位作者单位: |
|
| 备注: |
|
关闭窗口 |
|
