Computerized evaluation scheme to detect metastasis in sentinel lymph nodes using contrast-enhanced computed tomography before breast cancer surgery
- 45 Downloads
Sentinel lymph node (SLN) biopsy for evaluating lymph node metastasis during breast cancer surgery is associated with several problems, such as the consequent increase in operation time and the possibility of abrupt changes in the treatment plan during the operation. Although it is desirable to distinguish SLNs with and without cancer metastasis before surgery, there is no established examination for this purpose. This study aimed to develop a computerized scheme for evaluating metastasis in SLNs by analyzing computed tomography lymphography images and the three-dimensional versions of these images. Our database consisted of computed tomography lymphography images from 100 patients with breast cancer. Three subjective features of the nodes were assessed in the three-dimensional images: (1) the shape of the lymphoduct, (2) degree of signal enhancement in the nodes, and (3) shape of the nodes. Six objective features were also assessed in the computed tomography lymphography images: (4) the long axis, (5) area, (6) standard deviation of the signal values, (7) mean signal values, (8) maximum signal value, and (9) minimum signal value. Support vector machines were employed to evaluate cancer metastasis in SLNs. For the input, six of the nine features were selected in a stepwise method. The classification accuracy, sensitivity, and specificity were 98.0% (98/100), 97.8% (44/45), and 98.2% (54/55), respectively. The positive and negative predictive values were 97.8% (44/45) and 98.2% (54/55), respectively. This computerized method exhibited high classification accuracy and will be useful in determining the need for lymph node dissection before breast cancer surgery.
KeywordsSentinel lymph node CT lymphography Breast cancer Metastasis
The authors are grateful to the members of the Department of Radiology, Maruyama Memorial General Hospital, for supporting this work.
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
For this type of study, formal consent is not required at our Institution. This article does not contain any studies with animals performed.
Informed consent was obtained from all individual participants included in the study.
- 2.National Cancer Center. Cancer Registry and Statistics Japan. Cancer Inf Serv. 2014;105:1480–6.Google Scholar
- 3.Kurebayashi J, Miyoshi Y, Ishikawa T, Saji S, Sugie T, Suzuki T, Takahashi S, Nozaki M, Yamashita H, Tokuda Y, Nakamura S. Clinicopathological characteristics of breast cancer and trends in the management of breast cancer patients in Japan: based on the Breast Cancer Registry of the Japanese Breast Cancer Society between 2004 and 2011. Breast Cancer. 2015;22:235–44.CrossRefPubMedGoogle Scholar
- 7.Yamamoto S, Maeda N, Tamesa M, Nagashima Y, Yoshimura K, Oka M. Prospective ultrasonographic prediction of sentinel lymph node metastasis by real-time virtual sonography constructed with three-dimensional computed tomography-lymphography in breast cancer patients. Breast Cancer. 2012;19:77–82.CrossRefPubMedGoogle Scholar
- 12.Minato M, Hirose C, Sasa M, Nishitani H, Hirose Y, Morimoto T. 3-dimensional computed tomography lymphography-guided identification of sentinel lymph nodes in breast cancer patients using subcutaneous injection of nonionic contrast medium: a clinical trial. J Comput Assist Tomogr. 2004;28:46–51.CrossRefPubMedGoogle Scholar
- 14.Borgstein PJ, Meijer S, Pijpers R. Intradermal blue dye to identify sentinel lymph-node in breast cancer. Lancet. 1997;384:149–57.Google Scholar
- 15.Vapnik V, Lerner A. Pattern recognition using generalized portrait method. Automat Rem Contr. 1963;24:774–80.Google Scholar
- 16.Tsochantaridis I, Joachims T, Hofmann T, Altun Y. Large margin methods for structured and interdependent output variables. J Mach Learn Res. 2005;6:1453–84.Google Scholar
- 17.Cortes C, Vapnik V. Support-vector networks. Mach Learn. 1995;20:273–97.Google Scholar
- 18.Bergstra J, Bengio Y. Random search for hyper-parameter optimization. J Mach Learn Res. 2012;13:281–305.Google Scholar
- 22.Liu C, Ding J, Spuhler K, Gao Y, Serrano Sosa M, Moriarty M, Hussain S, He X, Liang C, Huang C. Preoperative prediction of sentinel lymph node metastasis in breast cancer by radiomic signatures from dynamic contrast-enhanced MRI. Magn Reson Imaging. 2018 (Epub ahead of print).Google Scholar
- 23.Dong Y, Feng Q, Yang W, Lu Z, Deng C, Zhang L, Lian Z, Liu J, Luo X, Pei S, Mo X, Huang W, Liang C, Zhang B, Zhang S. Preoperative prediction of sentinel lymph node metastasis in breast cancer based on radiomics of T2-weighted fat-suppression and diffusion-weighted MRI. Eur Radiol. 2018;28:582–91.CrossRefPubMedGoogle Scholar