The Clinical Detection of Breast Cancer by Spectrum Method
Gao Tian-xin,Fan Xiao-fei,Xuan Li-xue,Zhang Bao-ning,Li Xia,Bai Jing
DOI: https://doi.org/10.3964/j.issn.1000-0593(2008)11-2531-05
2008-01-01
Spectroscopy and spectral analysis
Abstract:Breast cancer is one of the most frequently encountered malignant tumors of women. Early detection can save lives successfully. A safe, effective detection method is needed. The detection of breast cancer based on the laser-tissue interactions is an international research focus. The prototype of the detection system in the authors' lab uses a 780 nm low frequency modulated laser to penetrate breast tissue. Two-dimensional scan is processed under the control of computer. A photomultiplier tube (PMT) is used to get the penetrated light and convert it to electrical signal. The signal of light intensity is sampled by the system and used to get the near infrared penetrating image of breast after data processing. In the present paper the signal processing method is discussed and the data processing results in the lab experiments are given. Clinical trials were carried out in the Cancer Institute and Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, using NIR (near infrared light) breast scanner developed by the authors' lab. The investigations were performed after approval by the ethic committee of Cancer Institute and Hospital, Chinese Academy of Medical Sciences. Written informed consent was obtained from each subject. None of the patients' names, initials, or hospital numbers was used in this paper. Fifty patients underwent the examination. Thirty four of them were malignant, and 13 were benign. The other 3 lacked pathology results. Analysis and comparison were executed to evaluate the result. NIR images, mammographs, and the ultrasound images were compared with both the pathology results and each other. The accuracy percentage of NIR image reaches 72.5%, which is between the accuracy percentage of ultrasound (77.50%) and that of mammography (71.88%). In this paper, the characteristics of different breast diseases were found in NIR images, which offers criterion for NIR diagnosis method in detail. The typical NIR images of different diseases, such as papillomatosis with local cancer and cancer, were shown. The clinical trial verified the validity of tumor diagnosis with the special absorption of NIR light by hemoglobin. Both the position and the benign/malignant property of tumor can be detected by NIR method. The improving aspects of the prototype were proposed. A new approach was put forward to the optical method.