Abstract
Cervical cancer is a serious threat to women's health due to malignant tumours, and early detection can greatly reduce mortality. In this paper, cervical tissue was used as the research object, and Raman spectroscopy analysis of cervical inflammation and precancerous tissues was used to detect cervical cancer. This provides a clinical basis for the use of Raman spectroscopy in analysis of cervical precancerous lesions. In this study, the actual Raman spectrum signal of precancerous cervical tissue was collected, and the PLS and Relief methods were used to extract the signal characteristics of the spectrum. Then, we established and compared KNN and ELM classification models and finally achieved the early diagnosis of cervical cancer. This experiment designed a novel feature fusion method in feature extraction, and we used the first and second derivative features that reflect more peak details of the original spectrum for fusion. The accuracy rate of KNN without feature fusion is 88.17%, and the accuracy rate after fusion is 93.55%. The accuracy rate of ELM without feature fusion is 90.81%, and the accuracy rate after fusion is 93.51%. The results show that the accuracy of feature fusion has been improved to a certain extent, and this method is expected to be used as a new method of spectral data fusion.
Highlights
Every year, more than 500000 women worldwide are diagnosed with cervical cancer, and the number of female cervical cancer deaths worldwide is as high as 300000 [1], [2]
All samples were randomly divided into a training set and test set at a ratio of 7:3: the inflammation training set contained 34 samples, and the test set contained 15 samples; the Low-grade squamous intraepithelial lesions (LSIL) training set contained 20 samples, and the test set contained 9 samples; and the high-grade squamous intraepithelial lesions (HSIL) training set contained 31 samples, and the test set contained 14 samples
For the data obtained after the fusion of the various derivative features, Relief is used for feature selection, and 50 features are still selected for the construction of the classification model
Summary
More than 500000 women worldwide are diagnosed with cervical cancer, and the number of female cervical cancer deaths worldwide is as high as 300000 [1], [2]. In the early stages of cervical cancer, if effective treatment can be obtained, the patient’s 5-year survival rate is as high as 90%. The Director-General of the World Health Organization (WHO) issued a global call for action in 2018 This is for the early screening of cervical cancer to eliminate cervical cancer [4], [5]. Low-grade squamous intraepithelial lesions (LSIL) (CIN1) are mainly caused by low-risk Human Papilloma Virus (HPV) infection, and high-grade squamous intraepithelial lesions (HSIL) (CIN2 + CIN3) remainly caused by persistent high-risk HPV infection. It takes approximately 5–8 years after HPV infection to cause cervical cancer from CIN. There is an urgent need to develop a stable, efficient and fast diagnostic method to provide help for the early diagnosis and pathological research of cervical cancer
Sign-in/Register to view highlights & summary Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
