Analysis Of Clinicopathological Characteristics Of Cervical Cancer And Construction Of Prognostic Prediction Model Based On Multi-omics Data

As a common female malignancy,cervical cancer(CC)remains one of the leading causes of cancer-related deaths worldwide.The pathological types of CC mainly include adenocarcinoma(AC)and squamous cell carcinoma(SCC).With the promotion of human papillomavirus(HPV)vaccines and cancer screening,the incidence of CC have decreased in decades,but the incidence of cervical adenocarcinoma(AC)is increasing yearly,especially in young women.Moreover,women with advanced or recurrent disease face a dismal prognosis with potentially considerable mortality.The treatment and prognostic prediction of CC patients still remain a big challenge.This study takes cervical cancer as the research object,which is divided into two parts.The first part aims to analyze the clinicopathological characteristics of young CC patients,especially AC patients.The information of young patients(35 years old or younger when diagnosed)diagnosed as SCC or AC of cervix in Peking Union Medical College Hospital from 2012 to 2017 was retrospectively collected.It was verified that the symptoms of CC are nonspecific.Bleeding after sexual intercourse is the most common complaint in young patients,especially in SCC group.AC diagnosis can be difficult since AC group has a lower abnormal rate in TCT and HPV test than SCC group.AC group has worse prognosis than SCC group.Histological type is an important factor affecting the prognosis of young cervical cancer.For young patients with early stage cervical cancer,fertility preservation treatment can be considered.There is no significant difference in the prognosis between fertility-sparing group and nonfertility-sparing group.The second part is a bioinformatics analysis,based on public data downloaded from TCGA database.The information of training dataset and test dataset were downloaded from TCGA database.The copy number variation(CNV)data and mutation data of training dataset were analyzed respectively,and genes related to prognosis of CC were identified out by transcriptome data.After the integration of transcriptome and genomic results,6 genes of prognosis value of CC were identified by random survival forest,which were SLC19A3,FURIN,SLC22A3,DPA GT1,CCL17,DES.Then a 6-gene signature was constructed through multivariate Cox regression analysis.Subsequently,it is proved that the signature has robustness and strong stability through the verification of test dataset and independent validation dataset.In the training dataset,test dataset and independent validation dataset,the risk score group were analyzed by multivariate Cox regression with other prognostic factors.It was proved that the 6-gene signature is clinically independent.Finally,the relationship between the expression of 6 genes and prognosis was confirmed by immunohistochemistry.