Exploration Of Humanized Diagnosis And Treatment Of Cervical High-grade Intraepithelial Lesions

OBJECTIVE:To introduce the artificial intelligence medical management system to monitor the patient information from admission to diagnosis and treatment to follow-up in a comprehensive and detailed way,so as to efficiently obtain patient diagnosis and treatment information and provide comprehensive and effective medical information resources for cervical cancer screening in China.Second,through cytology,histological double-stain technology to carry out disease shunt,the occurrence and development of the disease can be judged more accurately,improve the economic benefits of diagnosis and treatment for the country and patients,and avoid excessive medical treatment and resource waste.METHODS:Artificial intelligence medical management system was established,the whole medical process was input into the big data system,and the functions of patient medical record data collection,follow-up,analysis,remote consultation and so on were completed.Clinical cervical exfoliated cells were collected,and TCT and hrHPV tests were performed at the same time.Combined with the results of cervical histopathology,the remaining cervical epithelial exfoliated cells were performed P16/Ki-67 immunocytochemical double staining test,and the diagnostic efficacy of cytological P16/Ki-67 double staining test and hrHPV for CIN2 grade and above lesions was compared.Histological specimens of cervix were collected and p16/Ki-67 histological double staining was performed to compare the consistency of histological p16/Ki-67 double staining results with pathological results.Cervical cancer samples were collected for micro cutting,lesions of different degrees of the same tissue were collected,and DNA was extracted for exon sequencing.Copy number variation,missense mutation,driver gene,high frequency mutation,susceptibility gene and cloning analysis were carried out on the sequencing results to further clarify the relationship between components of different degrees.RESULTS:The positive rate of cytological p16/Ki-67 double staining was approximately proportional to the lesion degree,and the coincidence rate was higher than that of hrHPV detection(P=0.6963).The specificity(57%vs.15.8%,P=0.0014),overall coincidence rate(67.7%vs.38%,P=0.0058),positive predictive value(46.2%vs.30.4%,P=0.2829)of CIN2+were detected by double staining.The negative predictive value(97%vs.90%,P=0.4958)and Kappa coefficient(0.395 vs.0.068)were superior to the hrHPV test and could reduce coloscopy shunt by 29.7%.The positive rate of histological p16/Ki-67 double staining was approximately proportional to the degree of lesion.According to the degree of DNA degradation,22 patients were selected for exon sequencing.For the results of whole exon sequencing,somatic mutations common to CIN components of different lesions were defined as common mutations,while mutations unique to each were defined as unique mutations.We found a large number of unique somatic cell single nucleotides and unique copy number variations in the samples.However,a large number of common missense SNV and somatic copy number variations were found in patients' LSIL,HSIL and cervical cancer cells,and a large number of common missense mutations(including SNV,insertion deletion)were also found among the three components.CONCLUSION:Artificial intelligence medical whole process management system can facilitate clinical research data management,remote consultation,and improve the level of primary medical diagnosis.Cytological p16/Ki-67 double staining is a good shunt method for cervical cancer screening,which can improve the detection efficiency of CIN2+,reduce misdiagnosis rate and overtreatment.Histological p16/Ki-67 double staining is a good set of biological markers,which can assist in the diagnosis of cervical lesions.Cervical lesions in different patients have obvious tumor heterogeneity,and the same patient has similar mutation spectrum in different degrees of cervical lesions,but different degrees of cervical lesions in a single cervical are not always gradually derived from the same source of LSIL.