| BackgroundsEndometriosis is a common gynecological benign disease with a relatively high incidence.About one in ten women of childbearing age will have endometriosis in her lifetime,which brings financial and mental pressure to the patient.Ovarian endometriosis represents one of the most common types of endometriosis.Early detection of endometriosis plays an important role in the timely and effective treatment,but the gold standard of diagnosis depends on laparoscopic surgery,which brings great difficulties to the accurate diagnosis of ordinary patients in outpatient clinics.Therefore,the value of non-invasive diagnosis of endometriosis through imaging examinations or easily obtained biological specimens(such as blood,urine,saliva,etc.)cannot be ignored.At present,the stability of the accuracy of imaging examinations is poor,and biomarkers with high sensitivity,high specificity and high clinical relevance have not yet been discovered.With the continuous development of high-throughput technology,proteomics technology has become an important means to investigate biomarkers.Data independent acquisition(DIA)is a kind of newer method of mass spectrometry data acquisition for targeted quantitative proteomics.ObjectiveIn this study,serum samples from patients with ovarian endometriosis(EM group),other benign ovarian tumors(FEM group)and ovarian cancer(OC group)were collected,and DIA technology was used for proteomic analysis in order to screen protein markers.Experiments were further carried out to select and validate highly specific and sensitive protein markers.Besides,relative molecular pathways were explored to provide ideas for subsequent mechanism research.MethodsIn the stage of protein marker screening,a total of 30 serum samples from the three groups of EM,FEM,and OC patients were collected and detected with the method of the DIA high-throughput protein quantitative detection.Differential proteins were screened and bioinformatics analysis was performed to explore molecular pathways.In the stage of protein marker validation,another 203 patients' serum samples were collected to verify the selected protein markers.The proteins were detected by enzyme-linked immunosorbent assay(Elisa)and Western Blot assays.ROC curves were drawn and the sensitivity and specificity were calculated to find protein markers with strong differential diagnosis ability.ResultsIn the stage of protein marker screening,a total of 739 proteins were identified according to the DIA data comparison.After the preliminary screening,the number of identified proteins in the EM group was 722,in the FEM group was 728,and in the OC group was 731.There were 20 differential proteins between the EM group and the FEM group,and 88 differential proteins between the EM group and the OC group.Taking the EM group as a control,five overlapping differential proteins were found,namely transmembrane protein 198(TMEM198),L-selectin(SELL),leukocyte immunoglobulin-like receptor subfamily A member 3(LILRA3),immunoglobulin kappa constant(IGKC)and desmocollin-2(DSC2).The differential proteins between the EM group and the FEM group were mainly concentrated in the metabolism,the PPAR and the AMPK pathway,while the differential proteins between the EM group and the OC group were concentrated in the complement and coagulation cascades,and the MAPK pathway.In the stage of protein marker validation,five markers CA125,DSC2,LILRA3,SELL and IGKC were selected for experiments.We take the AUC,sensitivity and specificity in the consideration for the analysis.(1)For EM and FEM,the 5-protein-marker panel had a stronger diagnostic capability,with the AUC of 0.823,the sensitivity of 76.14%,and the specificity of 75.56%.For EM and OC,IGKC was of higher diagnostic value,while CA125+IGKC held the strongest diagnostic capability,with the AUC of 0.969,the sensitivity of 92.05%,and the specificity of 92.00%.(2)For EM and FEM patients aged<45 years,the AUC of CA125 was relatively larger(0.733),with the sensitivity of 70.51%,and the specificity of 70.97%;for patients aged ?45 years,the diagnostic capability of IGKC was stronger(AUC:0.900,sensitivity:100%,specificity:82.14%).For EM and OC patients aged ?45 years,CA125+IGKC had a higher diagnostic capability(AUC:0.980,sensitivity:100%,specificity:90.00%).(3)For EM and FEM patients with a tumor diameter of<5cm,CA125 was of higher diagnostic value,with the AUC of 0.854,the sensitivity of 71.43%,and the specificity of 85.00%;while for patients with a tumor diameter of?5cm,CA125 also had a stronger diagnostic capability(AUC:0.751,sensitivity:73.33%,specificity:72.86%).(4)For the differential diagnosis of EM and FEM with an age of<45 years and a tumor diameter of<5cm,the AUC of CA125+IGKC was 0.834,which was of high diagnostic value.ConclusionIn this study,DIA quantitative proteomic analysis technology was used to screen protein markers.Through experimental verification of a larger sample size,we found that the commonly used clinical marker CA125 had a strong ability to distinguish between EM and FEM,and the new marker IGKC had a strong ability to distinguish between EM and OC.The marker panel with several proteins especially CA125 and IGKC indicated promising diagnostic value.In addition,we found that metabolism and inflammation related factors may play an important role in the development of endometriosis,providing ideas for future research on related mechanisms.BackgroundsOvarian clear cell carcinoma is a subtype of epithelial ovarian cancer with unique biological characteristics and a high degree of malignancy.It poses a threat to the life and health of patients,and also brings a heavy psychological and economic burden to them.In view of the low incidence,the underlying mechanism of the occurrence and development of ovarian clear cell carcinoma is not clear.The relevant treatment plan follows the treatment standard of epithelial ovarian cancer,and there is a lack of effective and targeted treatment regimen.The clinical prediction model established by integrating multiple prognostic factors can help clinicians and patients design individualized treatment and follow-up plans together.However,the current researches on the factors affecting the prognosis of clear cell ovarian cancer were not quite systematic,and there is no accurate prediction model for predicting the prognosis of ovarian clear cell carcinoma.Objective:This study aims to combine the clinical data of public databases and data from our hospital to screen the prognostic factors of ovarian clear cell carcinoma and further establish prognostic prediction models.The clinical application of the prediction model is scientifically validated with various methods using external data to better facilitate clinical management.Methods(1)Prediction model for the overall survival:The clinical data of patients with ovarian clear cell carcinoma from 2010 to 2016 in the SEER database were obtained and randomly divided into the training set and the internal validation set in the ratio of 7:3.The records of patients with ovarian clear cell carcinoma from 2000 to 2011 in our hospital were also collected as an external validation set.The final variables were screened by Lasso Cox regression analysis method and multivariate Cox regression analysis to construct the model.(2)Prediction model for the progression-free survival:The data of patients with ovarian clear cell carcinoma from 2012 to 2017 were obtained as the training set,and the data from 2000 to 2011 were collected as the validation set.The final variables were screened by univariate Cox regression analysis and multivariate Cox regression analysis to construct the model.For the above two prediction models,the methods of the consistency index(C-index),the calibration curve,the Net Reclassification Index(NRI)and the Integrated Discrimination Improvement(IDI)and the decision curve were carried out for the validation.The risk score of each variable in the model was extracted to construct risk stratification system for patients,and the survival curve was drawn to compare the prognosis.Results(1)Prediction model for the overall survival:A total of 1079 patients in the training set,462 patients in the internal validation set,and 118 patients in the external validation set were enrolled.The initial variables included were:patient's age at diagnosis,race,tumor size,tumor location(unilateral or bilateral),tumor grade,tumor stage,lymph node status,radiotherapy,and chemotherapy.After screening,the age of diagnosis,tumor location,tumor stage,lymph node status and chemotherapy were integrated to construct the prediction model.The C-index is 0.801(0.786,0.816)in the training set,0.789(0.627,0.816)in the internal validation set,and 0.923(0.899,0.946)in the external validation set.(2)Prediction model for the progression-free survival:A total of 122 patients in the training set and 118 patients in the validation set were collected.The initial variables included were age at diagnosis,menopausal status,parity,preoperative CA125 level,tumor size,tumor location,tumor stage,lymph node status,postoperative residual disease,concurrent endometriosis and hypertension.After screening,preoperative CA125 level,tumor size,tumor stage and postoperative residual disease were integrated to construct the prediction model.The C-index is 0.703(0.617,0.790)in the training set,and 0.818(0.758,0.878)in the validation set.The calibration curves of both models indicated that little difference was noticed between the estimated prognosis and the actual prognosis.The results of NRI and IDI show that the performance of the model is prior to that of a single variable(tumor staging).The decision curve analysis revealed that obvious clinical benefits were observed from the two models.The two models exhibited strong distinguishing ability for high-and low-risk patients(both P values<0.05).ConclusionThis study used the SEER database and data of our hospital to establish prognostic models of the overall survival and the progression-free survival for patients with ovarian clear cell carcinoma.The validation results showed excellent performance in both models which indicated high clinical application value in patient management. |
PDF Full Text Request