Font Size: a A A

Prediction Model Of Biochemical Recurrence Within 1 Year After Radical Prostatectomy Based On Bayesian Network

Posted on:2023-06-15Degree:MasterType:Thesis
Country:ChinaCandidate:K X WangFull Text:PDF
GTID:2544306833454414Subject:Surgery
Abstract/Summary:PDF Full Text Request
Background: Prostate Cancer is the second leading cause of cancer in men worldwide after lung cancer.Radical prostatectomy is one of the most important and effective treatments for prostate cancer,yet about a third of prostate cancer patients experience postoperative biochemical recurrence(BCR).Although biochemical recurrence does not necessarily indicate tumor recurrence or metastasis,it often indicates clinical recurrence.Establish a predictive model of biochemical recurrence after radical prostatectomy is helpful to guide the selection of treatment regimens.More and more computer algorithms are gradually applied to the clinic,so as to better mine the relationship between data and provide a new way to solve clinical problems.Bayesian network(BN)is a kind of network structure based on the directed acyclic graph,which is good at dealing with the complex dependencies among many characteristic variables.It can make full use of the information of all variables to deal with the missing data more easily.Therefore,we chose the method based on the Bayesian network to establish models to predict the biochemical recurrence within 1 year after radical prostatectomy.Objective: To investigate the clinical value of a biochemical recurrence model based on a Bayesian network within 1 year after radical prostatectomy.Methods: The clinical data of 209 patients who underwent laparoscopic radical prostatectomy in Qingdao University Affiliated Hospital from January 2018 to March 2021 were retrospectively analyzed.According to postoperative examination,patients were divided into a biochemical recurrence group and the non-biochemical recurrence group.The general clinical data of the two groups,preoperative examination results,puncture and postoperative pathological information,treatment methods,MRI,and other related factors were recorded and univariate analysis was performed.A naive Bayesian prediction model for biochemical recurrence within 1 year after radical prostatectomy was constructed using univariate related risk factors,and confusion matrix,receiver operating characteristic curve,and area under the curve(AUC)were used to evaluate the prediction effect of the model.Based on the Bayesian network model and the Bayesia Lab validation function,a priori probability statistical analysis was performed on the predictors of biochemical recurrence within 1 year after radical prostatectomy.According to the results of the posterior analysis,the importance of polymorphic Birnbaum is analyzed and calculated,and the importance order of each attribute variable is given.Results: Biochemical recurrence occurred in 43 patients in 209 cases within 1 year after operation,and the rate of biochemical recurrence was 20.57%.Univariate analysis showed that PI-RADs score(P = 0.009),Gleason score for radical pathology(P = 0.026),encapsulation invasion(P = 0.023),seminal vesicle invasion(P = 0.001),lymphatic vessel and peripheral nerve invasion(P = 0.045),incision nature(P < 0.001),clinical stage(P =0.009),pathological stage(P = 0.023),operative mode(P = 0.047),postoperative adjuvant therapy(P = 0.014)were risk factors for biochemical recurrence within one year after radical prostatectomy.Based on the above 10 factors,the Naive Bayes model is established.The final model AUC is 81.43%,the accuracy is 80.95%,the sensitivity is 46.15%,the specificity is 96.55%,the positive predictive value is 85.71%,and the negative predictive value is 80.00%.The results showed that the nature of incision margin,clinical stage,operation method,and invasion of seminal vesicle were the top 4 prognostic factors.Conclusion: The prediction model of biochemical recurrence within 1 year after radical prostatectomy based on the Bayesian network has moderate or higher predictive value and high accuracy.It can be used as a method for predicting biochemical recurrence within 1 year after radical prostatectomy.The rank of the importance degree of prognostic factors affecting biochemical recurrence within 1 year after radical prostatic cancer was given,PSM was the most important factor affecting biochemical recurrence within 1 year after radical prostatic cancer.Clinical stage,surgical approach and seminal vesicle invasion are the more important influencing factors for biochemical recurrence within 1 year after radical prostatectomy.When the relevant factors appear,intervention should be actively carried out to guard against the occurrence of biochemical recurrence.
Keywords/Search Tags:Prostate cancer, biochemical recurrence, predictive models, Bayesian networks
PDF Full Text Request
Related items