Development Of Predictiong Model For Positive Surgical Margins In Robot-assisted Laparoscopic Radical Prostatectomy

Background: According to the 2020 global cancer statistics released by the International Agency for Research on Cancer,more than 1.4 million new cases of prostate cancer in the world in 2020,accounting for 7.3% of new cases of cancer,ranking the third.In the country,the incidence of prostate cancer has been on the rise since 2015.Radical prostatectomy(either laparoscopically or robotically assisted)is still the first-line treatment for prostate cancer.However,the effect of surgery is different for different patients.There are still 27-53% of patients with biochemical recurrence or even tumor metastasis after surgery,which is lifethreatening in serious cases.Multiple studies have shown that postoperative biochemical recurrence and tumor metastasis of prostate cancer are closely related to positive surgical margins.Therefore,it is of guiding significance to determine the operative margin of patients before operation.According to the information of preoperative prostate puncture,imaging,test and other examinations,the probability of positive surgical margin(PSM)after Robot-assisted laparoscopic radical prostatectomy(RARP),on the basis of which the corresponding operation and treatment plan is given,is predicted.And a scoring scheme which can be used clinically will be given.It can help clinicians to quantify the significance of various preoperative indicators and help junior doctors to adapt to clinical work more quickly.Methods: This is a retrospective cohort study.The data were provided in the database of Affiliated Drum Tower Hospital,Medical School of Nanjing University.Patients diagnosed with prostate cancer from January 2018 to December 2021 without preoperative neoadjuvant therapy and no history of prostate-related surgery,and whose pathology was reported as prostate cancer after RARP in Nanjing Drum Tower Hospital were selected as the research subjects.A total of 903 cases were collected in the study.After removing the missing data(21data,2.3%),882 complete data were used for logistic regression model establishment in total.Lasso method was used for variable screening,logistic regression to establish the final model,and strengthened bootstrap method for model internal verification.Finally,the model was visualized using nomograms.All the statistical analyses were based on the R-4.1.2.Based on the Lasso method,nine predictors were ultimately included in the model,which were age,percent of positive needles(PPN),international society of urological pathology(ISUP),percent of tumour(PT),maximal tumor diameter(MTD),prostate imaging reporting and data system(PI.RADS),tumour location(TL),clinical staging of the tumor provided by MRI(T-MRI),and PSA.The main outcome was a positive surgical margin(PSM)(extension of cancer cells in the ink section of the specimen)of the patient after RARP reported pathologically.Results: 1.Among the 903 patients,151 had PSM,and the overall positive rate was 151÷903*100%=16.7%.2.Univariate analysis showed that age,PPN,ISUP score,PT,MTD,PIRADS score,PSA,PSAD and TL were significantly different between the PSM group and the Negative surgical margin(NSM)group(P< 0.05).3.Multivariate analysis showed that ISUP score,PI-RADS score and PSA were independent risk factors for PSM(P< 0.05).4.The adjusted C statistic of the prediction model is 0.727(this is the model differentiation index,higher value means better differentiation);Brier value is 0.126(this is the model calibration index,smaller value indicates better calibration degree,generally less than 0.25).The calibration curve has a high fitting degree to the ideal curve in low-risk and medium-risk patients,but a general fitting degree in high-risk groups,which will overestimate the results to some extent.Conclusions: 1.The prediction model of PSM after RARP operation was established and verified.It can help clinicians to identify high-risk groups with positive surgical margins,so as to give a more personalized surgical and treatment plan.2.The indexes commonly used to evaluate PCa in clinic were integrated,and some new variables were included in the model,such as PPN,PT and MTD.3.An independent predictor of PSM after RARP was obtained:ISUP score.PI-RADS score,PSA.