Validation of Epstein criteria and development of a nomogram for active surveillance in a contemporary Chinese population.

PURPOSE: To evaluate the Epstein criteria for insignificant prostate cancer (CaP) prediction in a contemporary Chinese population, and to develop a risk model with combined clinical and systematic biopsy and targeted biopsy parameters for active surveillance. METHODS: A total of 249 CaP patients with biopsy Gleason score (GS) of 6 were included. One hundred and one patients were eligible for insignificant CaP on final pathology (GS ≤6 and organ-confined). Diagnostic tests were used to validate the ability of the 2 Epstein criteria. Univariate and multivariate regression analyses were performed to identify predictors of insignificant CaP for the development of predictive models. Receiver operating characteristics analysis was used to select the best model, followed by risk nomogram construction and internal validation. RESULTS: There were 47 patients met EC1 and 61 met EC2, with pathological upgrading rates of 36% and 41%, respectively, and 70% (71/101) and 64% (65/101) were missed, with areas under the curve of 0.591 and 0.594, respectively. Four prediction models were developed using regression analysis, and model 2 incorporating age, prostate-specific antigen density, maximum percentage of core involvement at systematic biopsy, and percentage of positive cores at targeted biopsy showed the best diagnostic value (area under the curve = 0.731, sensitivity 62.4%, specificity 77.0%) and was used to construct the nomogram. Calibration curves and decision curve analysis demonstrated favorable calibration (mean absolute error 0.048) and clinical benefits of the novel nomogram. CONCLUSION: The Epstein criteria need to be revised by incorporating targeted biopsy parameters to improve diagnostic performance, and a novel nomogram was developed with better prediction of insignificant CaP.

Development of a Prediction Model for Positive Surgical Margin in Robot-Assisted Laparoscopic Radical Prostatectomy.

A positive surgical margin (PSM) is reported to have some connection to the occurrence of biochemical recurrence and tumor metastasis in prostate cancer after the operation. There are no clinically usable models and the study is to predict the probability of PSM after robot-assisted laparoscopic radical prostatectomy (RALP) based on preoperative examinations. It is a retrospective cohort from a single center. The Lasso method was applied for variable screening; logistic regression was employed to establish the final model; the strengthened bootstrap method was adopted for model internal verification; the nomogram and web calculator were used to visualize the model. All the statistical analyses were based on the R-4.1.2. The main outcome was a pathologically confirmed PSM. There were 151 PSMs in the 903 patients, for an overall positive rate of 151/903 = 16.7%; 0.727 was the adjusted C statistic, and the Brier value was 0.126. Hence, we have developed and validated a predictive model for PSM after RALP for prostate cancer that can be used in clinical practice. In the meantime, we observed that the International Society of Urological Pathology (ISUP) score, Prostate Imaging Reporting and Data System (PI-RADS) score, and Prostate-Specific Antigen (PSA) were the independent risk factors for PSM.