Establishment and validation of nomogram for predicting prostate biopsy results based on pre-biopsy inflammatory markers / 中华泌尿外科杂志

ABSTRACT

Objective:To explore the predictive value of pre-biopsy serum inflammatory markers on positive prostate biopsy results, establish a nomogram model based on pre-biopsy inflammatory markers combined with other parameters, and evaluate its predictive ability for prostate biopsy results.Methods:The clinical data of 601 patients undergoing transperineal prostate biopsy who were admitted to the Second Hospital of Tianjin Medical University from August 2019 to August 2021 were retrospectively analyzed. The median age was 68(35, 89)years, and the median tPSA was 9.56(4.01, 19.95)ng/ml. The median fPSA was 1.36(0.88, 2.02)ng/ml, the median PSAD was 0.16(0.11, 0.26)ng/ml 2, and the median platelet-to-lymphocyte ratio(PLR)was 129.90(98.95, 169.89). PI-RADS v2.1 score<3 points in 189 cases(31.45%), 3 points in 174 cases(28.95%), 4 points in 190 cases(31.61%), and 5 points in 48 cases(7.99%). A simple randomization method was used to obtain 421 cases(70.00%)in the modeling group and 180 cases(30%)in the validation group.There was no significant difference in the clinical data between the two groups ( P>0.05). Univariate and multivariate logistic regression analysis were performed in the modeling group to screen independent influencing factors for the prediction of positive prostate biopsy results. A nomogram model was established and internal verification was conducted. External validation of the model was performed in the validation group. Receiver operating characteristic(ROC)curve was used to verify model discrimination, Hosmer-Lemeshow goodness-of-fit test was used to verify model calibration, and decision curve analysis (DCA) was used to evaluate the net benefit and clinical utility of the predictive model. Results:The results of univariate analysis showed that the age( OR=1.060, P<0.01), histological inflammation( OR=0.312, P<0.01), the number of biopsy needles( OR=0.949, P=0.009), f/tPSA( OR=0.954, P=0.003), PV( OR=0.973, P<0.01), PSAD( OR=29.260, P<0.01), PI-RADS v2.1 score(3-point OR=3.766, P=0.001; 4-point OR=11.800, P<0.01; 5-point OR=57.033, P<0.01), lymphocyte count( OR=1.535, P=0.013), NLR( OR=0.848, P=0.044), PLR( OR=0.994, P=0.005)and SII( OR=0.999, P=0.009)were statistically different between the prostate patients and non-prostate cancer patients in the modeling group; Multivariate analysis showed that age( OR=1.094, P<0.001), fPSA( OR=0.605, P=0.002), histological inflammation ( OR=0.241, P<0.001), PSAD ( OR=7.57, P=0.013), PLR ( OR=0.994, P=0.005) and PI-RADS v2.1 Score(3-point OR=2.737, P=0.016; 4-point OR=8.621, P<0.001; 5-point OR=47.65, P<0.001) was an independent influencing factor for prostate cancer at initial biopsy; a nomogram model based on age, fPSA, PSAD, PLR and PI-RADS v2.1 scores was established. The AUC of the modeling group was 0.849(95% CI 0.810-0.888), and the sensitivity was 80.9%, and the specificity was 76.1%; the AUC of the validation group was 0.862(95% CI 0.809-0.915), and the sensitivity was 91.9%, and the specificity was 67.8%, suggesting that the diagnostic prediction model had a good discrimination. The calibration curve showed that the prediction model was well calibrated ( χ2=6.137, P=0.632). The decision curve analysis (DCA) of the modeling and validation groups indicated a larger net benefit of the predictive model. Conclusions:The nomogram model established in this study based on age, fPSA, PSAD, PLR and PI-RADS v2.1 score showed good predictive efficacy for prostate biopsy in patients with PSA between 4-20 ng/ml.