建立前列腺特异性抗原灰区患者重复穿刺阳性的数学模型

Establishment of a nomogram for predicting positive repeat prostate biopsy in patients with total PSA 4 to 10 ng/ml

目的:建立预测前列腺特异性抗原(PSA)灰区患者重复穿刺阳性的数学模型。方法:选择2004~2016年158例血清PSA位于4~10ng/ml且首次穿刺病理结果为阴性的患者行重复穿刺,记录并分析患者的年龄、前列腺体积(PV)、PSA、游离PSA(fPSA)/总PSA(tPSA)、前列腺特异性抗原速率(PSAV)、前列腺特异抗原密度(PSAD)、前列腺移行带特异性抗原密度(PSAD-TZ)、超声检查(TRUS)、直肠指检(DRE)、高级别上皮内瘤变(HGPIN)、不典型小腺泡增生(ASAP)等重复活检结果的潜在预测指标。将有统计学意义的变量行二分类Logistic回归分析和建立数学模型,该模型的预测价值通过ROC曲线下面积(AUC)来评估。结果:158例前列腺重复穿刺活检患者中,前列腺癌的检出率为25.9%(41/158),单变量分析结果中统计学上有意义的指标包括Age、PV、f/tPSA、PSAD、PSAD-TZ、DRE、TRUS、Previous HGPIN、Previous ASAP(P<0.05),对以上所有变量进行二分类Logistic回归分析并建立数学模型,预测指标ASAP、HGPIN、f/tPSA、TRUS、DRE被纳入该模型。该模型AUC为89.8%,预测价值较高。结论:该数学模型可以很好的预测PSA患者重复穿刺阳性的概率,能够帮助临床医师判断哪些PSA灰区患者更适合行超声引导下前列腺重复穿刺活检术。

Objective：To establish a nomogram to predict the outcome of a repeat prostate biopsy in patients with serum total prostate-specific antigen（PSA）4 to 10 ng/ml.Method：From 2004 to 2016,158 patients with total PSA 4 to 10 ng/ml underwent repeat prostate biopsy.Potential predictors as age,prostate volume（PV）,total PSA（tPSA）,free-to-total PSA（f/tPSA）,PSA velocity（PSAV）,PSA density（PSAD）,PSA transitional zone density（PSAD-TZ）,transrectal ultrasound（TRUS）,digital rectal examination（DRE）,high-grade intraepithelial neoplasia（HGPIN）,atypical small acinar proliferation（ASAP）were obtained and subjected to univariate analysis.Then we entered the significant variables stepwise into logistic regression models,and developed a nomogram on the probability of prostate cancer in repeat biopsy,whose predictive value was evaluated by receiver operating characteristic（ROC）analysis.Result：Forty-one（25.9%）patients were diagnosed with prostate cancer.Of all the potential predictors,age,PV,f/tPSA,PSAD,PSAD-TZ,TRUS,DRE,HGPIN,ASAP were significantly statistical difference in univariate analysis.On a series of analysis,ASAP,HGPIN,f/tPSA,TRUS,DRE were included in the predictive nomogram.The nomogram exhibited a high predictive value,with the area under the ROC curve（AUC）of 89.8%,significantly greater than that of the prediction based single predictor.Conclusion：The nomogram is a valuable multi-variable predictive tool,which can be used by urologists to aid decisionmaking for patients with total PSA 4 to 10 ng/ml for whom a repeat biopsy is being considered.