Chun Felix K-H, Müller Imke, Lange Imke, Friedrich Martin G, Erbersdobler Andreas, Karakiewicz Pierre I, Graefen Markus, Pantel Klaus, Huland Hartwig, Schwarzenbach Heidi
Department of Urology, University Hospital Hamburg Epppendorf, Hamburg, Germany.
BJU Int. 2006 Sep;98(3):544-8. doi: 10.1111/j.1464-410X.2006.06352.x.
To investigate whether preoperative plasma levels of free DNA can discriminate between men with localized prostate cancer and benign prostatic hyperplasia (BPH).
In all, 161 referred patients suspicious for prostate cancer either by an elevated prostate-specific antigen (PSA) level and/or abnormal digital rectal examination (DRE) were included in this prospective study. Peripheral plasma was taken before prostate biopsy and genomic DNA was extracted from the plasma using the a commercial kit and a vacuum chamber. After controlling for age, PSA level, the percentage free/total (f/t) PSA and prostate volume, the median prostate cancer plasma DNA concentration served as diagnostic threshold in uni- and multivariate logistic regression models. Multivariate models were subjected to 200 bootstraps for internal validation and to reduce over-fit bias.
Subgroups consisted of 142 men with clinically localized prostate cancer and 19 with BPH. The median plasma concentration of cell-free DNA was 267 ng/mL in men with BPH vs 709 ng/mL in men with prostate cancer. In univariate analyses, plasma DNA concentration was a statistically significant and informative predictor (P = 0.032 and predictive accuracy 0.643). In multivariate analyses, it remained statistically significant after controlling for age, tPSA, f/tPSA and prostate volume, increasing the predictive accuracy by 5.6%.
Our data suggest that plasma DNA level is a highly accurate and informative predictor in uni- and multivariate models for the presence of prostate cancer on needle biopsy. The predictive accuracy was substantially increased by adding plasma DNA level. However, larger-scale studies are needed to further confirm its clinical impact on prostate cancer detection.
探讨术前血浆游离DNA水平能否区分局限性前列腺癌男性患者和良性前列腺增生（BPH）患者。
本前瞻性研究纳入了161例因前列腺特异性抗原（PSA）水平升高和/或直肠指检（DRE）异常而疑似前列腺癌的转诊患者。在前列腺活检前采集外周血，使用商业试剂盒和真空室从血浆中提取基因组DNA。在控制年龄、PSA水平、游离/总（f/t）PSA百分比和前列腺体积后，将前列腺癌血浆DNA浓度中位数作为单因素和多因素逻辑回归模型的诊断阈值。多因素模型进行200次自抽样以进行内部验证并减少过度拟合偏差。
亚组包括142例临床局限性前列腺癌男性患者和19例BPH患者。BPH患者无细胞DNA的血浆浓度中位数为267 ng/mL，而前列腺癌患者为709 ng/mL。在单因素分析中，血浆DNA浓度是具有统计学意义且信息丰富的预测指标（P = 0.032，预测准确率0.643）。在多因素分析中，在控制年龄、总PSA、f/tPSA和前列腺体积后，其仍具有统计学意义，预测准确率提高了5.6%。
我们的数据表明，在单因素和多因素模型中，血浆DNA水平是穿刺活检时前列腺癌存在的高度准确且信息丰富的预测指标。加入血浆DNA水平后预测准确率大幅提高。然而，需要更大规模的研究来进一步证实其对前列腺癌检测的临床影响。
