MRI和1997年版Partin表对前列腺癌病理特征预测准确性的对比研究

Accuracy of MRI and 1997 edition of Partin tables in predicting the pathological features of prostate cancer

    摘要
  • 摘要:
    目的 比较MRI与1997年版Partin表对前列腺癌病理特征预测结果的准确性。
    方法 随机选取2012年1月至2014年2月51例前列腺癌患者并行根治性前列腺切除术，统计资料具备术前血清前列腺特异性抗原结果、临床分期、Gleason评分、盆腔MRI资料以及术后病理报告。分别将术前MRI及Partin表对前列腺癌器官局限性癌、包膜侵犯、精囊侵犯以及淋巴结转移4种病理特征的预测结果与术后病理结果进行比较，通过受试者工作特征曲线（ROC）分析法检验MRI与Partin表预测前列腺癌病理特征的准确性并进行比较。
    结果 应用Partin表预测器官局限性癌、包膜侵犯、精囊侵犯以及淋巴结转移4种病理特征的曲线下面积分别为0.911、0.742、0.827、0.899；应用MRI预测的曲线下面积分别为0.902、0.765、0.563、0.500。MRI与Partin表预测器官局限性癌和包膜侵犯的ROC曲线下面积差异无统计学意义（Z=0.071、0.020，P均>0.05）；预测精囊侵犯和淋巴结转移的ROC曲线下面积差异有统计学意义（Z=0.286、0.499，P均 < 0.01）。
    结论 Partin表预测前列腺癌4种病理特征具有临床诊断价值；MRI与Partin表预测前列腺癌器官局限性癌、包膜侵犯2种病理特征的准确性差异无统计学意义，Partin表预测精囊侵犯、淋巴结转移2种病理特征的准确性高于MRI。

     

    Abstract:
    Objective To compare the accuracies of MRI and 1997 edition of Partin tables in predicting the pathological features of prostate cancer.
    Methods A total of 51 patients with prostate carcinoma underwent radical prostatectomy from January 2012 to February 2014. Preoperative serum prostate specific antigen, clinical stage, biopsy Gleason score, percentage of positive biopsy scores, pelvic MRI, and pathological report of prostatectomy specimen were collected. Postoperative pathological results were compared with MRI results and Partin tables. Receiver operating characteristic curves were plotted to determine the accuracies of MRI and Partin tables in predicting the pathological features of prostate cancer.
    Results The areas under the curve(AUCs)of the Partin tables to predict organ-confined disease(OCD), established capsular penetration(ECP), seminal vesicle involvement(SVI), and lymph nodal involvement(LNI) were 0.911, 0.742, 0.827, and 0.899, respectively. The corresponding AUCs of MRI were 0.902, 0.765, 0.563, and 0.5. The AUCs of OCD and ECP did not significantly differ between MRI and Partin tables(Z=0.071 and 0.020, both P > 0.05). By contrast, the AUCs of SVI and LNI significantly differed between MRI and Partin tables(Z=0.286 and 0.499, both P < 0.01).
    Conclusion Partin tables exhibit a high clinical diagnostic value in the prediction of the pathological feature of prostate cancer. The accuracies of predicting OCD and ECP did not significantly differ between MRI and Partin tables. However, Partin tables were more accurate in predicting SVI and LNI than MRI.

     

    • HTML全文
    • 参考文献(20)
    • 相关文章
    • 施引文献
  • 资源附件(0)

/

返回文章
返回