Kang Minyong, Jeong Chang Wook, Choi Woo Suk, Park Yong Hyun, Cho Sung Yong, Lee Sangchul, Lee Seung Bae, Ku Ja Hyeon, Hong Sung Kyu, Byun Seok-Soo, Jeong Hyeon, Kwak Cheol, Kim Hyeon Hoe, Lee Eunsik, Lee Sang Eun
Department of Urology, Seoul National University Hospital, Seoul, Republic of Korea.
Department of Urology, Seoul National University Bundang Hospital, Seongnam, Republic of Korea.
PLoS One. 2014 Jun 17;9(6):e100053. doi: 10.1371/journal.pone.0100053. eCollection 2014.
Although the incidence of prostate cancer (PCa) is rapidly increasing in Korea, there are few suitable prediction models for disease recurrence after radical prostatectomy (RP). We established pre- and post-operative nomograms estimating biochemical recurrence (BCR)-free probability after RP in Korean men with clinically localized PCa.
Our sampling frame included 3,034 consecutive men with clinically localized PCa who underwent RP at our tertiary centers from June 2004 through July 2011. After inappropriate data exclusion, we evaluated 2,867 patients for the development of nomograms. The Cox proportional hazards regression model was used to develop pre- and post-operative nomograms that predict BCR-free probability. Finally, we resampled from our study cohort 200 times to determine the accuracy of our nomograms on internal validation, which were designated with concordance index (c-index) and further represented by calibration plots.
Over a median of 47 months of follow-up, the estimated BCR-free rate was 87.8% (1 year), 83.8% (2 year), and 72.5% (5 year). In the pre-operative model, Prostate-Specific Antigen (PSA), the proportion of positive biopsy cores, clinical T3a and biopsy Gleason score (GS) were independent predictive factors for BCR, while all relevant predictive factors (PSA, extra-prostatic extension, seminal vesicle invasion, lymph node metastasis, surgical margin, and pathologic GS) were associated with BCR in the post-operative model. The c-index representing predictive accuracy was 0.792 (pre-) and 0.821 (post-operative), showing good fit in the calibration plots.
In summary, we developed pre- and post-operative nomograms predicting BCR-free probability after RP in a large Korean cohort with clinically localized PCa. These nomograms will be provided as the mobile application-based SNUH Prostate Cancer Calculator. Our nomograms can determine patients at high risk of disease recurrence after RP who will benefit from adjuvant therapy.
尽管韩国前列腺癌（PCa）的发病率正在迅速上升，但对于根治性前列腺切除术（RP）后疾病复发，合适的预测模型却很少。我们建立了术前和术后列线图，以估计韩国临床局限性PCa男性患者RP后无生化复发（BCR）的概率。
我们的抽样框架包括2004年6月至2011年7月在我们三级中心连续接受RP的3034例临床局限性PCa男性患者。在排除不适当的数据后，我们评估了2867例患者以制定列线图。采用Cox比例风险回归模型来制定预测无BCR概率的术前和术后列线图。最后，我们从研究队列中进行了200次重采样，以确定我们列线图在内部验证中的准确性，用一致性指数（c指数）表示，并进一步用校准图表示。
在中位47个月的随访中，估计的无BCR率为87.8%（1年）、83.8%（2年）和72.5%（5年）。在术前模型中，前列腺特异性抗原（PSA）、阳性活检核心比例、临床T3a和活检Gleason评分（GS）是BCR的独立预测因素，而在术后模型中所有相关预测因素（PSA、前列腺外扩展、精囊侵犯、淋巴结转移、手术切缘和病理GS）均与BCR相关。代表预测准确性的c指数术前为0.792，术后为0.821，在校准图中显示拟合良好。
总之，我们在一个大型韩国临床局限性PCa队列中开发了术前和术后列线图来预测RP后无BCR的概率。这些列线图将作为基于移动应用程序的SNUH前列腺癌计算器提供。我们的列线图可以确定RP后疾病复发高风险的患者，这些患者将从辅助治疗中获益。
