Karakiewicz Pierre I, Shariat Shahrokh F, Palapattu Ganesh S, Perrotte Paul, Lotan Yair, Rogers Craig G, Amiel Gilad E, Vazina Amnon, Gupta Amit, Bastian Patrick J, Sagalowsky Arthur I, Schoenberg Mark, Lerner Seth P
Cancer Prognostics and Health Outcomes Unit, University of Montreal, Montreal, Quebec, Canada.
Eur Urol. 2006 Dec;50(6):1254-60; discussion 1261-2. doi: 10.1016/j.eururo.2006.06.010. Epub 2006 Jun 23.
To evaluate precystectomy prediction of pT and pN stages at cystectomy.
Multivariate logistic regression analyses modelled variables of 726 evaluable patients treated with radical cystectomy and bilateral pelvic lymphadenectomy. The first set of models predicted pT(3-4) stage at cystectomy, and the second set predicted pN(1-3) stages at cystectomy. Transurethral resection (TUR) predictors consisted of 2002 T stage, 1973 WHO tumour grade, presence of carcinoma in situ, age, gender, and delivery of neo-adjuvant chemotherapy. The area under the ROC curve quantified nomogram accuracy. Two hundred bootstrap resamples were used to reduce overfit bias.
At TUR, 11% of patients were staged as pT(3-4) versus 42% at cystectomy. Lymph node metastases were found in 24% of patients at cystectomy (pN(1-3)). The multivariate pT(3-4) nomogram was 75.7% accurate versus 71.4% for TUR T stage. The multivariate pN(1-3) nomogram was 63.1% accurate versus 61.0% for TUR T stage.
Multivariate nomograms are not perfect, but they do predict more accurately than TUR T stage alone.
评估膀胱切除术前对膀胱切除时pT和pN分期的预测情况。
多变量逻辑回归分析对726例接受根治性膀胱切除术和双侧盆腔淋巴结清扫术的可评估患者的变量进行建模。第一组模型预测膀胱切除时的pT（3 - 4）期，第二组模型预测膀胱切除时的pN（1 - 3）期。经尿道切除术（TUR）的预测因素包括2002年T分期、1973年世界卫生组织肿瘤分级、原位癌的存在、年龄、性别以及新辅助化疗的实施情况。ROC曲线下面积量化了列线图的准确性。使用200次自抽样重采样来减少过度拟合偏差。
在TUR时，11%的患者被分期为pT（3 - 4）期，而在膀胱切除时这一比例为42%。在膀胱切除时，24%的患者发现有淋巴结转移（pN（1 - 3））。多变量pT（3 - 4）列线图的准确率为75.7%，而TUR T分期的准确率为71.4%。多变量pN（1 - 3）列线图的准确率为63.1%，而TUR T分期的准确率为61.0%。
多变量列线图并不完美，但它们比单独的TUR T分期预测更准确。
Zotero 插件