Cancer Prognostics and Health Outcomes Unit, Division of Urology, University of Montreal Health Center, Montreal, Quebec, Canada.
Department of Urology and Division of Experimental Oncology, URI, Urological Research Institute, IRCCS San Raffaele Scientific Institute, Milan, Italy.
J Surg Oncol. 2020 Jun;121(7):1154-1161. doi: 10.1002/jso.25877. Epub 2020 Feb 27.
To examine the effect of conditional survival on 5-year cancer-specific survival (CSS) probability after radical nephroureterectomy (RNU) in a contemporary cohort of patients with non-metastatic urothelial carcinoma of the upper urinary tract (UTUC).
Within the Surveillance, Epidemiology and End Results database (2004-2015), 6826 patients were identified. Conditional 5-year CSS estimates were assessed after event-free follow-up duration. Multivariable Cox regression (MCR) models predicted cancer-specific mortality (CSM) according to event-free follow-up length.
Overall, 956 (14.0%) were T low grade(LG)N , 1305 (19.1%) T high grade(HG)N , 1215 (17.8%) T N , 2249 (32.9%) T N and 1101 (16.1%) T N /T N . From baseline, 93.4% to 94.2% in T LGN provided 5-year CSS and, respectively, 86.2% to 95.3% in T HGN , 77.5% to 87.8% in T N , 63.0% to 91.1% in T N , and 38.8% to 88.2% in T N /T N . In MCR models, relative to T LGN , T HGN (Hazard ratio [HR] 1.7), T N (HR 3.0), T N (HR: 5.2), and T N /T N (HR 11.9) were independent predictors of higher CSM. Conditional HRs decreased to levels equivalent to T LGN at 3 years vs 5 years of event-free survival for T HGN and all other groups, respectively.
A direct relationship exists between event-free follow-up and survival probability after RNU. From a clinical perspective, such survival estimates may have particular importance during preoperative counseling.
本研究旨在探讨在当代非转移性上尿路上皮癌（UTUC）患者队列中，条件生存对根治性肾输尿管切除术（RNU）后 5 年癌症特异性生存（CSS）概率的影响。
在 Surveillance, Epidemiology and End Results 数据库（2004-2015 年）中，共确定了 6826 例患者。在无事件随访时间后，评估了 5 年 CSS 的条件估计值。多变量 Cox 回归（MCR）模型根据无事件随访长度预测癌症特异性死亡率（CSM）。
总体而言，956 例（14.0%）为 T 低级别（LG）N，1305 例（19.1%）为 T 高级别（HG）N，1215 例（17.8%）为 T N，2249 例（32.9%）为 T N 和 1101 例（16.1%）为 T N /T N。从基线开始，T LGN 患者 5 年 CSS 率为 93.4%至 94.2%，T HGN 患者 5 年 CSS 率分别为 86.2%至 95.3%，T N 患者 5 年 CSS 率为 77.5%至 87.8%，T N 患者 5 年 CSS 率为 63.0%至 91.1%，T N /T N 患者 5 年 CSS 率为 38.8%至 88.2%。在 MCR 模型中，与 T LGN 相比，T HGN（风险比 [HR] 1.7）、T N（HR 3.0）、T N（HR：5.2）和 T N /T N（HR 11.9）是 CSM 更高的独立预测因素。与 T LGN 相比，T HGN 和所有其他组在 3 年和 5 年无事件生存时，条件 HR 分别降低至与 T LGN 相当的水平。
无事件随访与 RNU 后生存概率之间存在直接关系。从临床角度来看，这些生存估计在术前咨询中可能具有特别重要的意义。
