Department of Surgery, Stanford University School of Medicine, Stanford, CA.
Department of Surgery, Memorial Sloan Kettering Cancer Center, New York, NY; Department of General, Visceral and Vascular Surgery, Charite Campus Benjamin Franklin, Berlin, Germany.
Surgery. 2022 Jun;171(6):1580-1587. doi: 10.1016/j.surg.2022.01.030. Epub 2022 Feb 25.
Prognostic stratification of patients with colorectal cancer liver metastasis based solely on tumor-related factors has only moderate discriminatory ability. We hypothesized that the inclusion of nontumor related factors can improve prediction of long-term prognosis of patients with colorectal cancer liver metastasis.
Nontumor related laboratory markers were assessed utilizing a training cohort from 2 U.S. institutions (n = 1,205). Factors independently associated with prognosis were used to develop a nontumor related prognostic score. The discriminatory ability, assessed by Harrell's C-statistics (C-index) and net reclassification improvement, was validated and compared with 3 commonly used tumor-related clinical risk scores: Fong clinical risk scores, m-clinical risk scores, and Genetic and Morphological Evaluation (GAME) score in an external validation cohort from 5 Asian (n = 1,307) and 3 European (n = 1,058) institutions. The discriminatory ability of nontumor related prognostic score combined with each of these 3 tumor-related prognostic scores was also estimated.
Alkaline phosphatase (hazard ratio 1.43; 95% confidence interval, 1.11-1.84), albumin (hazard ratio 0.71; 95% confidence interval, 0.57-0.89), and mean corpuscular volume (hazard ratio 19.0, per log unit; 95% confidence interval, 4.79-75.0) were each independently associated with increased risk of death after resection of colorectal cancer liver metastasis (all P < .05). In turn, alkaline phosphatase, albumin, and mean corpuscular volume were combined to form a nontumor related prognostic score (2.942 × mean corpuscular volume + 0.399 × alkaline phosphatase-0.339 × albumin-12) × 10 (median, 16; range, 1-30). The nontumor related prognostic score had good-to-modest discriminatory ability in the external cohort (C-index = 0.58), which was comparable to the 3 established tumor-related prognostic scores (C-index: Fong clinical risk scores, 0.53, m-clinical risk scores, 0.55, GAME, 0.58). The addition of the nontumor related prognostic score to the tumor-related prognostic scores enhanced the discriminatory ability in the entire study cohort (C-index: nontumor related score+Fong, 0.60, nontumor related score+m-clinical risk scores, 0.61, nontumor related score+GAME, 0.64), as well reclassification improvement (42.5, 42.7%, and 21.2%, respectively).
Nontumor related prognostic information may help improve the prognostic stratification of patients after resection of colorectal cancer liver metastasis. The nontumor related prognostic score may be combined with tumor-related prognostic tools to enhance prognostic stratification of patients with colorectal cancer liver metastasis.
基于肿瘤相关因素对结直肠癌肝转移患者进行预后分层的能力只有中等程度的区分能力。我们假设纳入非肿瘤相关因素可以提高对结直肠癌肝转移患者长期预后的预测能力。
利用来自 2 个美国机构的训练队列（n=1205）评估非肿瘤相关实验室标志物。与预后相关的独立因素被用来开发非肿瘤相关的预后评分。通过 Harrell 的 C 统计量（C 指数）和净重新分类改善来评估判别能力，并在来自 5 个亚洲（n=1307）和 3 个欧洲（n=1058）机构的外部验证队列中与 3 种常用的肿瘤相关临床风险评分（Fong 临床风险评分、m-clinical 风险评分和遗传与形态学评估（GAME）评分）进行验证比较。还估计了非肿瘤相关预后评分与这 3 种肿瘤相关预后评分中的每一种结合的判别能力。
碱性磷酸酶（风险比 1.43；95%置信区间，1.11-1.84）、白蛋白（风险比 0.71；95%置信区间，0.57-0.89）和平均红细胞体积（风险比 19.0，每对数单位；95%置信区间，4.79-75.0）均与结直肠癌肝转移切除后死亡风险增加独立相关（均 P<0.05）。反过来，碱性磷酸酶、白蛋白和平均红细胞体积被组合成一个非肿瘤相关的预后评分（2.942×平均红细胞体积+0.399×碱性磷酸酶-0.339×白蛋白-12）×10（中位数为 16；范围为 1-30）。非肿瘤相关的预后评分在外部队列中具有较好到中等的判别能力（C 指数=0.58），与 3 种已建立的肿瘤相关预后评分（C 指数：Fong 临床风险评分，0.53、m-clinical 风险评分，0.55、GAME，0.58）相当。在整个研究队列中，加入非肿瘤相关的预后评分可提高肿瘤相关预后评分的判别能力（C 指数：非肿瘤相关评分+Fong，0.60、非肿瘤相关评分+m-clinical 风险评分，0.61、非肿瘤相关评分+GAME，0.64），以及重新分类改善（分别为 42.5%、42.7%和 21.2%）。
非肿瘤相关的预后信息可能有助于改善结直肠癌肝转移患者切除后的预后分层。非肿瘤相关的预后评分可以与肿瘤相关的预后工具相结合，以增强结直肠癌肝转移患者的预后分层。
