Department of Surgery, Division of Surgical Oncology, The Ohio State University Wexner Medical Center and James Comprehensive Cancer Center, Columbus, OH, USA.
Department of Surgery, University of Verona, Verona, Italy.
Ann Surg Oncol. 2020 Dec;27(13):5224-5232. doi: 10.1245/s10434-020-08696-z. Epub 2020 Jun 3.
Patients with intrahepatic cholangiocarcinoma (ICC) generally have a poor prognosis, yet there can be heterogeneity in the patterns of presentation and associated outcomes. We sought to identify clusters of ICC patients based on preoperative characteristics that may have distinct outcomes based on differing patterns of presentation.
Patients undergoing curative-intent resection of ICC between 2000 and 2017 were identified using a multi-institutional database. A cluster analysis was performed based on preoperative variables to identify distinct patterns of presentation. A classification tree was built to prospectively assign patients into cluster assignments.
Among 826 patients with ICC, three distinct presentation patterns were noted. Specifically, Cluster 1 (common ICC, 58.9%) consisted of individuals who had a small-size ICC (median 4.6 cm) and median carbohydrate antigen (CA) 19-9 and neutrophil-to-lymphocyte ratio (NLR) levels of 40.3 UI/mL and 2.6, respectively; Cluster 2 (proliferative ICC, 34.9%) consisted of patients who had larger-size tumors (median 9.0 cm), higher CA19-9 levels (median 72.0 UI/mL), and similar NLR (median 2.7); Cluster 3 (inflammatory ICC, 6.2%) comprised of patients with a medium-size ICC (median 6.2 cm), the lowest range of CA19-9 (median 26.2 UI/mL), yet the highest NLR (median 13.5) (all p < 0.05). Median OS worsened incrementally among the three different clusters {Cluster 1 vs. 2 vs. 3; 60.4 months (95% confidence interval [CI] 43.0-77.8) vs. 27.2 months (95% CI 19.9-34.4) vs. 13.3 months (95% CI 7.2-19.3); p < 0.001}. The classification tree used to assign patients into different clusters had an excellent agreement with actual cluster assignment (κ = 0.93, 95% CI 0.90-0.96).
Machine learning analysis identified three distinct prognostic clusters based solely on preoperative characteristics among patients with ICC. Characterizing preoperative patient heterogeneity with machine learning tools can help physicians with preoperative selection and risk stratification of patients with ICC.
肝内胆管癌（ICC）患者预后一般较差，但临床表现和相关结局可能存在异质性。我们试图根据术前特征识别 ICC 患者的聚类，这些特征可能基于不同的表现模式而具有不同的结局。
使用多机构数据库确定 2000 年至 2017 年间接受根治性切除术治疗的 ICC 患者。根据术前变量进行聚类分析，以确定不同的表现模式。建立分类树，前瞻性地将患者分配到聚类中。
在 826 名 ICC 患者中，观察到三种不同的表现模式。具体而言，第 1 组（常见 ICC，58.9%）由肿瘤较小（中位直径 4.6cm）、CA19-9 中位值为 40.3UI/mL 和中性粒细胞与淋巴细胞比值（NLR）中位值为 2.6 的个体组成；第 2 组（增殖性 ICC，34.9%）由肿瘤较大（中位直径 9.0cm）、CA19-9 中位值较高（72.0UI/mL）、NLR 中位值相似（2.7）的患者组成；第 3 组（炎症性 ICC，6.2%）由肿瘤中等大小（中位直径 6.2cm）、CA19-9 中位数最低（26.2UI/mL）、NLR 中位数最高（13.5）的患者组成（均 P<0.05）。三组患者的中位 OS 逐渐恶化，第 1 组优于第 2 组优于第 3 组{第 1 组 vs. 第 2 组 vs. 第 3 组；60.4 个月（95%置信区间 [CI] 43.0-77.8）vs. 27.2 个月（95% CI 19.9-34.4）vs. 13.3 个月（95% CI 7.2-19.3）；P<0.001}。用于将患者分配到不同聚类的分类树与实际聚类分配具有极好的一致性（κ=0.93，95%CI 0.90-0.96）。
机器学习分析仅根据 ICC 患者的术前特征确定了三个不同的预后聚类。使用机器学习工具对术前患者异质性进行特征描述可以帮助医生对 ICC 患者进行术前选择和风险分层。
