A preoperative prognostic model to predict surgical success in patients with perihilar cholangiocarcinoma

J Surg Oncol. 2018 Sep;118(3):469-476. doi: 10.1002/jso.25174. Epub 2018 Aug 21.

Abstract

Background: Patients with resectable perihilar cholangiocarcinoma (PHC) on imaging have a substantial risk of metastatic or locally advanced disease, incomplete (R1) resection, and 90-day mortality. Our aim was to develop a preoperative prognostic model to predict surgical success, defined as a complete (R0) resection without 90-day mortality, in patients with resectable PHC on imaging.

Study design: Patients with PHC who underwent exploratory laparotomy in three tertiary referral centers were identified. Multivariable logistic regression was performed to identify preoperatively available prognostic factors. A prognostic model was developed using data from two European centers and validated in one American center.

Results: In total, 671 patients with PHC underwent exploratory laparotomy. In the derivation cohort, surgical success was achieved in 102 of 331 patients (30.8%). No resection was performed in 176 patients (53.2%) because of metastatic or locally advanced disease. Of the 155 patients (46.8%) who underwent a resection, 38 (24.5%) had an R1-resection. Of the remaining 117 (35.3%), 15 (12.8%) had 90-day mortality. Independent poor prognostic factors for surgical success were identified, and a preoperative prognostic model was developed with a concordance index of 0.71. External validation showed good concordance (0.70).

Conclusion: Surgical success was achieved in only 30% of patients with PHC undergoing exploratory laparotomy and could be predicted by age, cholangitis, hepatic artery involvement, lymph node metastases, and Blumgart stage.

Keywords: Klatskin; perihilar cholangiocarcinoma (PHC); prognostic model.

MeSH terms

  • Aged
  • Bile Duct Neoplasms / pathology
  • Bile Duct Neoplasms / surgery*
  • Female
  • Follow-Up Studies
  • Hepatectomy*
  • Humans
  • Klatskin Tumor / pathology
  • Klatskin Tumor / surgery*
  • Male
  • Middle Aged
  • Models, Statistical*
  • Preoperative Care*
  • Prognosis
  • Retrospective Studies
  • Risk Assessment