A Novel Prognostic Classifier for Intrahepatic Cholangiocarcinoma: Development and External and Prospective Validation Studies

Guang-Zhi Jin,Guan-Zhen Yu,Wen-Ming Cong
DOI: https://doi.org/10.1200/jco.2018.36.15_suppl.e16102
IF: 45.3
2018-01-01
Journal of Clinical Oncology
Abstract:e16102 Background: Predicting the prognosis for intrahepatic cholangiocarcinoma (ICC) is currently inadequate, and its prognostic value could be improved by the identification of suitable biomarkers. This study aimed to develop a novel prognostic classifier for patients with ICC after surgery. Methods: We screened tumour markers that over-expressed in ICC tissues using a proteomics approach. We narrowed down the tumour markers and developed prognostic classifiers based on clinical parameters and tumour markers using a Cox regression analyses from 519 patients (primary training cohort). We validated the predictive accuracy of this classifier in both the external validation cohort of 318 patients, and a prospective validation cohort of 178 patients. Results: We established classifiers based on serum CA199, serum alubumin, tumour size, tumour number, TNM, pyruvate kinase isozymes M1/M2 for overall survival (OS) and tumour size, tumour number, TNM, triosephosphate isomerise for time to recurrence (TTR). Classifiers were able to classify patients into low risk and high risk groups based on risk score. In primary training cohort, compared with patients with low risk scores, patients with high risk scores had shorter OS (p < 0.0001), and TTR (p < 0.0001). The equivalent findings were observed in the external validation cohort for OS (p < 0.0001) and TTR (p < 0.0001) and in the prospective validation cohort for OS (p < 0.0001) and TTR (p < 0.0001). ROC analyses also revealed that accuracy of the classifier was superior to that of any other independent prognostic factor for both OS and TTR in three cohorts. Two nomograms for predicts death or recurrence probability of ICC patient’s were developed, respectively. Conclusions: Our prognostic classifiers are promising predictors for survival or recurrence in ICC patients after surgical resection.
What problem does this paper attempt to address?