Development of a nomogram to predict survival in advanced biliary tract cancer

Abstract

The prognosis of advanced biliary tract cancer (BTC) patients remains poor due to limited efficacy of chemotherapy and difficulties in management. Thus, prediction of survival is crucial for the clinical management of advanced BTC. The aim was to develop and validate a nomogram to predict 6-month and 12-month survival in advanced BTC patients treated with chemotherapy. A multivariable Cox regression model was used to construct a nomogram in a training set (JCOG1113, a phase III trial comparing gemcitabine plus S-1 [GS] and gemcitabine plus cisplatin, n = 351). External validity of the nomogram was assessed using a test set (JCOG0805, a randomized, phase II trial comparing GS and S-1 alone, n = 100). Predictive performance was assessed in terms of discrimination and calibration. The constructed nomogram included lymph node metastasis, liver metastasis, carbohydrate antigen 19-9, carcinoembryonic antigen, albumin, and C-reactive protein. Uno’s concordance index was 0.661 (95% confidence interval [CI] 0.629–0.696) in the training set and 0.640 (95% CI 0.566–0.715) in the test set. The calibration plots for 6-month and 12-month survival showed good agreement in the two analysis sets. The present nomogram can facilitate prediction of the prognosis of advanced BTC patients treated with chemotherapy and help clinicians’ prognosis-based decision-making.