Establishment a nomogram model for preoperative prediction of the risk of cholangiocarcinoma with microvascular invasion

Objectives

The research aimed to create and verify a nomogram model that can predict the likelihood of cholangiocarcinoma with microvascular invasion (MVI).

Methods

The clinical data of 476 patients with surgically confirmed cholangiocarcinoma were collected retrospectively. This included 240 cases of intrahepatic cholangiocarcinoma (iCCA), 85 cases of perihilar cholangiocarcinoma (pCCA), and 151 cases of extrahepatic cholangiocarcinoma (eCCA). Using this data, we conducted forward multivariate regression analysis to identify the factors that influence the risk of preoperative MVI in patients with cholangiocarcinoma. And using these variables, we developed three nomogram models.

Results

The variables in the model for predicting MVI of iCCA were lymph node metastasis, distant metastases, carcinoembryonic antigen, and tumor size, all of which had a significance level of P < 0.05. The internal and external validation consistency index (C-index) were 0.831 and 0.781, respectively. The variables in the model for predicting MVI of pCCA were lymph node metastasis, carcinoembryonic antigen, and tumor size, all of which had a significance level of P < 0.05. The internal and external validation consistency index (C-index) were 0.791 and 0.747. And the variables in eCCA were lymph node metastasis, distant metastases, carcinoembryonic antigen, and tumor size, all of which had a significance level of P < 0.05. The internal and external validation consistency index (C-index) were 0.834 and 0.830.

Conclusions

we have developed and validated a preoperative nomogram model for predicting MVI in patients with iCCA, pCCA, and eCCA.