Table 2 Comparison of traditional and nomogram models for hCCA.

From: Development and validation of a machine learning-based nomogram for survival prediction of patients with hilar cholangiocarcinoma after curative-intent resection

Year

Method

Objective

Span (years)

Sample size

Sample source

Variables included in nomogram/model

C- index (training /testing set)

References

2016

Cox regression analysis

OS

9

235

Single center

Age, preoperative CA19-9 levels, portal vein involvement, hepatic artery invasion, lymph node metastases and surgical treatment outcomes (R0 or R1/2)

0.680 and 0.650

39

2021

Cox regression analysis

OS

5

319 and 109

SEER database and single center

Age, T stage, tumor size and LODDS

0.695 and 0.688

40

2021

Cox regression analysis

OS

12

1173

SEER database

Age, T stage, radiation, chemotherapy, tumor size and LODDS

0.665 and 0.650

41

2022

Cox regression analysis

OS

15

806

SEER database

Age, tumor size, tumor grade, lymph node ratio and T stage parameters

0.655 and 0.626

42

OS

11

340

Our study

The TNM staging system

0.621 and 0.612

OS

11

340

Our study

The Bismuth–Corlette classification

0.531 and 0.487

2024

ML

OS

11

340

Our study

positive margin, N stage, TLNC and tumor differentiation

0.731 and 0.714

  1. SEER surveillance, epidemiology, and end results, LODDS log odds of metastatic lymph nodes.