Eur Radiol: Use MRI texture analysis to identify immune esophageal cancer in the liver and predict survival

Objective: Clinical evidence shows that the response of immune checkpoint blocking depends on the immune status of tumor micro-environment.
study aims to predict immunosytotypes (IPs) and total lifetimes (OS) in patients with endoheid bile tube cell carcinoma (ICC) through preoperative magnetic resonance imaging (MRI) texture analysis.
: This study included a total of 78 ICC patients, divided into inflammatory type (n s 26) or non-inflammatory type (n s 52) immune estypes according to the density of CD8 plus T cells.
texture analysis using an arterial-enhanced T1-weighted MRI.
application logic regression analysis to select meaningful features related to IP.
OS-related characteristics are determined by the Cox Scale Risk Model and Kaplan-Meier analysis.
the IP and OS prediction models were developed using the selected features.
results: three small waves and a 3D feature with good IP recognition, the combination of which performed best, AUC is 0.919.
of the immune esmic of the inflammatory type is better than that of the non-inflammatory type.
5-year survival rates of the two groups were 48.5% and 25.3%, respectively.
unique wavelet-HLH_firstorder_Median characteristics are associated with OS and are used to build OS prediction models, with a C index of 0.70 (95% CI, 0.57, 0.82), which is a good way to divide ICC patients into high-risk and low-risk groups.
one-year, three-year and five-year survival rates in the two groups were 62.5 per cent and 30.0 per cent, 24.2 per cent and 89.5 per cent, 62.2 per cent and 42.1 per cent, respectively.
: MRI texture characteristics can be used as potential predictive biomarkers of IP and OS in ICC patients.
。