Computed Tomography‐Based Habitat Analysis for Prognostic Stratification in Colorectal Liver Metastases
ABSTRACT Background Colorectal liver metastasis (CRLM) has a poor prognosis, and traditional prognostic models have certain limitations in clinical application. This study aims to evaluate the prognostic value of CT‐based habitat analysis in CRLM patients and compare it with existing traditional pro...
Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Wiley
2025-04-01
|
| Series: | Cancer Innovation |
| Subjects: | |
| Online Access: | https://doi.org/10.1002/cai2.70000 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | ABSTRACT Background Colorectal liver metastasis (CRLM) has a poor prognosis, and traditional prognostic models have certain limitations in clinical application. This study aims to evaluate the prognostic value of CT‐based habitat analysis in CRLM patients and compare it with existing traditional prognostic models to provide more evidence for individualized treatment of CRLM patients. Methods This retrospective study included 197 patients with CRLM whose preoperative contrast‐enhanced CT images and corresponding DICOM Segmentation Objects (DSOs) were obtained from The Cancer Imaging Archive (TCIA). Tumor regions were segmented, and habitat features representing distinct subregions were extracted. An unsupervised K‐means clustering algorithm classified the tumors into two clusters based on their habitat characteristics. Kaplan–Meier analysis was used to evaluate overall survival (OS), disease‐free survival (DFS), and liver‐specific DFS. The habitat model's predictive performance was compared with the Clinical Risk Score (CRS) and Tumor Burden Score (TBS) using the concordance index (C‐index), Integrated Brier Score (IBS), and time‐dependent area under the curve (AUC). Results The habitat model identified two distinct patient clusters with significant differences in OS, DFS, and liver‐specific DFS (p < 0.01). Compared with CRS and TBS, the habitat model demonstrated superior predictive accuracy, particularly for DFS and liver‐specific DFS, with higher time‐dependent AUC values and improved model calibration (lower IBS). Conclusions CT‐based habitat analysis captures spatial tumor heterogeneity and provides enhanced prognostic stratification in CRLM. The method outperforms conventional models and offers potential for more personalized treatment planning. |
|---|---|
| ISSN: | 2770-9191 2770-9183 |