Finding Reproducible and Prognostic Radiomic Features in Variable Slice Thickness Contrast Enhanced CT of Colorectal Liver Metastases

2025-01-15

Machine Learning for Biomedical Imaging

AI Summary

This study addresses the challenge of poor reproducibility and limited clinical translation of radiomic features from CT scans for precise prognosis prediction in colorectal liver metastases (CRLM), particularly across multi-center settings and varying reconstruction slice thicknesses. We propose a high-reproducibility feature selection strategy based on the concordance correlation coefficient (CCC ≥ 0.85), integrated with multi-parameter image preprocessing and fusion of 93 pyradiomics features—preserving predictive performance while enhancing robustness. Prognostic performance was evaluated using Cox regression and the concordance index (C-index). Validation across multi-center data demonstrated that the high-reproducibility feature subset maintains equivalent discriminative ability (C-index = 0.629 vs. 0.630). To our knowledge, this is the first systematic demonstration that radiomics retains robust prognostic value under stringent reproducibility constraints, thereby substantially improving clinical translatability.

Technology Category

• Machine Learning: Feature Construction/Reformulation
• Computer Vision: Multi-modal Vision
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Establishing the reproducibility of radiomic signatures is a critical step in the path to clinical adoption of quantitative imaging biomarkers; however, radiomic signatures must also be meaningfully related to an outcome of clinical importance to be of value for per- sonalized medicine. In this study, we analyze both the reproducibility and prognostic value of radiomic features extracted from the liver parenchyma and largest liver metastases in contrast enhanced CT scans of patients with colorectal liver metastases (CRLM). A prospective cohort of 81 patients from two major US cancer centers was used to establish the reproducibility of radiomic features extracted from images reconstructed with different slice thicknesses. A publicly available, single-center cohort of 197 preoperative scans from patients who underwent hepatic resection for treatment of CRLM was used to evaluate the prognostic value of features and models to predict overall survival. A standard set of 93 features was extracted from all images using pyradiomics, with a set of eight different extractor settings. Our results show that the feature extraction settings producing the most reproducible, as well as the most prognostically discriminative feature values are highly dependent on both the region of interest and the specific feature in question. While the best overall predictive model was produced using features extracted with a particular setting, without accounting for reproducibility, (C-index = 0.630 (0.603–0.649)) an equivalent-performing model (C-index = 0.629 (0.605–0.645)) was produced by pooling features from all extraction settings, and thresholding features with low reproducibility (CCC ≥ 0.85), prior to feature selection. Our findings support a data-driven approach to feature extraction and selection, preferring the inclusion of many features, and narrowing feature selection based on feature reproducibility when relevant reproducibility data is available. Further research is needed to determine how to select reproducible feature sets when reproducibility data is not available.

Problem

• Colorectal Cancer Hepatic Metastases
• Radiomics Features
• Treatment Response Prediction

Innovation

• Radiomics Features
• Colorectal Liver Metastases
• Multi-feature Processing