Risk Prediction in Cancer Imaging Using Enriched Radiomics Features

2026-03-06

AI Summary

This study addresses the limitation of conventional radiomics in adequately capturing intratumoral heterogeneity in hepatocellular carcinoma by proposing a novel integrative framework that combines structural and functional radiomic features. Specifically, it fuses classical structural features extracted from T1-weighted MRI with pixel-level hemodynamic perfusion characteristics derived from enhancement pattern maps (EPMs) to construct an enriched imaging representation. The method employs L1-regularized logistic regression for cross-sectional diagnosis and Bayesian tensor response regression for longitudinal spatiotemporal modeling, enabling automatic feature weighting and integration. Experimental results demonstrate that the proposed framework significantly outperforms existing approaches in both tumor diagnosis (AUC = 0.96, sensitivity > 0.8) and grading (AUC = 0.87, sensitivity = 0.8). Moreover, aggressive lesions exhibit a significantly higher median proportion of regions with decreased EPM values (0.12) compared to stable lesions (0.025), confirming its sensitivity to underlying tumor biological behavior.

Technology Category

• Machine Learning: Feature Construction/Reformulation
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• 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
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs

Abstract

Background: We aim to develop enriched radiomics features that integrate classical structural radiomics with novel functional radiomics derived from liver MRI for diagnosis and risk stratification in liver cancer. The proposed framework leverages enhancement pattern mapping (EPM) images to provide an automated and robust radiomics representation that captures intratumoral heterogeneity through pixel-level functional information. Methods: Pixel-wise EPM data reflecting blood perfusion were extracted from T1-weighted MRI scans. Classical structural radiomics features were extracted via existing software such as PyRadiomics. In addition, empirical quantiles of EPM values over all pixels within the image, and then smoothed using suitable basis. The smoothed quantiles, along with the classical structural quantiles, are used as functional radiomics features for diagnostic classification and tumor grade stratification, using L1-penalized logistic model that automatically downweights the contribution of the irrelevant features. Further, we conducted longitudinal analyses using Bayesian tensor response regression, which enables spatial smoothing and parsimonious modeling of temporally evolving imaging patterns. Results: The enriched radiomics features illustrate higher diagnostic classification performance (AUC=0.96, sensitivity>0.8) and superior tumor grade stratification accuracy (AUC=0.87, sensitivity=0.8) compared to alternate radiomics features. Moreover, we find that the proportion of lesion pixels with significant reduction in EPM values over time is considerably higher (median = 0.12) in aggressive lesions versus stable or mildly aggressive lesions (median = 0.025). Conclusion: The enriched novel radiomics features can potentially replace classical radiomics analysis and be used for imaging biomarkers in cross-sectional and in longitudinal cancer imaging studies.

Problem

• radiomics
• liver cancer
• risk stratification
• intratumoral heterogeneity
• MRI

Innovation

• enriched radiomics
• enhancement pattern mapping
• functional radiomics
• Bayesian tensor regression
• intratumoral heterogeneity