PLUS: Plug-and-Play Enhanced Liver Lesion Diagnosis Model on Non-Contrast CT Scans

2025-07-04

AI Summary

Accurate benign-malignant differentiation of focal liver lesions (FLLs) on non-contrast CT (NCCT) remains challenging, and existing methods rely heavily on multiphase contrast-enhanced CT or MRI—limiting clinical applicability. Method: We propose PLUS, a plug-and-play enhancement framework that requires no backbone modification and seamlessly integrates with any 3D segmentation model. PLUS enables end-to-end malignancy classification solely from NCCT by jointly optimizing segmentation and classification tasks using large-scale real-world NCCT data. Crucially, it leverages precise 3D lesion segmentation to fuse local contextual features and explicitly model inter-lesion discriminative patterns. Results: Evaluated on 8,651 patient cases, PLUS achieves a 5.66% improvement in lesion-level F1 score, and boosts case-level F1 scores by 6.26% for malignant and 4.03% for benign cases—significantly enhancing the diagnostic utility of routine abdominal NCCT.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Multi-instance/Multi-view Learning
• Natural Language Processing: Safety and Robustness

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• 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

Abstract

Focal liver lesions (FLL) are common clinical findings during physical examination. Early diagnosis and intervention of liver malignancies are crucial to improving patient survival. Although the current 3D segmentation paradigm can accurately detect lesions, it faces limitations in distinguishing between malignant and benign liver lesions, primarily due to its inability to differentiate subtle variations between different lesions. Furthermore, existing methods predominantly rely on specialized imaging modalities such as multi-phase contrast-enhanced CT and magnetic resonance imaging, whereas non-contrast CT (NCCT) is more prevalent in routine abdominal imaging. To address these limitations, we propose PLUS, a plug-and-play framework that enhances FLL analysis on NCCT images for arbitrary 3D segmentation models. In extensive experiments involving 8,651 patients, PLUS demonstrated a significant improvement with existing methods, improving the lesion-level F1 score by 5.66%, the malignant patient-level F1 score by 6.26%, and the benign patient-level F1 score by 4.03%. Our results demonstrate the potential of PLUS to improve malignant FLL screening using widely available NCCT imaging substantially.

Problem

• Distinguishing malignant and benign liver lesions on NCCT scans
• Improving liver lesion diagnosis without contrast-enhanced imaging
• Enhancing 3D segmentation models for better FLL classification accuracy

Innovation

• Plug-and-play framework for liver lesion analysis
• Enhances 3D segmentation on non-contrast CT scans
• Improves malignant and benign lesion differentiation