Registration of Longitudinal Liver Examinations for Tumor Progress Assessment

2025-01-24

AI Summary

In longitudinal CT imaging for hepatocellular carcinoma, inaccurate registration and anatomical/pathological variability lead to erroneous tumor volume estimation. Method: We propose a purely anatomy-guided non-rigid registration paradigm that leverages only geometric and anatomical structural information encoded in liver segmentation masks—explicitly excluding image intensity features—to avoid deformation distortion in tumor regions inherent to conventional feature-driven approaches. Our method integrates a geometry-constrained module, designed from segmentation masks, into a deep registration network, jointly optimized with smoothness regularization and a tumor-burden consistency loss. Results: Evaluated on 317 training and 53 test cases, our approach reduces overall tumor volume error by 37% and improves deformation field smoothness by 2.1× over state-of-the-art methods, significantly enhancing the reliability and clinical applicability of tumor burden assessment.

Technology Category

• Computer Vision: Segmentation
• Machine Learning: Learning with Manifolds
• Search and Optimization: Non-convex Optimization

Application Category

• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation
• Graph Algorithms and Modeling for the Web: Efficient manipulation of static and dynamic Web-related graphs
• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking

Abstract

Assessing cancer progression in liver CT scans is a clinical challenge, requiring a comparison of scans at different times for the same patient. Practitioners must identify existing tumors, compare them with prior exams, identify new tumors, and evaluate overall disease evolution. This process is particularly complex in liver examinations due to misalignment between exams caused by several factors. Indeed, longitudinal liver examinations can undergo different non-pathological and pathological changes due to non-rigid deformations, the appearance or disappearance of pathologies, and other variations. In such cases, existing registration approaches, mainly based on intrinsic features may distort tumor regions, biasing the tumor progress evaluation step and the corresponding diagnosis. This work proposes a registration method based only on geometrical and anatomical information from liver segmentation, aimed at aligning longitudinal liver images for aided diagnosis. The proposed method is trained and tested on longitudinal liver CT scans, with 317 patients for training and 53 for testing. Our experimental results support our claims by showing that our method is better than other registration techniques by providing a smoother deformation while preserving the tumor burden (total volume of tissues considered as tumor) within the volume. Qualitative results emphasize the importance of smooth deformations in preserving tumor appearance.

Problem

• Liver CT Image Alignment
• Tumor Size Misjudgment
• Clinical Decision Impact

Innovation

• Liver Shape-based Registration
• Tumor Change Assessment
• Accurate Tumor Sizing