TotalRegistrator: Towards a Lightweight Foundation Model for CT Image Registration

2025-08-06

AI Summary

Existing CT registration methods are typically designed for single organs, exhibiting poor generalizability and limited transferability across anatomical regions. To address this, we propose a lightweight, unified registration framework—the first to enable simultaneous, multi-organ registration across the thoracic, abdominal, and pelvic regions. Methodologically, we introduce a novel deformation field decomposition strategy integrated with a standardized U-Net architecture, trained end-to-end on large-scale longitudinal CT data without task-specific fine-tuning. On an internal multi-organ abdominal dataset, our method significantly outperforms baseline approaches; on heterogeneous, multi-center external datasets, it achieves performance comparable to state-of-the-art single-organ models. With only 11 GB GPU memory consumption, the framework balances computational efficiency and strong cross-domain generalization, establishing a new paradigm for clinically deployable, organ-agnostic CT registration.

Technology Category

• Computer Vision: Multi-modal Vision
• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Natural Language Processing: Machine Translation, Multilinguality, Cross-Lingual NLP

Application Category

• Systems and Infrastructure for Web, Mobile and WoT: Experiences and lessons learnt from Web-based algorithms and system deployments
• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

Image registration is a fundamental technique in the analysis of longitudinal and multi-phase CT images within clinical practice. However, most existing methods are tailored for single-organ applications, limiting their generalizability to other anatomical regions. This work presents TotalRegistrator, an image registration framework capable of aligning multiple anatomical regions simultaneously using a standard UNet architecture and a novel field decomposition strategy. The model is lightweight, requiring only 11GB of GPU memory for training. To train and evaluate our method, we constructed a large-scale longitudinal dataset comprising 695 whole-body (thorax-abdomen-pelvic) paired CT scans from individual patients acquired at different time points. We benchmarked TotalRegistrator against a generic classical iterative algorithm and a recent foundation model for image registration. To further assess robustness and generalizability, we evaluated our model on three external datasets: the public thoracic and abdominal datasets from the Learn2Reg challenge, and a private multiphase abdominal dataset from a collaborating hospital. Experimental results on the in-house dataset show that the proposed approach generally surpasses baseline methods in multi-organ abdominal registration, with a slight drop in lung alignment performance. On out-of-distribution datasets, it achieved competitive results compared to leading single-organ models, despite not being fine-tuned for those tasks, demonstrating strong generalizability. The source code will be publicly available at: https://github.com/DIAGNijmegen/oncology_image_registration.git.

Problem

• Develops lightweight model for multi-organ CT image registration
• Addresses limited generalizability of single-organ registration methods
• Evaluates robustness on diverse datasets without task-specific fine-tuning

Innovation

• Lightweight UNet for multi-organ registration
• Novel field decomposition strategy
• Large-scale whole-body CT dataset