Masked Registration and Autoencoding of CT Images for Predictive Tibia Reconstruction

2025-12-10

AI Summary

Preoperative planning for complex tibial fractures lacks patient-specific ideal reduction templates, hindering accurate surgical guidance. Method: We propose a mask-robust neural registration–autoencoding joint modeling framework. It introduces a novel 3D adaptive spatial transformer network for global rigid-deformable co-registration and uniquely extends both registration and autoencoding to mask-guided input scenarios, enabling fracture-region-driven healthy bone reconstruction. We systematically evaluate multiple 3D autoencoder architectures and introduce mask-aware training with a joint prototype learning framework. Results: On clinical CT data, our method achieves registration error <1.2 mm and reconstruction SSIM >0.91. It significantly improves surgeons’ spatial understanding of 3D reduction targets and enhances preoperative planning efficiency.

Technology Category

• Machine Learning: Deep Generative Models & Autoencoders
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Computer Vision: 3D Computer Vision

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Surgical planning for complex tibial fractures can be challenging for surgeons, as the 3D structure of the later desirable bone alignment may be diffi- cult to imagine. To assist in such planning, we address the challenge of predicting a patient-specific reconstruction target from a CT of the fractured tibia. Our ap- proach combines neural registration and autoencoder models. Specifically, we first train a modified spatial transformer network (STN) to register a raw CT to a standardized coordinate system of a jointly trained tibia prototype. Subsequently, various autoencoder (AE) architectures are trained to model healthy tibial varia- tions. Both the STN and AE models are further designed to be robust to masked input, allowing us to apply them to fractured CTs and decode to a prediction of the patient-specific healthy bone in standard coordinates. Our contributions include: i) a 3D-adapted STN for global spatial registration, ii) a comparative analysis of AEs for bone CT modeling, and iii) the extension of both to handle masked inputs for predictive generation of healthy bone structures. Project page: https://github.com/HongyouZhou/repair

Problem

• Predict patient-specific tibia reconstruction from fractured CT scans
• Combine neural registration and autoencoders for robust bone modeling
• Handle masked inputs to generate healthy bone structure predictions

Innovation

• Masked registration using spatial transformer network
• Autoencoder modeling of healthy tibia variations
• Robust prediction from fractured CT to healthy bone