Towards MR-Based Trochleoplasty Planning

2025-08-08

AI Summary

Preoperative planning for trochlear dysplasia (TD) relies on low-resolution MRI and subjective expert assessment, resulting in poor surgical consistency and limited minimally invasive applicability. To address this, we propose the first integrated framework combining implicit neural representations (INRs) and wavelet diffusion models (WDMs): INRs enable isotropic super-resolution MRI reconstruction; a multi-label segmentation network precisely localizes lower-limb bony anatomy; and the WDM synthesizes patient-specific, submillimeter-resolution 3D “pseudo-healthy” trochlear geometry directly from routine MRI—eliminating the need for CT and associated ionizing radiation. Validated on 25 clinical cases, our method significantly improves sulcus angle (SA) and trochlear groove depth (TGD) metrics. The generated high-fidelity trochlear models are directly deployable for preoperative planning and intraoperative navigation. This approach enhances the accuracy, reproducibility, and minimally invasive potential of trochleoplasty, establishing a novel paradigm for precision surgical intervention in patellofemoral instability.

Technology Category

• Computer Vision: Motion & Tracking
• Planning, Routing, and Scheduling: Replanning and Plan Repair
• Machine Learning: Multi-instance/Multi-view Learning

Application Category

• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Web Mining and Content Analysis: Large pretrained models with web data
• User Modeling, Personalization and Recommendation: On-Device user modeling, personalization, and recommendation

Abstract

To treat Trochlear Dysplasia (TD), current approaches rely mainly on low-resolution clinical Magnetic Resonance (MR) scans and surgical intuition. The surgeries are planned based on surgeons experience, have limited adoption of minimally invasive techniques, and lead to inconsistent outcomes. We propose a pipeline that generates super-resolved, patient-specific 3D pseudo-healthy target morphologies from conventional clinical MR scans. First, we compute an isotropic super-resolved MR volume using an Implicit Neural Representation (INR). Next, we segment femur, tibia, patella, and fibula with a multi-label custom-trained network. Finally, we train a Wavelet Diffusion Model (WDM) to generate pseudo-healthy target morphologies of the trochlear region. In contrast to prior work producing pseudo-healthy low-resolution 3D MR images, our approach enables the generation of sub-millimeter resolved 3D shapes compatible for pre- and intraoperative use. These can serve as preoperative blueprints for reshaping the femoral groove while preserving the native patella articulation. Furthermore, and in contrast to other work, we do not require a CT for our pipeline - reducing the amount of radiation. We evaluated our approach on 25 TD patients and could show that our target morphologies significantly improve the sulcus angle (SA) and trochlear groove depth (TGD). The code and interactive visualization are available at https://wehrlimi.github.io/sr-3d-planning/.

Problem

• Generates high-resolution 3D pseudo-healthy trochlear morphologies from low-resolution MR scans
• Eliminates need for CT scans, reducing patient radiation exposure
• Improves surgical outcomes by enhancing sulcus angle and groove depth

Innovation

• Uses Implicit Neural Representation for super-resolved MR
• Employs Wavelet Diffusion Model for pseudo-healthy morphologies
• Generates sub-millimeter 3D shapes without needing CT