ML-Inferred Normal Bone Model for Surgical Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current orthopedic surgery techniques face challenges in accurately determining the material to be removed during procedures for conditions like femoral acetabular impingement, especially when using robotic assistance, due to limited accessibility and the need for precise surgical planning.
Innovation Solution
A method utilizing machine learning models, specifically convolutional neural networks, to infer a normalized bone representation from an abnormal bone, identifying regions of deformity, and generating a surgical plan for altering the abnormal bone by comparing anatomical features, which can be used to control surgical tools for precise bone reshaping.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional surgical techniques are used for bone deformity correction, then the surgical procedure can be performed with conventional tools, but the accuracy in determining material to be removed is insufficient
Solution Approach 1:
The patent creates a digital copy (3D model) of the patient's abnormal bone structure from medical images. This digital replica allows for precise measurement and planning of material removal without requiring complex physical measurement tools during surgery. The virtual model serves as an accurate template for determining exactly what material needs to be removed.
Solution Approach 2:
The surgical plan is created in advance by comparing the abnormal bone model with a normalized healthy bone model. This preliminary planning identifies the exact regions of deformity and calculates the precise material removal required before the actual surgery begins, eliminating the need for complex real-time decision-making during the procedure.
2Manufacturing precision
If robotic assistance is used in orthopedic surgery, then the precision of bone reshaping can be improved, but the difficulty in accessing the bone increases due to limited surgical approach
Solution Approach 1:
The patent introduces a robotic arm as an intermediary tool that can access difficult-to-reach bone regions through narrow surgical incisions. The robotic system translates the pre-planned material removal paths into precise physical actions, allowing complex bone reshaping in areas that would be inaccessible to conventional surgical tools.
Solution Approach 2:
The patent replaces manual surgical techniques with a robotic system that executes pre-planned bone reshaping. The robotic arm with its controlled cutting tool substitutes for the surgeon's hand, enabling precise material removal in hard-to-reach areas while maintaining ease of operation through automated control.
3Ease of manufacture
If statistical modeling is used to model normal bone anatomy, then the modeling process can be simplified, but the accuracy of the modeled anatomy decreases
Solution Approach 1:
Instead of using statistical averages to represent normal bone anatomy, the patent creates an actual copy of the patient's specific healthy bone structure. The normalized bone model is derived directly from the patient's own imaging data, preserving their unique anatomical characteristics while removing only the pathological portions. This patient-specific approach maintains high accuracy without requiring complex statistical modeling.
Data Source
AI summary
The present disclosure provides a machine learning model to model a normal version of a bone from an abnormal version of the bone. The machine learning model can be trained with a training set including abnormal bone images and corresponding normalized, or post-operative, bone images. The abnormal bone image and the inferred normal bone image can be used to plan a surgery to correct the abnormal bone with a surgical navigation system.