Bony Landmark Classification for Orthopedic Shoulder Surgery Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Planning orthopedic shoulder surgeries is complicated due to difficulties in determining soft tissue qualities and properties around the shoulder joint, which are crucial for deciding between different surgical procedures, and existing imaging techniques like CT and MRI are computationally intensive and costly.
Innovation Solution
A surgical assistance system generates bony landmark data from medical imaging of bones to characterize relationships between landmarks, using a classifier algorithm trained on training data to classify surgical options, thereby serving as a proxy for soft tissue information, reducing computational demands and costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If CT or MRI imaging techniques are used to obtain soft tissue information, then measurement precision of soft tissue qualities is improved, but device complexity and cost increase
Solution Approach 1:
The patent creates a simplified copy of soft tissue information by using bony landmark relationships as a proxy. Instead of directly imaging soft tissue with complex CT or MRI systems, the system captures bone geometry through simpler imaging and uses machine learning to infer soft tissue qualities from these bony landmarks, effectively copying the essential diagnostic information in a more accessible format
Solution Approach 2:
The patent introduces bony landmarks as an intermediary between the imaging system and soft tissue assessment. Rather than directly measuring soft tissue properties, the system measures easily obtainable bony landmarks and uses them as intermediaries to infer soft tissue qualities through trained classification algorithms, bridging the gap between simple bone imaging and complex soft tissue evaluation
2Measurement precision
If CT or MRI imaging techniques are used to obtain soft tissue information, then measurement precision of soft tissue qualities is improved, but cost increases
Solution Approach 1:
The patent creates a simplified copy of soft tissue information by using bony landmark relationships as a proxy. Instead of directly imaging soft tissue with complex CT or MRI systems, the system captures bone geometry through simpler imaging and uses machine learning to infer soft tissue qualities from these bony landmarks, effectively copying the essential diagnostic information in a more accessible format
Solution Approach 2:
The patent replaces expensive, resource-intensive imaging modalities with a cheaper alternative. By using standard X-ray or simple CT imaging of bones combined with computational analysis, the system achieves soft tissue assessment capability at a fraction of the cost of direct MRI or advanced CT imaging, making the diagnostic tool more economically sustainable
3Reliability
If direct soft tissue imaging is performed, then reliability of surgical planning is improved, but productivity decreases due to computational intensity
Solution Approach 1:
The patent performs preliminary action by pre-training classification algorithms on large datasets of bony landmark measurements and corresponding soft tissue outcomes. This pre-computational work creates ready-to-use models that can quickly classify new patients without requiring intensive real-time processing, moving the computational burden from the clinical decision moment to the earlier training phase
Solution Approach 2:
The patent extracts only the essential features needed for soft tissue assessment from full imaging datasets. By identifying and isolating specific bony landmark measurements that correlate with soft tissue qualities, the system removes unnecessary computational complexity while retaining the diagnostic signal, processing only the critical geometric features rather than entire volumetric datasets
Data Source
AI summary
A computing system generates, based on medical imaging data of bones of a joint of a patient, bony landmark data that characterizes relationships between two or more landmarks on one or more of the bones of the joint of the patient. Additionally, the computing system applies a classifier algorithm that has been trained using training data to select a class associated with the patient from among a plurality of classes. The classifier algorithm takes the bony landmark data of the patient as input.