Synthetic CT Parameterization for Low-Radiation Scan Planning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current CT scan planning relies on coarse patient measurements and visual inspection, leading to potential overexposure to ionizing radiation, and existing methods to reduce radiation, such as tube current modulation and automatic exposure control, still require visual inspection or scouting scans, which may not accurately represent internal anatomy for precise planning.
Innovation Solution
A method for predicting three-dimensional CT representations from surface data using machine-learned generative networks, incorporating segmentation and landmark information to generate accurate synthetic CT volumes, allowing for precise anatomical estimation and reduced radiation exposure by using depth camera information and parameterization to correct and refine the predicted CT volumes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If visual inspection or scouting scans are used for CT scan planning, then radiation exposure can be controlled through tube current modulation, but the internal anatomy representation remains coarse and inaccurate
Solution Approach 1:
The patent creates a synthetic CT volume (a copy) from surface geometry data using generative adversarial networks. This synthetic volume replicates the internal anatomy information needed for scan planning without requiring actual CT scanning, thereby avoiding ionizing radiation while providing accurate anatomical representation for treatment planning.
Solution Approach 2:
The system performs preliminary action by generating a synthetic CT volume before the actual CT scan. This pre-generated volume provides sufficient anatomical information for scan planning and attenuation correction, eliminating or reducing the need for scouting scans and enabling more accurate tube current modulation based on predicted internal structures.
2Object-affected harmful factors
If tube current modulation or automatic exposure control is used to reduce radiation dose, then radiation exposure is reduced, but visual inspection or scouting scans are still required which limit planning precision
Solution Approach 1:
The synthetic CT volume serves as a radiation-free copy that contains detailed internal anatomy information. This copy enables precise scan planning and automatic exposure control setup without requiring actual scouting scans, thereby achieving both radiation dose reduction and high anatomical detail accuracy simultaneously.
Solution Approach 2:
The synthetic CT volume acts as an intermediary between surface geometry and actual CT scanning. It provides the detailed anatomical information needed for precise planning and accurate radiation dose calculation, enabling better tube current modulation without requiring the patient to undergo additional scanning.
3Measurement precision
If synthetic CT is generated only from surface geometry, then radiation is avoided, but the internal anatomy information is insufficient for specific planning requirements
Solution Approach 1:
The generative adversarial network is segmented into two distinct components: a generator that creates the synthetic CT volume from surface geometry, and a discriminator that evaluates the realism and anatomical accuracy of the generated volume. This segmentation allows each component to be optimized independently, improving overall performance while managing system complexity.
Solution Approach 2:
The discriminator provides feedback to the generator during the training process, evaluating whether the generated CT volume looks realistic and contains accurate anatomical structures. This feedback mechanism enables iterative improvement of the synthetic volume quality, ensuring sufficient internal anatomy information for clinical planning without requiring excessive system complexity.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Synthetic CT is estimated for planning or other purposes from surface data (e.g., depth camera information). The estimation uses parameterization, such as landmark and/or segmentation information, in addition to the surface data. In training and/or application, the parameterization may be used to correct the predicted CT volume. The CT volume may be predicted as a sub-part of the patient, such as estimating the CT volume for scanning one system, organ, or type of tissue separately from other system, organ, or type of tissue.