AI Respiratory Cycle Prediction for Medical Scan Triggering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for acquiring imaging data during medical scans, such as MRI, often require breath-holds, which can be challenging for certain patient groups, leading to blurred scans due to respiratory motion, and existing free-breathing techniques require long exposition times and laborious image processing.
Innovation Solution
A computing device that uses artificial intelligence to predict respiratory cycles and determine optimal triggering times for data acquisition, reducing respiratory-associated motion by triggering data acquisition around end-expiratory or end-inspiratory peaks, thereby improving scan quality.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If breath-holds are used to limit body motion during scanning, then image quality is improved, but the method becomes challenging or impossible for certain patient groups (pediatric, geriatric, psychiatric, or patients with respiratory conditions)
Solution Approach 1:
The system performs a learning phase before the actual scan to characterize the patient's respiratory pattern by detecting maxima and minima of the respiratory signal. This preliminary characterization enables the system to predict future respiratory cycles and determine optimal triggering times without requiring the patient to perform breath-holds during the scan.
Solution Approach 2:
The trigger threshold is updated regularly based on the detected respiratory signal, typically at the conclusion of each respiratory cycle. This feedback mechanism allows the system to adapt to variations in the patient's breathing pattern and maintain accurate triggering times throughout the scan, ensuring consistent image quality despite natural breathing variations.
2Ease of operation
If free-breathing techniques with motion state extraction are used, then patient compliance is improved, but exposition time increases and image processing becomes laborious
Solution Approach 1:
The system performs a learning phase before the actual scan to characterize the patient's respiratory pattern by detecting maxima and minima of the respiratory signal. This preliminary characterization enables the system to predict future respiratory cycles and determine optimal triggering times without requiring the patient to perform breath-holds during the scan.
Solution Approach 2:
The trigger threshold is updated regularly based on the detected respiratory signal, typically at the conclusion of each respiratory cycle. This feedback mechanism allows the system to adapt to variations in the patient's breathing pattern and maintain accurate triggering times throughout the scan, ensuring consistent image quality despite natural breathing variations.
3Device complexity
If conventional trigger algorithms with heuristic approaches are used, then device complexity is reduced, but measurement precision of respiratory pattern determination deteriorates
Solution Approach 1:
The system performs a learning phase before the actual scan to characterize the patient's respiratory pattern by detecting maxima and minima of the respiratory signal. This preliminary characterization enables the system to predict future respiratory cycles and determine optimal triggering times without requiring the patient to perform breath-holds during the scan.
Solution Approach 2:
The trigger threshold is updated regularly based on the detected respiratory signal, typically at the conclusion of each respiratory cycle. This feedback mechanism allows the system to adapt to variations in the patient's breathing pattern and maintain accurate triggering times throughout the scan, ensuring consistent image quality despite natural breathing variations.
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
A computing device for providing triggering times to a medical scanning apparatus adapted to perform triggered imaging data acquisition, said computing device comprising an input data interface, configured to obtain real-time data of respiratory cycles of a patient; a computation module, configured to implement an artificial intelligence entity, which is trained and adapted to generate a prediction of a number of respiratory cycles of the patient based on the obtained real-time data; a triggering module, configured to determine triggering times corresponding to the predicted respiratory cycles; and an output data interface, configured to output the determined triggering times to the medical scanning apparatus.