Bladder Motion Compensation in MR Imaging via Deformable Registration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional MR imaging methods face challenges in bladder cancer detection due to low Signal-to-Noise Ratio (SNR) in short-time scanning and motion artifacts from bladder motion in long-time scanning, which are invasive, costly, and limited in field of view.
Innovation Solution
An MR imaging method using short-time scan acquisition and deformable registration to compensate for bladder motion, involving the acquisition of multiple short-time MR images, selection of a reference image, definition of transformation images, registration, and calculation of a motion-corrected image through averaging aligned images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If long-time scanning is used to improve SNR, then image quality improves, but motion artifacts increase due to bladder motion
Solution Approach 1:
The patent divides the long scanning process into multiple short-time scans, each capturing a brief moment in time. By segmenting the acquisition into discrete temporal segments and subsequently registering them, the method avoids motion artifacts while maintaining acceptable SNR through computational integration of the segmented data.
Solution Approach 2:
The patent performs deformable registration on the short-time scan images before averaging them. This preliminary alignment action compensates for motion that occurs during scanning, ensuring that when the images are averaged, the anatomical structures are properly aligned, thus preventing motion artifacts in the final composite image.
2Object-affected harmful factors
If short-time scanning is used to reduce motion artifacts, then motion artifacts decrease, but SNR decreases
Solution Approach 1:
The patent combines multiple short-time scan images through deformable registration and averaging. By merging the signal from multiple short scans while aligning them to a reference image, the method accumulates signal strength (improving SNR) while maintaining the motion artifact benefits of short scanning through the registration process.
Solution Approach 2:
The patent performs deformable registration as a preliminary step before averaging the short-time scan images. This preliminary alignment ensures that when the low-SNR images are combined, their signals constructively interfere rather than cancel out, thereby improving the SNR of the final composite image while preserving the motion artifact reduction benefits.
3Measurement precision
If conventional fiber-optic cystoscopy is used to improve diagnostic accuracy, then diagnostic accuracy improves, but invasiveness and cost increase
Solution Approach 1:
The patent creates a virtual copy of the cystoscopic examination using non-invasive MR imaging. By using deformable registration to warp the MR images to match the anatomy, the method generates a detailed visual representation of the bladder that serves as a substitute for invasive cystoscopy, achieving similar diagnostic accuracy without physical invasion.
Solution Approach 2:
The patent replaces the mechanical invasive cystoscope with a non-invasive magnetic resonance imaging system. By using computational image processing and deformable registration algorithms, the system substitutes the physical mechanical examination with a field-based imaging approach that achieves comparable diagnostic results without invasion or infection risk.
Data Source
AI summary
Provided is a method for compensating for tissue motion during magnetic resonance (MR) imaging, and an apparatus for use thereof. The method includes acquiring a plurality of short-time MR scan images; selecting a reference scan image from the acquired plurality of short-time MR scan images; defining a set of transformation images based on the acquired plurality of short-time MR scan images other than the selected reference scan image; registering the reference scan image and the defined set of transformation images; calculating an average of aligned, registered images of the defined set of transformation images; and generating a motion-corrected image based on the calculated average.