Quantitative Material Characterization of an Object with Magnetic Resonance Imaging

Simple SummaryContent extracted from patent full text and abstract with AI.

This invention is a computer-implemented method that extracts quantitative tissue properties—such as T1 and T2 relaxation times and proton density—from standard MRI scans without requiring specially designed acquisition sequences. It works by repeatedly comparing MRI data simulated from a virtual tissue model against two or more real MRI scans taken with different acquisition settings, then iteratively adjusting the model until the simulated and real data match closely. The final, optimized model directly yields spatially resolved maps of the tissue's physical parameters. Because it can use routine clinical MRI scans as input, no additional or specialized scanning protocols are needed.

Use CasesContent extracted from patent full text and abstract with AI.

• Deriving quantitative T1, T2, and proton-density maps from routine clinical brain MRI scans (e.g., T1-weighted and T2-weighted sequences already acquired for diagnosis) without extra scan time.
• Enabling multi-site and multi-scanner comparability of tissue measurements in neurological studies by producing objective, scanner-independent parameter maps from standard hospital protocols.
• Supporting radiological assessment of white matter diseases, tumors, or neurodegeneration by providing quantitative biomarkers (e.g., T2* or apparent diffusion coefficient maps) derived from existing qualitative scans.
• Retrospective quantitative analysis of archived MRI datasets where only conventional qualitative images and their sequence parameters were stored, enabling new quantitative insights without re-scanning patients.
• Integration into MRI system software to automatically generate quantitative tissue maps as a post-processing step after any standard clinical acquisition, augmenting routine radiology workflows.
• Accelerating clinical adoption of quantitative MRI by removing the need for dedicated, time-consuming qMRI protocols such as inversion-recovery T1 mapping or multi-echo T2 mapping.

BenefitsContent extracted from patent full text and abstract with AI.

• Eliminates the need for dedicated quantitative MRI acquisition sequences, allowing quantitative tissue characterization from standard clinical scans and saving valuable scanner time.
• Improves reproducibility and objectivity of tissue measurements compared to conventional qualitative MRI, facilitating reliable comparisons across patients, scanners, and imaging sites.
• Using at least two complementary acquisition sequences in the optimization significantly reduces the ill-posedness of the inverse problem, increasing accuracy of the recovered material parameters.
• Working directly in k-space (raw data) avoids image reconstruction artifacts and reduces computational memory and processing demands.
• A differentiable MRI simulator enables efficient gradient-based optimization, making the iterative model fitting computationally tractable for spatially resolved 3-D parameter maps.
• The framework is extensible to any number of additional acquisition sequences, allowing accuracy to be progressively improved by incorporating more complementary MRI contrasts without changing the core algorithm.

Technical Classifications (CPCs)

Inventors & Applicants

Patent Abstract

For material characterization, first and second measured MRI data (20, 23, 26) representing an object (6) and corresponding first and second sequence descriptions (21, 24, 27) are received. For each iteration, first and second simulated MRI data (22, 25, 28) is generated according to the sequence descriptions (21, 24, 27) based on a model (29) comprising model values for at least one material parameter. An error (31) is determined, which depends on a deviation of the simulated MRI data (22, 25, 28) from the measured MRI data (20, 23, 26). The model (29) is adapted depending on the error (31). For an initial iteration, the model (29) corresponds to an initial model (29) and otherwise it corresponds to the adapted model (29) of the preceding iteration. The quantitative material characterization is determined depending on the adapted model (29) of a final iteration of the two or more iterations.