Back to results
US 12,136,217 B2Granted
Magnetic-resonance-based method for measuring microscopic histologic soft tissue textures
AssigneebioProtonics, L.L.C
InventorsTimothy W. James, Kristin James
Life of this patent
Estimating the term…
Abstract
A method for measuring soft tissue texture to identify diseased as opposed to normal tissue by identifying textural markers that distinguish diseased tissue from normal tissue and selecting a MRμT excitation sequence and associated parameters to reveal those markers. Data is then acquired in an MR scanner responsive to the selected MRμT excitation sequence to establish a multipoint time series data set. The acquired data is then analyzed for presence of the markers.
Claims9
A method for measuring soft tissue texture to identify diseased as opposed to normal tissue, the method comprising:
- identifying textural markers in a ground truth that distinguish diseased tissue from normal tissue;
- selecting a MRμT excitation sequence comprising a Multiple TR series and associated parameters to reveal said markers;
- acquiring data in an MR scanner responsive to the selected MRμT excitation sequence to establish a multipoint time series data set by performing a selective internal excitation of a Volume of Interest (VOI) within a targeted tissue region;
- imposing a spatial-frequency phase encode for a targeted tissue textural-wavelength, λ (or k-value), of interest along a texture analysis direction within the VOI;
- and
- recording a resulting signal;
- and
- analyzing the acquired data for presence of said markers by
- performing a FFT on the multipoint time series data set acquired to generate an NMR spectrum of signal intensity vs. PPM chemical shift with a system center frequency as 0 PPM;
- identifying a center frequency of the NMR spectrum from a selected TR measurement of the Multiple TR series of excitations of the targeted VOI by plotting and locating the maximum of the NMR signal vs. chemical shift;
- selecting a span in chemical shift and summing the signal over the span to generate a magnitude measurement as a raw signal for each wavelength encode and for the k0 encoded TR;
- selecting the span and summing the signal from an ultimate TR to generate a measure of noise level (noise);
- calculating the signal for each of the encoded wavelength TR increments as
- signal(λ)=(raw signal(λ)-noise)/(raw signal(∞)-noise);
- and
- binning the multiple measurements of the signal to highlight distribution of feature sizes.
13 more elements
A method for obtaining soft tissue texture to identify biomarkers for diseased as opposed to normal tissue in a prostate, the method comprising:
- acquiring data of a signal using a Multiple TR method with a targeted VOI for
- a multi-point time series data set for range of wavelengths each with a single k-value (textural wavelength) encoded during each excitation (TR), wherein subsequent TRs encode different k-values, with spin echo recorded for a full T2* period;
- a normalizing dataset in a penultimate TR at k0;
- and
- a multi-point dataset with no encode or excitation in an ultimate TR to generate a measure of noise;
- performing a FFT on the multi-point time series data set acquired including ko to generate an NMR spectrum of signal intensity vs. PPM chemical shift with the system center frequency as 0 PPM;
- identifying a center frequency of the NMR spectrum from a penultimate TR measurement of the Multiple TR series of excitations of the targeted VOI by plotting and locating the maximum of the NMR signal vs. chemical shift-;
- selecting a span in chemical shift and summing the signal to generate a magnitude measurement as a raw signal for each wavelength encode and for the TR encoded at k0;
- selecting the span of and summing the signal from the ultimate TR with no excitation as the noise level (noise);
- calculating the signal for each of the encoded wavelength TR increments as
- signal(λ)-(raw signal(λ)-noise)/(raw signal(∞)-noise);
- binning the multiple measurements of the signal by segmenting measurements into wavelength bands (bins) and then calculating an average value and range for each wavelength band;
- identifying distinctions based on markers defined in a ground truth between uniform spectra indicating cancerous regions and variable spectra indicating normal tissue;
- and,
- designating feature signal ranges corresponding to the identified distinctions as biomarkers for disease identification.
12 more elements
Dependencies are inferred from the claim text.