DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Priority
Receipt is acknowledged of certified copies of papers submitted under 35 U.S.C. 119(a)-(d), which papers have been placed of record in the file.
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 06/09/2026 has been entered.
Response to Arguments
Applicant’s arguments, filed 06/09/2026, with respect to claims 1-7, 9, 11-15 and 17-18 have been fully considered but are moot because the arguments do not apply to the current references and current combinations of references being used in the current rejection.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102 of this title, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-3, 6-7, 9, 11-13, and 15 are rejected under 35 U.S.C. 103 as being unpatentable over HILBERT et al. (US 20200333414 A1), hereinafter referenced as HILBERT in view of XING et al. (US 20210313046 A1), hereinafter referenced as XING and further in view of BANERJEE et al. (US 20210027436 A1), hereinafter referenced as BANERJEE.
Regarding claim 1, HILBERT explicitly teaches a method for generating a magnetic resonance image (Fig. 1. Paragraph [0025]-HILBERT discloses the present invention proposes to use, for the generation of quantitative maps, a quantitative acquisition strategy which measures quantitative parameters in a way that additional contrast information is sampled together with the quantitative parameters for generating a corresponding synthetic (i.e. simulated) image based on physical signal models), comprising:
performing image conversion on the plurality of quantitative maps on the basis of the plurality of scan parameters (Fig. 2. Paragraph [0027]-HILBERT discloses at step 101, the system uses a first quantitative MRI acquisition technique. From the use of the first quantitative acquisition technique, the system generates a first quantitative map for the first quantitative parameter, e.g. a T1 map. In paragraph [0029]-HILBERT discloses at step 102, the system uses a second quantitative MRI acquisition technique, preferentially a T2 mapping acquisition technique the acquisition of the second quantitative map (wherein the first acquisition technique also generates at the same time a quantitative proton-density map with additional weighting and a quantitative map free of the additional weighting, and the second acquisition technique generates a second proton-density weighted image or quantitative map)) to generate a first converted image and a second converted image (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first quantitative proton-density map, and d) the second quantitative proton-density map, as inputs in a contrast synthetization module which contains a physical signal model given by Eq. (2), wherein the inputs are used to generate a synthetic image M with arbitrary TE, TR and TI. The system contains a user interface with a contrast switch enabling to automatically switch between the first contrast component, the second contrast component and the initial contrast component (i.e. no contrast—initial image) when displaying, at step 104, a synthetic image of the biological object. The system is configured for displaying on a display at least two different contrasts at the same time for the biological object. The fat signal or MT-weighting could be turned on and off by switching between M0.sub.P, M0.sub.W, and M0.sub.M when using equation (4) for the physical signal model used to generate the synthetic image M);
generating a fused image of the first converted image and the second converted image (Fig. 1. Paragraph [0031]-HILBERT discloses at least 3 synthetic images M might be displayed by the system, either at the same time, or by switching from one of the synthetic images to the other one by selecting the appropriate initial magnetization M0.sub.P, M0.sub.W, or M0.sub.M via the contrast switch. Further in paragraph [0034]-HILBERT discloses the obtained maps and images are used as input in a contrast synthetization module 24 of the processing unit 203 (wherein the contrast synthetization module 24 contains a physical signal model (contrast mechanism) as shown in Eq. 4 configured for generating a synthetic image M of the biological object from said inputs). The contrast switch is preferentially configured for enabling a switch between a first synthetic image generated by using the contrast component C.sub.i, with i≥1, and a second synthetic image generated by using the contrast component C.sub.0, in order to switch on/off the corresponding contrast);
and generating a plurality of quantitative weighted images on the basis of the fused image (Fig. 1. Paragraph [0030]-HILBERT discloses the system is configured for displaying on a display at least two different contrasts at the same time for the biological object. Further in paragraph [0034]-HILBERT discloses the user interface 205 is further configured for enabling a user to choose the desired synthetic sequence parameters TE, TR, TI. By means of the user interface 25 and its contrast switch, a user may choose to display any of the weighted contrast on a map of the biological object shown then on the display 204 of the system 200, like T2 weighted image, T2 weighted image WE, T1 weighted image, T1 weighted image WE, PD image, PD WE image, or STIR image. A user may switch between different types of preparation contrast by turning the fat signal and a MT-weighting in synthetic contrasts “on” or “off”).
HILBERT fails to explicitly teach simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters.
However, XING explicitly teaches simultaneously generating a plurality of quantitative maps (Fig. 1A-B and 4A-C. Paragraph [0027]-XING discloses the method may be performed by a conventional MRI scanner using standard imaging protocols, adapted with a neural network to generate the quantitative MRI map(s) from the qualitative image acquired by the scanner using conventional clinical imaging techniques. The deep learning network derives quantitative relaxation parametric maps from a single qualitative MR image) on the basis of a single raw image, the single raw image (Fig. 1B. Paragraph [0025]-XING discloses for T.sub.1 mapping, the method uses a deep neural network 108 to derive quantitative T.sub.1 and proton density maps from a single conventional T.sub.1 weighted image 106 acquired in routine clinical practice, as illustrated in FIG. 1B. With the use of the deep neural network, only one T.sub.1 weighted image 106 is required for the generation of a quantitative T.sub.1 map. Further in paragraph [0026]-XING discloses a T.sub.2 map can be produced from a single T.sub.2 or T.sub.2/T.sub.1 weighted image using a trained deep neural network. In paragraph [0028]-XING discloses pulse sequences included in standard clinical imaging protocols may be used for acquisition of the qualitative image. Therefore, it would have been obvious to a person of ordinary skill in the art to use a single raw image given all Magnetic Resonance Imaging scanners acquire a raw image (i.e. K-space, which is a raw digital matrix that stores signal data) and all modern MRI scanners have both i-channel and q-channels (i.e. real and imaginary components). This would improve the ability to accurately analyze the imaging data) being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters (Fig. 1B. Paragraph [0002]-XING discloses in conventional clinical MRI, a qualitative magnetic resonance (MR) image with a specific weighting (T.sub.1 weighting, T.sub.2 weighting, or other) is acquired using a particular value of every imaging parameter (time of repetition, echo time, flip angle, etc.). In paragraph [0004]-XING discloses traditionally, a quantitative MR image (or map) is obtained from multiple qualitative MR images that are acquired with different values of parameters (time of repetition, echo time, flip angle, spin-lock time) to gain variable contrasts. In paragraph [0028]-XING discloses for T.sub.1 mapping, T.sub.1 weighted images can be acquired using the Spoiled Gradient (SPGR) sequence. For T.sub.2 mapping, T.sub.2/T.sub.1 weighted images can be obtained using the Steady State Free Precession (SSFP) sequence, or T.sub.2 weighted images can be obtained using the Fast Spin Echo image (FSE) sequence. Please also see claim 1-3 and read paragraph [0005, 0018-0021 and 0023]);
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a raw image, the raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters; performing image conversion on the plurality of quantitative maps on the basis of the plurality of scan parameters to generate a first converted image and a second converted image; generating a fused image of the first converted image and the second converted image; and generating a plurality of quantitative weighted images on the basis of the fused image, with the teachings of XING of having simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters.
Wherein HILBERT’s method having simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters.
The motivation behind the modification would have been to obtain a method that improves the generation of synthetic quantitative MRI images and the performance of neural networks, since both HILBERT and XING concern processing quantitative MRI images and mappings. Wherein HILBERT systems and methods improves the ability for radiologists to form diagnoses and improves the generation of synthetic images based on quantitative maps by using additional weightings and providing a large variety of contrasts based on short acquisition times on top of the quantitative information, while XING’s systems and methods improves the accuracy and efficiency for generating quantitative maps, requires a single image as the initial input and implements a neural network architecture that achieves a balance between computational workload and performance. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015 and 0025] and XING et al. (US 20210313046 A1), Abstract and paragraph [0030 and 0039-0040].
Although XING explicitly teaches simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters.
HILBERT in view of XING is silent on wherein the single raw image comprises at least one of a real image or an imaginary image.
However, BANERJEE explicitly teaches wherein the single raw image (Fig. 2. Paragraph [0022]-BANERJEE discloses the present disclosure is related to synthetic MRI. Synthetic MRI can reconstruct multiple image contrasts (e.g., T1- and T2-weighted, T1- and T2-FLAIR, proton density-weighted, STIR images) from MR signals acquired with a quantification sequence (e.g., MDME sequences) in a single scan. In paragraph [0031]-BANERJEE discloses an MR scan is complete when an array of raw k-space data has been acquired and stored. This raw k-space data is rearranged into separate k-space data arrays for each image to be reconstructed. In paragraph [0035]-BANERJEE discloses referring to FIG. 3, a schematic diagram of a single block of a pulse sequence 300 for performing a quantification scan is shown. A two-dimensional (2D) fast spin echo (FSE) multi-delay multi-echo (MDME) sequence may be performed for the quantification scan, which includes an interleaved slice selective saturation RF pulse and multi-echo acquisition. The quantitative acquisition 210 may be raw MRI signals in k-space in some embodiments, such as raw MDME MRI) comprises at least one of a real image or an imaginary image (Fig. 2. Paragraph [0036]-BANERJEE discloses the deep neural network 220 may be a multi-scale U-Net convolutional neural network (CNN) with an encoder-decoder structure. An input layer 221 receives the quantitative acquisition 210. The input layer 221 includes multiple channels to accommodate the input of the multiple images as the result of a quantitative acquisition at once. Each of the multiple images is split up into real and imaginary components and stacked as a corresponding channel).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, with the teachings of BANERJEE of having wherein the single raw image comprises at least one of a real image or an imaginary image.
Wherein HILBERT’s method having simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image.
The motivation behind the modification would have been to obtain a system that enhances processing speed, accuracy and image quality, since HILBERT and BANERJEE concern systems and methods for processing and generating quantitative magnetic resonance maps and images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while BANERJEE provides systems and methods that reduces image artifacts and improves reconstruction accuracy, resolution restoration, and training time. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and BANERJEE et al. (US 20210027436 A1), Abstract and Paragraph [0037, 0042, and 0046].
Regarding claim 2, HILBERT in view of XING and in further view of WANG explicitly teach the method according to claim 1, HILBERT fail to explicitly teach wherein the generating a plurality of quantitative maps on the basis of the single raw image comprises: generating a plurality of quantitative maps by performing deep learning processing on the single raw image on the basis of a first deep learning network, wherein the plurality of quantitative maps comprise at least one of a quantitative T1 value, a quantitative T2 value, and a quantitative proton density value.
However, XING explicitly teaches wherein the generating a plurality of quantitative maps on the basis of the single raw image (Fig. 1B. Paragraph [0024]-XING discloses the inventors have discovered that deep learning enables the acquisition and utilization of generic a priori information to predict quantitative MRI data from a single qualitative MR image) comprises: generating a plurality of quantitative maps by performing deep learning processing (Fig. 1B. Paragraph [0025]-XING discloses for T.sub.1 mapping, the method uses a deep neural network 108 to derive quantitative T.sub.1 and proton density maps from a single conventional T.sub.1 weighted image 106 acquired in routine clinical practice, as illustrated in FIG. 1B. With the use of the deep neural network, only one T.sub.1 weighted image 106 is required for the generation of a quantitative T.sub.1 map. Further in paragraph [0026]-XING discloses a T.sub.2 map can be produced from a single T.sub.2 or T.sub.2/T.sub.1 weighted image using a trained deep neural network. In this way, qualitative and quantitative MR images can be obtained in the routine clinical practice without changing the imaging protocol or performing multiple scans) on the single raw image on the basis of a first deep learning network (Fig. 1B, #108 called a Deep learning network. Paragraph [0036]. Further in paragraph [0040]-XING discloses generative adversarial networks with various architectures may be used), wherein the plurality of quantitative maps comprise at least one of a quantitative T1 value (Fig. 5A-E. Paragraph [0019]-XING discloses FIG. 5A-E show images illustrating prediction of quantitative T.sub.1 map from a T.sub.1 weighted image using T-net), a quantitative T2 value (Fig. 7A-E. Paragraph [0021]-XING discloses FIG. 7A-E shows images illustrating prediction of quantitative T.sub.2 map from a single T.sub.2/T.sub.1 weighted image using T-net), and a quantitative proton density value (Fig. 6A-E. Paragraph [0020]-XING discloses FIG. 6A-E show images illustrating prediction of quantitative proton density (PD) map from a T.sub.1 weighted image using T-net).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of XING of having wherein the generating a plurality of quantitative maps on the basis of the single raw image comprises: generating a plurality of quantitative maps by performing deep learning processing on the single raw image on the basis of a first deep learning network, wherein the plurality of quantitative maps comprise at least one of a quantitative T1 value, a quantitative T2 value, and a quantitative proton density value.
Wherein HILBERT’s method having wherein the generating a plurality of quantitative maps on the basis of the single raw image comprises: generating a plurality of quantitative maps by performing deep learning processing on the single raw image on the basis of a first deep learning network, wherein the plurality of quantitative maps comprise at least one of a quantitative T1 value, a quantitative T2 value, and a quantitative proton density value.
The motivation behind the modification would have been to obtain a method that improves the generation of synthetic quantitative MRI images and the performance of neural networks, since both HILBERT and XING concern processing quantitative MRI images and mappings. Wherein HILBERT systems and methods improves the ability for radiologists to form diagnoses and improves the generation of synthetic images based on quantitative maps by using additional weightings and providing a large variety of contrasts based on short acquisition times on top of the quantitative information, while XING’s systems and methods improves the accuracy and efficiency for generating quantitative maps, requires a single image as the initial input and implements a neural network architecture that achieves a balance between computational workload and performance. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015 and 0025] and XING et al. (US 20210313046 A1), Abstract and paragraph [0030 and 0039-0040].
Regarding claim 3, HILBERT in view of XING and in further view of BANERJEE explicitly teach the method according to claim 1, although HILBERT explicitly teaches wherein the plurality of quantitative images are generated by the fused image (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first contrast component, which is preferentially the first proton-density image or quantitative proton-density map, e.g. the proton-density image M0.sub.P with fat signal, d) the second contrast component, which is preferentially the second proton-density image or quantitative proton-density map, e.g. the proton-density image with additional magnetization transfer contrast M0.sub.M, and e) the initial contrast component, which is preferentially the proton-density image or quantitative proton-density map. Please also read paragraph [0027-0029 and 0034]).
HILBERT fails to explicitly teach wherein the plurality of quantitative images are generated by performing deep learning processing on the fused image on the basis of a second deep learning network.
However, XING explicitly teaches wherein the plurality of quantitative images are generated by performing deep learning processing on the fused image (Fig. 1A-B and 4A-C. Paragraph [0029]-XING discloses after a qualitative image is acquired 800 it is then applied 802 as input to the neural network to obtain the quantitative image. In paragraph [0026]-XING discloses a T.sub.2 map can be produced from a single T.sub.2/T.sub.1 weighted image using a trained deep neural network. Please also read paragraph [0025, 0028 and 0032]) on the basis of a second deep learning network (Fig. 1A-B and 4A-C. Paragraph [0036]-XING discloses convolutional neural networks or generative adversarial networks with various architectures may be used (wherein one or more convolutional neural networks may be used or a generative adversarial network, which contains a neural network for the generator and the discriminator). Please also read paragraph [0037-0040]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of WANG of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of XING of having wherein the plurality of images are generated by performing deep learning processing on the image on the basis of a second deep learning network.
Wherein HILBERT’s method having wherein the plurality of quantitative images are generated by performing deep learning processing on the fused image on the basis of a second deep learning network.
The motivation behind the modification would have been to obtain a method that improves the generation of synthetic quantitative MRI images and the performance of neural networks, since both HILBERT and XING concern processing quantitative MRI images and mappings. Wherein HILBERT systems and methods improves the ability for radiologists to form diagnoses and improves the generation of synthetic images based on quantitative maps by using additional weightings and providing a large variety of contrasts based on short acquisition times on top of the quantitative information, while XING’s systems and methods improves the accuracy and efficiency for generating quantitative maps, requires a single image as the initial input and implements a neural network architecture that achieves a balance between computational workload and performance. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015 and 0025] and XING et al. (US 20210313046 A1), Abstract and paragraph [0030 and 0039-0040].
Regarding claim 6, HILBERT in view of XING and in further view of BANERJEE explicitly teach the method according to claim 1, HILBERT further explicitly teaches wherein the plurality of scan parameters comprise echo time, repetition time, and inversion recovery time (Fig. 2. Paragraph [0005]-HILBERT discloses the image contrast of the synthetic image S, usually called “synthetic contrast”, depends on sequence parameters (inversion time TI, repetition time TR, echo time TE) and tissue properties (longitudinal relaxation T1, transverse relaxation T2, and initial magnetization M0). In paragraph [0013]-HILBERT discloses the quantitative maps Q1=T1, Q2=T2 and the sequence parameters P1=TI, P2=TR, P3=TE can be used to create synthetic maps in conjunction with the known contrast mechanisms of relaxation (i.e. T1 and T2). Please also read paragraph [0030 and 0034]).
Regarding claim 7, HILBERT in view of XING and in further view of BANERJEE explicitly teaches the method according to claim 6, HILBERT further explicitly teaches wherein the performing image conversion on the plurality of quantitative maps on the basis of the plurality of scan parameters to generate a first converted image and a second converted image comprises:
generating the first converted image (Fig. 2. Paragraph [0027]-HILBERT discloses at step 101, the system uses a first quantitative MRI acquisition technique, which is preferentially a T1 mapping acquisition technique. Further in paragraph [0028]-HILBERT discloses the first quantitative MRI acquisition technique enables at the same time to generate at least a first contrast component which is for instance a first contrast-weighted image for the biological object (e.g. a first proton-density image or quantitative proton-density map) with additional weighting and optionally an initial contrast component, which is for instance an initial image (e.g. a proton-density image or map) free of the additional weighting (wherein the first contrast component image is a proton-density image M0.sub.P). Please also read paragraph [0033]) on the basis of a first formula, the first formula having the echo time and the plurality of quantitative maps as variables (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first contrast component, e.g. the proton-density image M0.sub.P with fat signal, d) the second contrast component, e.g. the proton-density image with additional magnetization transfer contrast M0.sub.M, and e) the initial contrast component, which is preferentially the proton-density image or quantitative proton-density map, e.g. M0.sub.W, as inputs in a contrast synthetization module which contains a physical signal model given by Equation (2), wherein the inputs are used to generate a synthetic image M with arbitrary TE, TR and TI. Optionally, the system is configured for displaying on a display at least two different contrasts at the same time for the biological object. The fat signal or MT-weighting could be turned on and off by switching between M0.sub.P, M0.sub.W, and M0.sub.M when using the following equation for the physical signal model used to generate the synthetic image M by means of the system according to the invention: (wherein TR, TI and TE in both the equation below and equation (2) represent repetition time, inversion time and echo time, respectively):
PNG
media_image1.png
166
1108
media_image1.png
Greyscale
Please also read paragraph [0004-0005 and 0033]); and
generating the second converted image (Fig. 2. Paragraph [0029]-HILBERT discloses at step 102, the system uses a second quantitative MRI acquisition technique, preferentially a T2 mapping acquisition technique, configured for measuring a value for a second quantitative parameter, e.g. T2, for the biological object. The acquisition of the second quantitative map, e.g. T2 map, is also used for acquiring a second contrast component which is a second contrast-weighed image or quantitative map with an additional weighting different from the additional weighting of the first contrast-weighted image or map (wherein the second contrast component image is a proton-density image M0.sub.M). Please also read paragraph [0033])) on the basis of a second formula, the second formula having the echo time, the repetition time, the inversion recovery time, and the plurality of quantitative maps as variables (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first contrast component, e.g. the proton-density image M0.sub.P with fat signal, d) the second contrast component, e.g. the proton-density image with additional magnetization transfer contrast M0.sub.M, and e) the initial contrast component, which is preferentially the proton-density image or quantitative proton-density map, e.g. M0.sub.W, as inputs in a contrast synthetization module which contains a physical signal model given by Equation (2), wherein the inputs are used to generate a synthetic image M with arbitrary TE, TR and TI. Optionally, the system is configured for displaying on a display at least two different contrasts at the same time for the biological object. The fat signal or MT-weighting could be turned on and off by switching between M0.sub.P, M0.sub.W, and M0.sub.M when using the following equation for the physical signal model used to generate the synthetic image M by means of the system according to the invention: (wherein TR, TI and TE in both the equation below and equation (2) represent repetition time, inversion time and echo time, respectively):
PNG
media_image2.png
135
1169
media_image2.png
Greyscale
Please also read paragraph [0004-0005 and 0033]).
Regarding claim 9, HILBERT in view of XING and in further view of BANERJEE explicitly teach the method according to claim 1, HILBERT further teaches wherein the single raw image is obtained by executing a synthesized magnetic resonance scan sequence (Fig. 2. Paragraph [0034]-HILBERT discloses the contrast switch is configured for enabling a switch between a first synthetic image generated by using the contrast component C.sub.i, with i≥1, and a second synthetic image generated by using the contrast component C.sub.0, in order to switch on/off the corresponding contrast. The user interface 205 is further configured for enabling a user to choose the desired synthetic sequence parameters TE, TR, TI. A user may choose to display any of the weighted contrast on a map of the biological object shown then on the display 204 of the system 200, like T2 weighted image, T2 weighted image WE, T1 weighted image, T1 weighted image WE, PD image, PD WE image, or STIR image. The invention enables a user to switch between different types of preparation contrast in the example by turning the fat signal and a MT-weighting in synthetic contrasts “on” or “off”).
Regarding claim 11, HILBERT explicitly teaches a magnetic resonance imaging system (Fig. 2, #200 called a system. Paragraph [0032]-HILBERT discloses a system 200 for generating synthetic images with switchable image contrasts for a biological object, like a brain), comprising:
an image processing module (Fig. 2, called #203 called a processing unit. Paragraph [0032]-HILBERT discloses a processing unit 203 configured for processing the data required for generating the synthetic image, the processing unit 203 being connected to the device 201 for acquiring imaging data and to the database 202; d) a display 204 for displaying the synthetic image, the display 204 being connected to the processing unit 203), comprising:
a conversion processor (Fig. 2, called #203 called a processing unit. Paragraph [0032]), configured to perform image conversion on the plurality of quantitative maps on the basis of the plurality of scan parameters to generate a first converted image and a second converted image (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first quantitative proton-density map, and d) the second quantitative proton-density map, as inputs in a contrast synthetization module which contains a physical signal model given by Eq. (2), wherein the inputs are used to generate a synthetic image M with arbitrary TE, TR and TI. The system contains a user interface with a contrast switch enabling to automatically switch between the first contrast component, the second contrast component and the initial contrast component (i.e. no contrast—initial image) when displaying, at step 104, a synthetic image of the biological object. The system is configured for displaying on a display at least two different contrasts at the same time for the biological object. The fat signal or MT-weighting could be turned on and off by switching between M0.sub.P, M0.sub.W, and M0.sub.M when using equation (4) for the physical signal model used to generate the synthetic image M);
an image fusion processor (Fig. 2, called #203 called a processing unit. Paragraph [0032]), configured to generate a fused image of the first converted image and the second converted image (Fig. 1. Paragraph [0031]-HILBERT discloses at least 3 synthetic images M might be displayed by the system, either at the same time, or by switching from one of the synthetic images to the other one by selecting the appropriate initial magnetization M0.sub.P, M0.sub.W, or M0.sub.M via the contrast switch. Further in paragraph [0034]-HILBERT discloses the obtained maps and images are used as input in a contrast synthetization module 24 of the processing unit 203 (wherein the contrast synthetization module 24 contains a physical signal model (contrast mechanism) as shown in Eq. 4 configured for generating a synthetic image M of the biological object from said inputs). The contrast switch is preferentially configured for enabling a switch between a first synthetic image generated by using the contrast component C.sub.i, with i≥1, and a second synthetic image generated by using the contrast component C.sub.0, in order to switch on/off the corresponding contrast);
a second processor (Fig. 2, called #203 called a processing unit. Paragraph [0032]), configured to generate a quantitative weighted image on the basis of the fused image (Fig. 1. Paragraph [0030]-HILBERT discloses the system is configured for displaying on a display at least two different contrasts at the same time for the biological object. Further in paragraph [0034]-HILBERT discloses the user interface 205 is further configured for enabling a user to choose the desired synthetic sequence parameters TE, TR, TI. By means of the user interface 25 and its contrast switch, a user may choose to display any of the weighted contrast on a map of the biological object shown then on the display 204 of the system 200, like T2 weighted image, T2 weighted image WE, T1 weighted image, T1 weighted image WE, PD image, PD WE image, or STIR image. A user may switch between different types of preparation contrast by turning the fat signal and a MT-weighting in synthetic contrasts “on” or “off”. Therefore, it would have been obvious to a person of ordinary skill in the art to use a second processor. Hilbert explicitly teaches a processing unit to perform the same functions of image processing, conversion, fusion and generation. Thus, it would be obvious to use one or more additional dedicated processors for these tasks given this would increase speed and computing power).
HILBERT fails to explicitly teach and a first processor, configured to simultaneously generate a plurality of quantitative maps on the basis of the single raw image;
However, XING explicitly teaches and a first processor (Fig. 1A-B and 4A-C. Paragraph [0027]-XING discloses the deep learning network derives quantitative relaxation parametric maps from a single qualitative MR image, which gives flexibility to input qualitative images. The network may be implemented in the MRI scanner or on an external computer Nvidia GPU GeForce GTX1070.), configured to simultaneously generate a plurality of quantitative maps on the basis of the single raw image (Fig. 1A-B and 4A-C. Paragraph [0025]-XING discloses for T.sub.1 mapping, the method uses a deep neural network 108 to derive quantitative T.sub.1 and proton density maps from a single conventional T.sub.1 weighted image 106 acquired in routine clinical practice, as illustrated in FIG. 1B. With the use of the deep neural network, only one T.sub.1 weighted image 106 is required for the generation of a quantitative T.sub.1 map. Further in paragraph [0026]-XING discloses a T.sub.2 map can be produced from a single T.sub.2 or T.sub.2/T.sub.1 weighted image using a trained deep neural network. In this way, qualitative and quantitative MR images can be obtained in the routine clinical practice without changing the imaging protocol or performing multiple scans. In paragraph [0036]-XING discloses generative adversarial networks with various architectures may be used. Therefore, it would have been obvious to a person of ordinary skill in the art to use a single raw image given all Magnetic Resonance Imaging scanners acquire a raw image (i.e. K-space, which is a raw digital matrix that stores signal data) and all modern MRI scanners have both i-channel and q-channels (i.e. real and imaginary components). This would improve the ability to accurately process and analyze the data).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT of having a magnetic resonance imaging system, comprising: a scanner, configured to execute a magnetic resonance scan sequence to generate a raw image, the magnetic resonance scan sequence having a plurality of scan parameters; a conversion processor, configured to perform image conversion on the plurality of quantitative maps on the basis of the plurality of scan parameters to generate a first converted image and a second converted image; an image fusion processor, configured to generate a fused image of the first converted image and the second converted image; and a second processor, configured to generate a plurality of quantitative weighted images on the basis of the fused image, with the teachings of XING of having and a first processor, configured to simultaneously generate a plurality of quantitative maps on the basis of the single raw image.
Wherein HILBERT’s system having and a first processor, configured to simultaneously generate a plurality of quantitative maps on the basis of the single raw image.
The motivation behind the modification would have been to obtain a system that improves the generation of synthetic quantitative MRI images and the performance of neural networks, since both HILBERT and XING concern processing quantitative MRI images and mappings. Wherein HILBERT systems and methods improves the ability for radiologists to form diagnoses and improves the generation of synthetic images based on quantitative maps by using additional weightings and providing a large variety of contrasts based on short acquisition times on top of the quantitative information, while XING’s systems and methods improves the accuracy and efficiency for generating quantitative maps, requires a single image as the initial input and implements a neural network architecture that achieves a balance between computational workload and performance. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015 and 0025] and XING et al. (US 20210313046 A1), Abstract and paragraph [0030 and 0039-0040].
HILBERT in view of XING is silent on a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image.
However, BANERJEE explicitly teaches a scanner (Fig. 1, #100 called an MRI system. Paragraph [0025]. In paragraph [0022]-BANERJEE discloses the present disclosure is related to synthetic MRI. Synthetic MRI can reconstruct multiple image contrasts (e.g., T1- and T2-weighted, T1- and T2-FLAIR, proton density-weighted, STIR images) from MR signals acquired with a quantification sequence (e.g., MDME sequences) in a single scan), configured to execute a magnetic resonance scan sequence to generate a single raw image (Fig. 1. Paragraph [0029]-BANERJEE discloses an object or patient 170 undergoing a MRI scan may be positioned within the open cylindrical imaging volume 146 of the resonance assembly 140. In paragraph [0031]-BANERJEE discloses an MR scan is complete when an array of raw k-space data has been acquired and stored. This raw k-space data is rearranged into separate k-space data arrays for each image to be reconstructed), the magnetic resonance scan sequence having a plurality of scan parameters (Fig. 1. Paragraph [0035]-BANERJEE discloses referring to FIG. 3, a schematic diagram of a single block of a pulse sequence 300 for performing a quantification scan is shown. A two-dimensional (2D) fast spin echo (FSE) multi-delay multi-echo (MDME) sequence may be performed for the quantification scan, which includes an interleaved slice selective saturation RF pulse and multi-echo acquisition. The saturation acts on a slice n, whereas the acquisition acts on a slice m, using the 90° excitation pulse and multiple 180° pulses. n and m are different slices. Four (4) different choices of n and m are performed, resulting in four different delay times. The number of echoes may be set as two (2), at two different echo times. The result of the quantitative acquisition 210 is eight (8) complex images per slice. It should be understood that the example of 4 delays at 2 echoes is described herein for illustration, not for limitation. Any appropriate combination of delays and echoes can be used to perform the quantification scan. The quantitative acquisition 210 may be raw MRI signals in k-space, such as raw MDME MRI), wherein the single raw image comprises at least one of a real image or an imaginary image (Fig. 1. Paragraph [0036]-BANERJEE discloses the deep neural network 220 may be a multi-scale U-Net convolutional neural network (CNN) with an encoder-decoder structure. An input layer 221 receives the quantitative acquisition 210. The input layer 221 includes multiple channels to accommodate the input of the multiple images as the result of a quantitative acquisition at once. Each of the multiple images is split up into real and imaginary components and stacked as a corresponding channel).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING of having a magnetic resonance imaging system, comprising: a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, with the teachings of BANERJEE of having a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image.
Wherein HILBERT’s system having a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image.
The motivation behind the modification would have been to obtain a system that enhances processing speed, accuracy and image quality, since HILBERT and BANERJEE concern systems and methods for processing and generating quantitative magnetic resonance maps and images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while BANERJEE provides systems and methods that reduces image artifacts and improves reconstruction accuracy, resolution restoration, and training time. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and BANERJEE et al. (US 20210027436 A1), Abstract and Paragraph [0037, 0042, and 0046].
Regarding claim 12, HILBERT in view of XING and in further view of BANERJEE explicitly teach the system according to claim 11, HILBERT fail to explicitly teach wherein the first processor is configured to perform deep learning processing on the single raw image on the basis of a first deep learning network to generate the plurality of quantitative maps.
However, XING explicitly teaches wherein the first processor (Fig. 1B. Paragraph [0027]-XING discloses the method may be performed by a conventional MRI scanner using standard imaging protocols, adapted with a neural network to generate the quantitative MRI map(s) from the qualitative image acquired by the scanner using conventional clinical imaging techniques. The deep learning network derives quantitative relaxation parametric maps from a single qualitative MR image. The network may be implemented in the MRI scanner or on an external computer Nvidia GPU GeForce GTX1070) is configured to perform deep learning processing on the single raw image on the basis of a first deep learning network (Fig. 1B, #108 called a deep learning network. Paragraph [0025]. In paragraph [0036]-XING discloses generative adversarial networks with various architectures may be used (wherein a generative adversarial network consists of two networks). Please also see FIG. 3 and read paragraph [0037-0040]) to generate the plurality of quantitative maps (Fig. 1B. Paragraph [0025]-XING discloses for T.sub.1 mapping, the method uses a deep neural network 108 to derive quantitative T.sub.1 and proton density maps from a single conventional T.sub.1 weighted image 106 acquired in routine clinical practice, as illustrated in FIG. 1B. With the use of the deep neural network, only one T.sub.1 weighted image 106 is required for the generation of a quantitative T.sub.1 map. Further in paragraph [0026]-XING discloses a T.sub.2 map can be produced from a single T.sub.2 or T.sub.2/T.sub.1 weighted image using a trained deep neural network. Please also see Fig. 4-7).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of WANG of having a magnetic resonance imaging system, comprising: a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of XING of having wherein the first processor is configured to perform deep learning processing on the single raw image on the basis of a first deep learning network to generate the plurality of quantitative maps.
Wherein HILBERT’s system having wherein the first processor is configured to perform deep learning processing on the single raw image on the basis of a first deep learning network to generate the plurality of quantitative maps.
The motivation behind the modification would have been to obtain a system that improves the generation of synthetic quantitative MRI images and the performance of neural networks, since both HILBERT and XING concern processing quantitative MRI images and mappings. Wherein HILBERT systems and methods improves the ability for radiologists to form diagnoses and improves the generation of synthetic images based on quantitative maps by using additional weightings and providing a large variety of contrasts based on short acquisition times on top of the quantitative information, while XING’s systems and methods improves the accuracy and efficiency for generating quantitative maps, requires a single image as the initial input and implements a neural network architecture that achieves a balance between computational workload and performance. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015 and 0025] and XING et al. (US 20210313046 A1), Abstract and paragraph [0030 and 0039-0040].
Regarding claim 13, HILBERT in view of XING and in further view of BANERJEE explicitly teach the system according to claim 11, although HILBERT explicitly teaches wherein the second processor (Fig. 2, called #203 called a processing unit. Paragraph [0032]-HILBERT discloses a processing unit 203 configured for processing the data required for generating the synthetic image. Therefore, it would have been obvious to a person of ordinary skill in the art to use a second processor. Hilbert explicitly teaches a processing unit to perform the same functions of image processing, conversion, fusion and generation. Thus, it would be obvious to use one or more additional dedicated processors for these tasks given this would increase speed and computing power) performs processing on the fused image to generate a plurality of quantitative weighted images (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first contrast component, which is preferentially the first proton-density image or quantitative proton-density map, e.g. the proton-density image M0.sub.P with fat signal, d) the second contrast component, which is preferentially the second proton-density image or quantitative proton-density map, e.g. the proton-density image with additional magnetization transfer contrast M0.sub.M, and e) the initial contrast component, which is preferentially the proton-density image or quantitative proton-density map. Please also read paragraph [0027-0029 and 0034]).
HILBERT fails to explicitly teach wherein the second processor performs deep learning processing on the fused image on the basis of a second deep learning network to generate a plurality of quantitative weighted images.
However, XING explicitly teaches wherein the processor (Fig. 1A-B and 4A-C. Paragraph [0027]-XING discloses the network may be implemented in the MRI scanner or on an external computer Nvidia GPU GeForce GTX1070) performs deep learning processing on the image on the basis of a second deep learning network (Fig. 1A-B and 4A-C. Paragraph [0036]-XING discloses convolutional neural networks or generative adversarial networks with various architectures may be used (wherein one or more convolutional neural networks may be used or a generative adversarial network, which contains a neural network for the generator and the discriminator). Please also read paragraph [0037-0040]) to generate a plurality of quantitative weighted images (Fig. 1A-B and 4A-C. Paragraph [0024]-XING discloses the inventors have discovered that deep learning enables the acquisition and utilization of generic a priori information to predict quantitative MRI data from a single qualitative MR image. In paragraph [0029]-XING discloses after a qualitative image is acquired 800 it is then applied 802 as input to the neural network to obtain the quantitative image. In paragraph [0025]-XING discloses for T.sub.1 mapping, the method uses a deep neural network 108 to derive quantitative T.sub.1 and proton density maps no from a single conventional T.sub.1 weighted image 106. In paragraph [0026]-XING discloses a T.sub.2 map can be produced from a single T.sub.2/T.sub.1 weighted image using a trained deep neural network. Please also read paragraph [0025, 0028 and 0032]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a magnetic resonance imaging system, comprising: a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of XING of having wherein the processor performs deep learning processing on the image on the basis of a second deep learning network to generate a plurality of images.
Wherein HILBERT’s system having wherein the second processor performs deep learning processing on the fused image on the basis of a second deep learning network to generate a plurality of quantitative weighted images.
The motivation behind the modification would have been to obtain a system that improves the generation of synthetic quantitative MRI images and the performance of neural networks, since both HILBERT and XING concern processing quantitative MRI images and mappings. Wherein HILBERT systems and methods improves the ability for radiologists to form diagnoses and improves the generation of synthetic images based on quantitative maps by using additional weightings and providing a large variety of contrasts based on short acquisition times on top of the quantitative information, while XING’s systems and methods improves the accuracy and efficiency for generating quantitative maps, requires a single image as the initial input and implements a neural network architecture that achieves a balance between computational workload and performance. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015 and 0025] and XING et al. (US 20210313046 A1), Abstract and paragraph [0030 and 0039-0040].
Regarding claim 15, HILBERT in view of XING and in further view of BANERJEE explicitly teach the system according to claim 11, HILBERT further teaches wherein the single raw image is obtained by executing a synthesized magnetic resonance scan sequence (Fig. 2. Paragraph [0032]-HILBERT discloses a processing unit 203 configured for processing the data required for generating the synthetic image, the processing unit 203 being connected to the device 201 for acquiring imaging data and to the database 20. The system 200 according to the invention is configured for performing the steps of the previously described method for generating the synthetic image with switchable image contrasts. Please also read paragraph [0030 and 0033]).
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over HILBERT et al. (US 20200333414 A1), hereinafter referenced as HILBERT in view of XING et al. (US 20210313046 A1), hereinafter referenced as XING and further in view of BANERJEE et al. (US 20210027436 A1), hereinafter referenced as BANERJEE and further in view of SHI et al. (US 20220207791 A1), hereinafter referenced as SHI.
Regarding claim 4, HILBERT in view of XING and in further view of BANERJEE explicitly teach the method according to claim 1, although HILBERT explicitly teaches wherein the fused image is generated by the first converted image and the second converted image as a preprocessing step (Fig. 1. Paragraph [0027]-HILBERT discloses at step 101, the system uses a first quantitative MRI acquisition technique, which is preferentially a T1 mapping acquisition techniqu [0031]-HILBERT discloses at least 3 synthetic images M might be displayed by the system, either at the same time, or by switching from one of the synthetic images to the other one by selecting the appropriate initial magnetization M0.sub.P, M0.sub.W, or M0.sub.M via the contrast switch. Further in paragraph [0034]-HILBERT discloses the obtained maps and images are used as input in a contrast synthetization module 24 of the processing unit 203 (wherein the contrast synthetization module 24 contains a physical signal model (contrast mechanism) as shown in Eq. 4 configured for generating a synthetic image M of the biological object from said inputs). The contrast switch is preferentially configured for enabling a switch between a first synthetic image generated by using the contrast component C.sub.i, with i≥1, and a second synthetic image generated by using the contrast component C.sub.0, in order to switch on/off the corresponding contrast. Please also read paragraph [0003-0008, 0025 and 0028-0033]).
HILBERT in view of XING fails to explicitly teach wherein the fused image is generated by performing channel concatenation on the first converted image and the second converted image as a preprocessing step prior to input into the second deep learning network.
However, SHI explicitly teaches wherein the fused image is generated by performing channel concatenation on the first image (Fig. 1, #12a’ called a Primary SPECT patch. Paragraph [0040]) and the second image (Fig. 1, #12b’ called a Scatter SPECT patch. Paragraph [0040]) as a preprocessing step prior to input into the second deep learning network (Fig. 1. Paragraph [0040]-SHI discloses a system 100 is disclosed that employs a machine learning system based upon artificial neural networks to estimate attenuation maps for SPECT emission data, wherein the machine learning system includes a generator network 10 and a discriminator network 16. The artificial neural network is in the form of a deep convolutional neural network (CNN) the artificial neural network is in the form of a deep convolutional neural network (CNN) and training of the deep CNN is described. Images reconstructed from photopeak window (126 keV-155 keV) 12a (that is, the primary window) and scatter window (114 keV-126 keV) 12b are concatenated as a multi-channel image and fed into a generator network 10. Please also read paragraph [0042]).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of SHI of having wherein the fused image is generated by performing channel concatenation on the first image and the second image as a preprocessing step prior to input into the second deep learning network.
Wherein HILBERT’s method having with the teachings of SHI of having wherein the fused image is generated by performing channel concatenation on the first converted image and the second converted image as a preprocessing step prior to input into the second deep learning network.
The motivation behind the modification would have been to obtain a method that enhances processing speed and image quality, since both HILBERT and SHI concern systems and methods for processing and generating medical images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while SHI provides systems and methods that allow for the production of realistic attenuation maps with speed and high accuracy. Furthermore, as SHI states in paragraph [0007], deep learning-based approaches have been proposed to estimate images of one modality from another. For example, “initial success was obtained for the task of generating attenuation maps for nuclear images. In “MR-based synthetic CT generation using a deep convolutional neural network method,” convolutional neural networks were used to convert magnetic resonance imaging (MRI) images to attenuation CT images for PET/MRI systems. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and SHI et al. (US 20220207791 A1), Abstract and Paragraph [0003-0007 and 0097-0102].
Claim 5 is rejected under 35 U.S.C. 103 as being unpatentable over HILBERT et al. (US 20200333414 A1), hereinafter referenced as HILBERT in view of XING et al. (US 20210313046 A1), hereinafter referenced as XING and further in view of BANERJEE et al. (US 20210027436 A1), hereinafter referenced as BANERJEE and further in view of WANG et al. (US 20110044524 A1), hereinafter referenced as WANG (2011).
Regarding claim 5, HILBERT in view of XING and in further view of BANERJEE explicitly teach the method according to claim 1, although HILLBERT explicitly teaches wherein the plurality of quantitative maps comprise a quantitative T1 map, a quantitative T2 map, and a quantitative PD map (Fig. 2. Paragraph [0030]-HILBERT discloses at step 103, the system is configured for using: a) the first quantitative map, e.g. the T1 map, b) the second quantitative map, e.g. the T2 map, c) the first contrast component, which is preferentially the first proton-density image or quantitative proton-density map, e.g. the proton-density image M0.sub.P with fat signal, d) the second contrast component, which is preferentially the second proton-density image or quantitative proton-density map, e.g. the proton-density image with additional magnetization transfer contrast M0.sub.M, and e) the initial contrast component, which is preferentially the proton-density image or quantitative proton-density map), and the plurality of quantitative weighted images comprise a T1 weighted image, a T2 weighted image (Fig. 2. Paragraph [0028]-HILLBERT discloses the user interface 205 is further configured for enabling a user to choose the desired synthetic sequence parameters TE, TR, TI. A user may choose to display any of the weighted contrast on a map of the biological object shown then on the display 204 of the system 200, like T2 weighted image, T2 weighted image WE, T1 weighted image, T1 weighted image WE, PD image, PD WE image, or STIR image. A user may switch between different types of preparation contrast in the example by turning the fat signal and a MT-weighting in synthetic contrasts “on” or “off”. In paragraph [0036]-HILBERT discloses other quantitative parameters than T1 and/or T2 may be acquired (e.g. T2*, multi compartment T2/T1, MT)).
HILLBERT in view of XING fail to explicitly teach a T2 weighted-fluid attenuated inversion recovery image.
However, WANG (2011) explicitly teaches a T2 weighted-fluid attenuated inversion recovery image (Fig. 1. Paragraph [0665]-WANG discloses all subjects will be imaged on a 3T MR scanner. Standard T2-weighted fluid attenuated inversion recovery imaging will also be included. The T2* multiple echo imaging data will generate iron maps, standard T2* magnitude images and their phase masked SWI images for analysis).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of WANG (2011) of having a T2 weighted-fluid attenuated inversion recovery image.
Wherein HILBERT’s method wherein the plurality of quantitative maps comprise a quantitative T1 map, a quantitative T2 map, and a quantitative PD map, and the plurality of quantitative weighted images comprise a T1 weighted image, a T2 weighted image, and a T2 weighted-fluid attenuated inversion recovery image.
The motivation behind the modification would have been to obtain a method for generating a magnetic resonance that enhances processing speed and image quality, since both HILBERT and WANG (2011) concern systems and methods for processing and generating quantitative magnetic resonance maps and images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while WANG (2011)provides systems and methods that improve the performance and precision of measurements in magnetic resonance imaging. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and WANG et al. (US 20110044524 A1), Abstract and Paragraph [0493-0494, 0617, 0834, 0926].
Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over HILBERT et al. (US 20200333414 A1), hereinafter referenced as HILBERT in view of XING et al. (US 20210313046 A1), hereinafter referenced as XING and further in view of BANERJEE et al. (US 20210027436 A1), hereinafter referenced as BANERJEE and further in view of AKCAKAYA et al. (US 20210090306 A1), hereinafter referenced as AKCAKAYA.
Regarding claim 14, HILBERT in view of XING and in further view of BANERJEE explicitly teach the system according to claim 11, HILBERT fails to explicitly teach wherein the image fusion processor is configured to perform channel concatenation on the first converted image and the second converted image to generate the fused image.
However, AKCAKAYA explicitly teaches wherein the image fusion processor is configured to perform channel concatenation on the first converted image and the second converted image to generate the fused image (Fig. 1. Paragraph [0026]-AKCAKAYA discloses a complex-valued k-space dataset, s, of size n.sub.x×n.sub.y×n.sub.c, can be embedded into a real-valued space as a dataset of size n.sub.x×n.sub.y×2n.sub.c, where the real part of s is concatenated with the imaginary part of s along the third (channel) dimension).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a magnetic resonance imaging system, comprising: a scanner, configured to execute a magnetic resonance scan sequence to generate a single raw image, the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of AKCAKAYA of having wherein the image fusion processor is configured to perform channel concatenation on the first converted image and the second converted image to generate the fused image.
Wherein HILBERT’s system having wherein the image fusion processor is configured to perform channel concatenation on the first converted image and the second converted image to generate the fused image.
The motivation behind the modification would have been to obtain a system that enhances processing speed and image quality, since HILBERT and XING concern systems and methods for processing and generating quantitative magnetic resonance maps and images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while AKCAKAYA provides systems and methods that improved performance of MRI systems and the noise performance of reconstructions. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and AKCAKAYA et al. (US 20210090306 A1), Abstract and Paragraph [0013, 0022 and 0060-0062].
Claims 17-18 are rejected under 35 U.S.C. 103 as being unpatentable over HILBERT et al. (US 20200333414 A1), hereinafter referenced as HILBERT in view of XING et al. (US 20210313046 A1), hereinafter referenced as XING and further in view of BANERJEE et al. (US 20210027436 A1), hereinafter referenced as BANERJEE and further in view of WANG et al. (US 20190162805 A1), hereinafter referenced as WANG.
Regarding claim 17, HILBERT in view of XING and in further view of BANERJEE explicitly teach method according to claim 1, HILBERT in view of XING is silent on wherein the single raw image comprises at least one of the real image.
However, WANG explicitly teaches wherein the single raw image (Fig. 1. Paragraph [0057]-WANG discloses FIG. 1 depicts an MRI system 10. In paragraph [0060]-WANG discloses the computing device 100 may receive MR signals from the receiving coil of the transmission and receiving unit 16 and reconstruct an MRI image based on the received MR signals. In paragraph [0061]-WANG discloses the computing device 100 may convert the MR signals received from the transmitting and receiving unit 16 into k-space data. The computing device 100 may generate MR image data from the k-space data with image reconstruction processing (wherein sequences may include gradient echo, spin echo, fast gradient echo, fast spin echo, and their variations with or without magnetization preparation and/or specific tissue suppression, parallel imaging technique, under-sampling technique, and/or the administration of contrast agent and parameters may include T.sub.1 relaxation, T.sub.2 relaxation, T.sub.2 star relaxation, but not limited to, proton density, diffusion, magnetic susceptibility, or magnetization transfer). Please also read paragraph [0004, 0079, 0090, 0094-0098, and 0100]) comprises at least one of the real image (Fig. 1. Paragraph [0099]-WANG discloses the acquired image may be at least one of a magnitude image, a phase image, a real image, an imaginary image, or a complex image. In paragraph [0080]-WANG discloses the images may be at least one of magnitude images, phase images, real images, imaginary images, complex images, including combinations thereof)).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of WANG of having wherein the single raw image comprises at least one of the real image.
Wherein HILBERT’s method having wherein the single raw image comprises at least one of the real image.
The motivation behind the modification would have been to obtain a method that enhances processing speed and image quality, since HILBERT and WANG concern systems and methods for processing and generating quantitative magnetic resonance maps and images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while WANG provides systems and methods that increase MRI value by improving image quality, increasing MRI efficiency, reducing scan time, suppressing the non-targeted signal intensity, increasing accuracy of diagnoses, and reducing image artifacts. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and WANG et al. (US 20190162805 A1), Abstract and Paragraph [0091-0092 and 0099].
Regarding claim 18, HILBERT in view of XING and in further view of BANERJEE explicitly teach the method according to claim 1, HILBERT in view of XING is silent on wherein the single raw image comprises at least one of the imaginary image.
However, WANG explicitly teaches wherein the single raw image (Fig. 1. Paragraph [0057]-WANG discloses FIG. 1 depicts an MRI system 10. In paragraph [0060]-WANG discloses the computing device 100 may receive MR signals from the receiving coil of the transmission and receiving unit 16 and reconstruct an MRI image based on the received MR signals. In paragraph [0061]-WANG discloses the computing device 100 may convert the MR signals received from the transmitting and receiving unit 16 into k-space data. The computing device 100 may generate MR image data from the k-space data with image reconstruction processing (wherein sequences may include gradient echo, spin echo, fast gradient echo, fast spin echo, and their variations with or without magnetization preparation and/or specific tissue suppression, parallel imaging technique, under-sampling technique, and/or the administration of contrast agent and parameters may include T.sub.1 relaxation, T.sub.2 relaxation, T.sub.2 star relaxation, but not limited to, proton density, diffusion, magnetic susceptibility, or magnetization transfer). Please also read paragraph [0004, 0079, 0090, 0094-0098, and 0100]) comprises at least one of the imaginary image (Fig. 1. Paragraph [0099]-WANG discloses the acquired image may be at least one of a magnitude image, a phase image, a real image, an imaginary image, or a complex image. In paragraph [0080]-WANG discloses the images may be at least one of magnitude images, phase images, real images, imaginary images, complex images, including combinations thereof)).
Therefore, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention was made to combine the teachings of HILBERT in view of XING and in further view of BANERJEE of having a method for generating a magnetic resonance image, comprising: simultaneously generating a plurality of quantitative maps on the basis of a single raw image, the single raw image being obtained by executing a magnetic resonance scan sequence, and the magnetic resonance scan sequence having a plurality of scan parameters, wherein the single raw image comprises at least one of a real image or an imaginary image, with the teachings of WANG of having wherein the single raw image comprises at least one of the imaginary image.
Wherein HILBERT’s method having wherein the single raw image comprises at least one of the imaginary image.
The motivation behind the modification would have been to obtain a method that enhances processing speed and image quality, since HILBERT and WANG concern systems and methods for processing and generating quantitative magnetic resonance maps and images. Wherein HILBERT systems and methods provides a large variety of contrasts based on short acquisition times on top of the quantitative information and improves the ability for radiologists to form diagnoses, while WANG provides systems and methods that increase MRI value by improving image quality, increasing MRI efficiency, reducing scan time, suppressing the non-targeted signal intensity, increasing accuracy of diagnoses, and reducing image artifacts. Please see HILBERT et al. (US 20200333414 A1), Abstract and Paragraph [0015] and WANG et al. (US 20190162805 A1), Abstract and Paragraph [0091-0092 and 0099].
Conclusion
Listed below are the prior arts made of record and not relied upon but are considered pertinent to applicant`s disclosure.
Koch et al. (US 20210033688 A1)- Systems and methods for quantitative susceptibility mapping (“QSM”) using magnetic resonance imaging (“MRI”) are described. Localized magnetic field information is used when performing the inversion to compute quantitative susceptibility maps. The localized magnetic field information can include multi-resolution subvolumes obtained by segmenting, or dividing, a field shift map. In some instances, a trained machine learning algorithm, such as a trained neural network, can be implemented to convert the localized magnetic field information into quantitative susceptibility data. These local susceptibility maps can be combined to form a composite quantitative susceptibility map of the imaging volume..............................Please see Fig. 1-4. Abstract.
COHEN et al. (US 20180203081 A1)- Disclosed is a system and method for estimating quantitative parameters of a subject using a magnetic resonance (“MR”) system using a dictionary. The dictionary may include a plurality of signal templates that sparsely sample acquisition parameters used when acquiring data. The acquired data is compared with the dictionary using a neural network. Thus, systems and methods are provided that are more computationally efficient, and have reduced data storage requirements than traditional MRF reconstruction systems and methods............................Please see Fig. 1 and Para. [0086-0087]. Abstract.
CHEN et al. (US 20220308147 A1)- Systems and methods providing enhancements to quantitative imaging systems and techniques are described herein. In one aspect, a system for tissue quantification in magnetic resonance fingerprinting (MRF) comprises a feature extraction module operable to convert pixel input high-dimensional signal evolution in to a low-dimensional feature map. The system also comprises a spatially constrained quantification module operable to capture spatial information from the low-dimensional feature map and generate an estimated tissue property map..........................Please see Fig. 1. Abstract.
HILBERT et al. (US 20190371465 A1)-A system and a method determine a value for a parameter. Reference values for the parameter are determined from a group of objects. A first technique is used by the system for determining for each object the reference value from a first set of data. A learning dataset is created by associating for each object of the group of objects a second set of data and the reference value.......................Please see Fig. 1-2. Abstract.
Hilbert et al. (US 20180286088 A1)- The disclosure includes a method for generating quantitative magnetic resonance (MR) images of an object under investigation. A first MR data set of the object under investigation is captured in an undersampled raw data space, wherein the object under investigation is captured in a plurality of 2D slices, in which the resolution in a slice plane of the slices is in each case higher than perpendicular to the slice plane, wherein the plurality of 2D slices are in each case shifted relative to one another by a distance which is smaller than the resolution perpendicular to the slice plane. Further MR raw data points of the first MR data set are reconstructed with the assistance of a model using a cost function which is minimized. The cost function takes account of the shift of the plurality of 2D slices perpendicular to the slice plane........................Please see Fig. 1-2. Abstract.
JARA et al. (US 20190365273 A1)- Methods of making a white matter fibrogram representing the connectome of the brain of a subject, comprising: (a) performing a multispectral multislice magnetic resonance scan on the brain of a subject, (b) storing image data indicative of a plurality of magnetic resonance weightings of each of a plurality of slices of the brain of the subject to provide directly acquired images, (c) processing the directly acquired images to generate a plurality of quantitative maps of the brain indicative of a plurality of qMRI parameters of the subject, (d) constructing a plurality of magnetic resonance images indicative of white matter structure from the quantitative maps, and (e) rendering a white matter fibrogram of the brain of the subject from the plurality of magnetic resonance images.......................Please see Fig. 4-7. Abstract.
SHIH et al. (US 20240230810 A1)- A method for generating magnetic resonance imaging (MRI) quantitative parameter maps includes receiving at least one multi-contrast magnetic resonance (MR) image of a subject, providing the image to an artifact suppression deep learning network of a two-stage deep learning network and generating at least one multi-contrast MR image with suppressed undersampling artifacts using the artifact suppression deep learning network. The method further includes providing the at least one multi-contrast MR image with suppressed undersampling artifacts to a parameter mapping deep learning network of the two-stage deep learning network, generating at least one quantitative MR parameter map and generating an uncertainty estimation map for the at least one quantitative MR parameter map using the parameter mapping deep learning network. The method further includes displaying at least one multicontrast MR image with suppressed undersampling artifacts, at least one quantitative MR parameter map, and the corresponding uncertainty estimation map on a display..............................Please see Fig. 2-5 and Para. [0044-0045]. Abstract.
Zhang et al. (US 20210042883 A1)- A computer-implemented method is provided for improving image quality with shortened acquisition time. The method comprises: determining an accelerated image acquisition scheme for imaging a subject using a medical imaging apparatus; acquiring a medical image of the subject according to the accelerated image acquisition scheme using the medical imaging apparatus; applying a deep network model to the medical image to improve the quality of the medical image; and outputting an improved quality image of the subject, for analysis by a physician...........................Please see Fig. 4-5 and Para. [0086-0087]. Abstract.
Any inquiry concerning this communication or earlier communications from the examiner
should be directed to Aaron Bonansinga whose telephone number is (703) 756-5380 The examiner can normally be reached on Monday-Friday, 9:00 a.m. - 6:00 p.m. ET.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s
supervisor, Chineyere Wills-Burns can be reached by phone at (571) 272-9752. The fax phone number for the organization where this application or proceeding is assigned is (571) 273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/AARON TIMOTHY BONANSINGA/Examiner, Art Unit 2673
/CHINEYERE WILLS-BURNS/Supervisory Patent Examiner, Art Unit 2673
Prosecution Insights