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 .
Comments
The Amendment – After Non-Final Rejection filed on February 27, 2026 has been entered and made of record.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim 23 is rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Claim 23 recites the limitation “the at least one pre-training image includes a plurality of pre-training images obtained from a plurality of perspectives of a region of interest” in lines 1-3. However, the originally filed specification does not provide proper support for the claimed subject matter.
Appropriate correction is required.
Claim Rejections - 35 USC § 103
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, 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 1-4, 9, 11, 12, 14, 16-20, and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Lenga et al. (U.S. Pub. No. 2024/0050054) in view of Gordon et al. (U.S. Pub. No. 2022/0405915).
As to claim 1, Lenga et al. teaches a computer-implemented method (i.e., “method for providing a prediction of a representation of an examination region that was generated using a medical image technique involving a contrast agent”, Abstract), comprising:
generating a pre-trained machine-learning model (i.e., “radiological machine learning system 102 may analyze the training data to generate the predictive machine learning model”, Paragraph [0145]) by unsupervised pre-training of a machine-learning model (i.e., “The machine learning techniques may include, for example, supervised and/or unsupervised techniques”, Paragraph [0145]) for predicting time-related information from at least one pre-training image (i.e., “prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]), the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period (i.e., “at three different time points t1, t2, and t3, radiological machine learning system 102 may generate three representations of an examination region of an examination object”, Paragraph [0153]), and the time-related information being associated with one or more points of time during the observation period (See for example, Paragraph [0153]; “The third radiological image, O3, may include an image of the examination region that was generated with a third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
However, Lenga et al. does not explicitly disclose building a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model being for predicting semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Gordon et al. teaches building a further machine-learning model (i.e., “ML model”, Paragraph [0104]) using at least part of a pre-trained machine-learning model (See for example, “The ML model is then trained by further training the baseline ML model, using the transfer learning approach”, Paragraph [0104]; see in addition “The ML model is trained on a training dataset of records, each record including a respective contrast enhanced medical image(s) and optional additional parameter(s), labelled with a ground truth label of calcification parameter(s), for example, as described with reference to FIG. 2”, Paragraph [0108]; see also, “At 206, images with a specific contrast phase may be selected … respective contrast enhanced medical image(s) are analyze to determine a respective contrast phase. The contrast phase depicted in the respective contrast enhanced medical image may be determined, for example, by extracting the contrast phase from metadata associated with the scan, and/or by a trained ML model that outputs an indication of the contrast phase. Optionally, contrast enhanced medical images designed one or more specific phases (e.g., arterial, portal venous, delayed) may be selected for including in a training dataset, along with the corresponding non-contrast enhanced image”, Paragraphs [0075]-[0076]), the further machine-learning model being for predicting semantic context information (See for example, “a target calcification parameter(s) is obtained for a target calcification depicted in the target anatomical structure as an outcome of the ML model … a fine mask delineating the target calcification depicted in the target anatomical structure is obtained as an outcome of the ML model”, Paragraphs [0109]-[0110]) from at least one inference image (i.e., “a target contrast enhanced image of a target subject is accessed”, Paragraph [0105]) acquired by the contrast-enhanced medical imaging system (i.e., “Imaging device 112 and/or data repository 114 that store images acquired by imaging device 112. The acquired images may be fed into trained ML model(s) 122A for inference thereof”, Paragraph [0061]) using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Lenga et al. and Gordon et al. are analogous art because they are from the field of digital image processing for processing contrast enhanced images.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify Lenga et al. by incorporating the building of a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model being for predicting semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement, as taught by Gordon et al.
The suggestion/motivation for doing so would have been to increase the performance of a machine-learning model when identifying a calcification.
Therefore, it would have been obvious to combine Gordon et al. with Lenga et al. to obtain the invention as specified in claim 1.
As to claim 2, Lenga et al. teaches wherein said at least one pre-training image is acquired by the contrast-enhanced medical imaging system at a first point in time during the observation period (See for example, O1 at t1, Paragraphs [0153]-[0154]), and
wherein said time-related information includes information associated with an acquisition by the contrast-enhanced medical imaging system at a second point in time during the observation period, the second point in time being different than the first point in time (i.e., O2 at t2, Paragraphs [0153]-[0154]).
As to claim 3, Lenga et al. teaches wherein said unsupervised pre-training of a machine-learning model for predicting time-related information comprises:
obtaining said at least one pre-training image of the patient acquired by the contrast-enhanced medical imaging system at a first point in time during the observation period (i.e., “the first representation, F1, and the second representation, F2, may be provided as input to the predictive machine learning model”, Paragraph [0159]);
applying said machine-learning model to the at least one pre-training image, wherein said time-related information is predicted for a second point in time during the observation period (i.e., “The output of the predictive machine learning model may include a prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]);
obtaining ground-truth information based on a further image acquired by the contrast-enhanced medical imaging system at the second point in time (i.e., O3 at t3, Paragraph [0153]); and
training the machine-learning model based on comparing the ground-truth information and the time-related information predicted for the second point in time (i.e., “radiological machine learning system 102 may compare the prediction of the representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique to the third representation in the frequency space of the examination region, F3, with the third amount of the contrast agent administered during the medical imaging technique …”, Paragraph [0159]).
As to claim 4, Lenga et al. teaches wherein the first point in time corresponds to a pre-contrast phase of the observation period prior to a contrast agent being introduced into the patient (i.e., “The first radiological image, O1, may include an image of the examination region that was generated without an amount of a contrast agent administered during the medical imaging technique”, Paragraph [0154]), and
wherein the second point in time corresponds to a post-injection phase of the observation period after the contrast agent is introduced into the patient (i.e., “The second radiological image, O2, may include an image of the examination region that was generated with a second amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
As to claim 9, Gordon et al. teaches wherein said using of said at least part of the pre-trained machine-learning model comprises: incorporating said at least part of the pre-trained machine-learning model into the further machine-learning model (i.e. “A baseline ML model is first trained on a baseline training dataset that includes the non-contrast medical images, each labelled with the ground truth label of the respective calcification parameter which was determined for that respective non-contrast medical image, as described herein. The ML model is then trained by further training the baseline ML model, using the transfer learning approach, on the training dataset that includes the contrast enhanced medical images, each labelled with ground truth label of the respective calcification parameter which were determined for the corresponding non-contrast medical image”, Paragraph [0104]).
As to claim 11, Gordon et al. teaches wherein said using of said at least part of the pre-trained machine-learning model comprises:
supervised training of said at least part of the pre-trained machine-learning model (i.e., “Machine learning models may be trained using supervised approaches”, Paragraph [0056]) using further training images which are annotated with ground-truth semantic context information (i.e., “The ML model is then trained by further training the baseline ML model, using the transfer learning approach, on the training dataset that includes the contrast enhanced medical images, each labelled with ground truth label of the respective calcification parameter which were determined for the corresponding non-contrast medical image”, Paragraph [0104]).
As to claim 12, Gordon et al. teaches wherein the semantic context information comprises at least one of information about presence of a region of interest in the inference image or segmentation information related to the region of interest (See for example, “a fine mask delineating the target calcification depicted in the target anatomical structure is obtained as an outcome of the ML model”, Paragraph [0110]).
As to claim 14, Lenga et al. teaches a computing device (i.e., “device 200”, Paragraph [0089]) comprising a processor (i.e., “processor 204”, Paragraph [0095]) and a memory (i.e., “memory 206”, Paragraph [0094]), the memory comprising instructions executable by the processor, wherein when executing the instructions at the processor (i.e., “Device 200 may perform these processes based on processor 204 executing software instructions stored on a computer-readable medium, such as memory 206”, Paragraph [0094]), the computing device is configured to perform the computer-implemented method of (i.e., “When executed, software instructions stored in memory 206 and/or storage component 208 may cause processor 204 to perform one or more processes described herein”, Paragraph [0095]; also, refer to claim 1 above):
generating a pre-trained machine-learning model (i.e., “radiological machine learning system 102 may analyze the training data to generate the predictive machine learning model”, Paragraph [0145]) by unsupervised pre-training of a machine-learning model (i.e., “The machine learning techniques may include, for example, supervised and/or unsupervised techniques”, Paragraph [0145]) for predicting time-related information from at least one pre-training image (i.e., “prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]), the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period (i.e., “at three different time points t1, t2, and t3, radiological machine learning system 102 may generate three representations of an examination region of an examination object”, Paragraph [0153]), and the time-related information being associated with one or more points of time during the observation period (See for example, Paragraph [0153]; “The third radiological image, O3, may include an image of the examination region that was generated with a third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
However, Lenga et al. does not explicitly disclose building a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model being for predicting semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Gordon et al. teaches building a further machine-learning model (i.e., “ML model”, Paragraph [0104]) using at least part of a pre-trained machine-learning model (See for example, “The ML model is then trained by further training the baseline ML model, using the transfer learning approach”, Paragraph [0104]; see in addition “The ML model is trained on a training dataset of records, each record including a respective contrast enhanced medical image(s) and optional additional parameter(s), labelled with a ground truth label of calcification parameter(s), for example, as described with reference to FIG. 2”, Paragraph [0108]; see also, “At 206, images with a specific contrast phase may be selected … respective contrast enhanced medical image(s) are analyze to determine a respective contrast phase. The contrast phase depicted in the respective contrast enhanced medical image may be determined, for example, by extracting the contrast phase from metadata associated with the scan, and/or by a trained ML model that outputs an indication of the contrast phase. Optionally, contrast enhanced medical images designed one or more specific phases (e.g., arterial, portal venous, delayed) may be selected for including in a training dataset, along with the corresponding non-contrast enhanced image”, Paragraphs [0075]-[0076]), the further machine-learning model being for predicting semantic context information (See for example, “a target calcification parameter(s) is obtained for a target calcification depicted in the target anatomical structure as an outcome of the ML model … a fine mask delineating the target calcification depicted in the target anatomical structure is obtained as an outcome of the ML model”, Paragraphs [0109]-[0110]) from at least one inference image (i.e., “a target contrast enhanced image of a target subject is accessed”, Paragraph [0105]) acquired by the contrast-enhanced medical imaging system (i.e., “Imaging device 112 and/or data repository 114 that store images acquired by imaging device 112. The acquired images may be fed into trained ML model(s) 122A for inference thereof”, Paragraph [0061]) using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Therefore, in view of Gordon et al., it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Lenga et al. by incorporating the building of a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model being for predicting semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement, as taught by Gordon et al., in order to increase the performance of a machine-learning model when identifying a calcification.
As to claim 16, Lenga et al. teaches a non-transitory computer-readable storage medium (i.e., “computer-readable medium”, Paragraph [0094]) storing instructions that, when executed by a computer, cause the computer to carry out the computer-implemented method of:
generating a pre-trained machine-learning model (i.e., “radiological machine learning system 102 may analyze the training data to generate the predictive machine learning model”, Paragraph [0145]) by unsupervised pre-training of a machine-learning model (i.e., “The machine learning techniques may include, for example, supervised and/or unsupervised techniques”, Paragraph [0145]) for predicting time-related information from at least one pre-training image (i.e., “prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]), the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period (i.e., “at three different time points t1, t2, and t3, radiological machine learning system 102 may generate three representations of an examination region of an examination object”, Paragraph [0153]), and the time-related information being associated with one or more points of time during the observation period (See for example, Paragraph [0153]; “The third radiological image, O3, may include an image of the examination region that was generated with a third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
However, Lenga et al. does not explicitly disclose building a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model being for predicting semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Gordon et al. teaches building a further machine-learning model (i.e., “ML model”, Paragraph [0104]) using at least part of a pre-trained machine-learning model (See for example, “The ML model is then trained by further training the baseline ML model, using the transfer learning approach”, Paragraph [0104]; see in addition “The ML model is trained on a training dataset of records, each record including a respective contrast enhanced medical image(s) and optional additional parameter(s), labelled with a ground truth label of calcification parameter(s), for example, as described with reference to FIG. 2”, Paragraph [0108]; see also, “At 206, images with a specific contrast phase may be selected … respective contrast enhanced medical image(s) are analyze to determine a respective contrast phase. The contrast phase depicted in the respective contrast enhanced medical image may be determined, for example, by extracting the contrast phase from metadata associated with the scan, and/or by a trained ML model that outputs an indication of the contrast phase. Optionally, contrast enhanced medical images designed one or more specific phases (e.g., arterial, portal venous, delayed) may be selected for including in a training dataset, along with the corresponding non-contrast enhanced image”, Paragraphs [0075]-[0076]), the further machine-learning model being for predicting semantic context information (See for example, “a target calcification parameter(s) is obtained for a target calcification depicted in the target anatomical structure as an outcome of the ML model … a fine mask delineating the target calcification depicted in the target anatomical structure is obtained as an outcome of the ML model”, Paragraphs [0109]-[0110]) from at least one inference image (i.e., “a target contrast enhanced image of a target subject is accessed”, Paragraph [0105]) acquired by the contrast-enhanced medical imaging system (i.e., “Imaging device 112 and/or data repository 114 that store images acquired by imaging device 112. The acquired images may be fed into trained ML model(s) 122A for inference thereof”, Paragraph [0061]) using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Therefore, in view of Gordon et al., it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Lenga et al. by incorporating the building of a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model being for predicting semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement, as taught by Gordon et al., in order to increase the performance of a machine-learning model when identifying a calcification.
As to claim 17, Lenga et al. teaches wherein said unsupervised pre-training of a machine-learning model for predicting time-related information comprises:
obtaining said at least one pre-training image of the patient acquired by the contrast-enhanced medical imaging system at the first point in time during the observation period (i.e., “the first representation, F1, and the second representation, F2, may be provided as input to the predictive machine learning model”, Paragraph [0159]);
applying said machine-learning model to the at least one pre-training image, wherein said time-related information is predicted for the second point in time during the observation period (i.e., “The output of the predictive machine learning model may include a prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]);
obtaining ground-truth information based on a further image acquired by the contrast-enhanced medical imaging system at the second point in time (i.e., O3 at t3, Paragraph [0153]); and
training the machine-learning model based on comparing the ground-truth information and the time-related information predicted for the second point in time (i.e., “radiological machine learning system 102 may compare the prediction of the representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique to the third representation in the frequency space of the examination region, F3, with the third amount of the contrast agent administered during the medical imaging technique …”, Paragraph [0159]).
As to claim 18, Lenga et al. teaches wherein the first point in time corresponds to a pre-contrast phase of the observation period prior to a contrast agent being introduced into the patient (i.e., “The first radiological image, O1, may include an image of the examination region that was generated without an amount of a contrast agent administered during the medical imaging technique”, Paragraph [0154]), and
wherein the second point in time corresponds to a post-injection phase of the observation period after the contrast agent is introduced into the patient (i.e., “The second radiological image, O2, may include an image of the examination region that was generated with a second amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
As to claim 19, Lenga et al. teaches wherein the time-related information comprises at least one further pre-training image at the one or more points of time during the observation period (See for example, “radiological machine learning system 102 may receive the output, F’3*, of the predictive machine learning model based on the input and convert the output of the predictive machine learning model to a predicted representation in real space, O’3*, of the examination region of the examination object using an iFT”, Paragraph [0162]).
As to claim 20, Lenga et al. teaches a computing device (i.e., “environment 100”, Paragraph [0083]) comprising:
a memory (i.e., “memory 206”, Paragraph [0094]) storing computer-executable instructions; and
at least one processor (i.e., “processor 204”, Paragraph [0095]) configured to execute the computer-executable instructions to cause the computing device to
generate a pre-trained machine-learning model (i.e., “radiological machine learning system 102 may analyze the training data to generate the predictive machine learning model”, Paragraph [0145]) by unsupervised pre-training of a machine-learning model (i.e., “The machine learning techniques may include, for example, supervised and/or unsupervised techniques”, Paragraph [0145]) for predicting time-related information from at least one pre-training image (i.e., “prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]),
the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period (i.e., “at three different time points t1, t2, and t3, radiological machine learning system 102 may generate three representations of an examination region of an examination object”, Paragraph [0153]), the time-related information being associated with one or more points of time during the observation period (See for example, Paragraph [0153]; “The third radiological image, O3, may include an image of the examination region that was generated with a third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
However, Lenga et al. does not explicitly disclose the at least one processor configured to execute the computer-executable instructions to cause the computing device to build a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model configured to predict semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Gordon et al. teaches at least one processor (i.e., “hardware processor(s) 102”, Paragraph [0046]) configured to execute computer-executable instructions to cause a computing device (i.e., “System 100”, Paragraph [0046]) to build a further machine-learning model (i.e., “ML model”, Paragraph [0104]) using at least part of a pre-trained machine-learning model (See for example, “The ML model is then trained by further training the baseline ML model, using the transfer learning approach”, Paragraph [0104]; see in addition “The ML model is trained on a training dataset of records, each record including a respective contrast enhanced medical image(s) and optional additional parameter(s), labelled with a ground truth label of calcification parameter(s), for example, as described with reference to FIG. 2”, Paragraph [0108]; see also, “At 206, images with a specific contrast phase may be selected … respective contrast enhanced medical image(s) are analyze to determine a respective contrast phase. The contrast phase depicted in the respective contrast enhanced medical image may be determined, for example, by extracting the contrast phase from metadata associated with the scan, and/or by a trained ML model that outputs an indication of the contrast phase. Optionally, contrast enhanced medical images designed one or more specific phases (e.g., arterial, portal venous, delayed) may be selected for including in a training dataset, along with the corresponding non-contrast enhanced image”, Paragraphs [0075]-[0076]), the further machine-learning model configured to predict semantic context information (See for example, “a target calcification parameter(s) is obtained for a target calcification depicted in the target anatomical structure as an outcome of the ML model … a fine mask delineating the target calcification depicted in the target anatomical structure is obtained as an outcome of the ML model”, Paragraphs [0109]-[0110]) from at least one inference image (i.e., “a target contrast enhanced image of a target subject is accessed”, Paragraph [0105]) acquired by the contrast-enhanced medical imaging system (i.e., “Imaging device 112 and/or data repository 114 that store images acquired by imaging device 112. The acquired images may be fed into trained ML model(s) 122A for inference thereof”, Paragraph [0061]) using the contrast-enhanced measurement or a further contrast-enhanced measurement.
Therefore, in view of Gordon et al., it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify Lenga et al. by incorporating the at least one processor configured to execute the computer-executable instructions to cause the computing device to build a further machine-learning model using at least part of the pre-trained machine-learning model, the further machine-learning model configured to predict semantic context information from at least one inference image acquired by the contrast-enhanced medical imaging system using the contrast-enhanced measurement or a further contrast-enhanced measurement, as taught by Gordon et al., in order to increase the performance of a machine-learning model when identifying a calcification.
As to claim 25, Lenga et al. teaches wherein the time-related information is based on contrast agent distribution phases associated with the at least one pre-training image (See for example, “The third representation, F3, may include a representation of the examination region with a third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0155]; and “As part of a training procedure, radiological machine learning system 102 may compare the prediction of the representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique to the third representation in the frequency space of the examination region, F3, with the third amount of the contrast agent administered during the medical imaging technique. A deviation between the prediction of the representation, F3*, and the third representation, F3, may be used in a backpropagation method to train the predictive machine learning model to reduce deviations to a defined minimum”, Paragraph [0159]).
Claims 6 and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Lenga et al. in view of Gordon et al. as applied to claim 1 above, and further in view of Ma et al. (U.S. Pub. No. 2021/0279868). The teachings of Lenga et al. and Gordon et al. have been discussed above.
As to claim 6, Lenga et al. and Gordon et al. do not explicitly disclose wherein the time-related information comprises statistical information for image pixel intensities across the observation period.
Ma et al. teaches time-related information that comprises statistical information for image pixel intensities across the observation period (See for example, “The horizontal axis indicates time in minutes and the vertical axis indicates density (average amplitude within the image region defined as an anatomic feature in HU”, Paragraph [0116]).
Lenga et al., Gordon et al. and Ma et al. are analogous art because they are from the field of digital image processing for processing contrast enhanced images.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to further modify Lenga et al. and Gordon et al. by incorporating the time-related information comprises statistical information for image pixel intensities across the observation period, as taught by Ma et al.
The suggestion/motivation for doing so would have been to accurately determine contrast phases in contrast enhanced images.
Therefore, it would have been obvious to combine Ma et al. with Lenga et al. and Gordon et al. to obtain the invention as specified in claim 6.
As to claim 21, Ma et al. teaches wherein the statistical information includes at least one of a variance or a standard deviation of the image pixel intensities across the observation period (See for example, Paragraph [0119]).
Claims 7 and 8 are rejected under 35 U.S.C. 103 as being unpatentable over Lenga et al. in view of Gordon et al. as applied to claim 1 above, and further in view of Goshen (U.S. Pub. No. 2024/0312086). The teachings of Lenga et al. and Gordon et al. have been discussed above.
As to claim 7, Lenga et al. and Gordon et al. do not explicitly disclose wherein the time-related information comprises a mask or a map for pixels of the at least one pre-training image.
Goshen teaches the time-related information comprises a mask or a map for pixels of the at least one pre-training image (See for example, Paragraph [0112]; and “the output of network G1=CR=GrĪ→I could be used to calculate an contrast agent-map”, Paragraph [0249]).
Lenga et al., Gordon et al. and Goshen are analogous art because they are from the field of digital image processing for processing contrast enhanced images.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to further modify Lenga et al. and Gordon et al. by incorporating the time-related information comprises a mask or a map for pixels of the at least one pre-training image, as taught by Goshen.
The suggestion/motivation for doing so would have been to provide enhanced training images.
Therefore, it would have been obvious to combine Goshen with Lenga et al. and Gordon et al. to obtain the invention as specified in claim 7.
As to claim 8, Lenga et al. and Gordon et al. do not explicitly disclose wherein the pre-trained machine-learning model comprises at least one of an autoencoder neural network architecture or a u-net neural network architecture.
Goshen teaches a pre-trained machine-learning model that comprises at least one of an autoencoder neural network architecture or a u-net neural network architecture (See for example, “The architecture is thus one of an autoencoder”, Paragraph [0150]).
Therefore, in view of Goshen, it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify Lenga et al. and Gordon et al. by incorporating the pre-trained machine-learning model comprises at least one of an autoencoder neural network architecture or a u-net neural network architecture, as taught by Goshen, in order to learn features to be used for training with reduced dimensions.
Claims 10 and 24 are rejected under 35 U.S.C. 103 as being unpatentable over Lenga et al. in view of Gordon et al. as applied to claims 1 and 9 above, and further in view of Rohrer et al. (U.S. Pub. No. 2022/0414972). The teachings of Lenga et al. and Gordon et al. have been discussed above.
As to claim 10, Lenga et al. and Gordon et al. do not explicitly disclose wherein said at least part of the pre-trained machine-learning model generates embedded features from the at least one inference image in the further machine-learning model, and wherein the semantic context information for the at least one inference image is determined based on the embedded features.
Rohrer et al. teaches at least part of the pre-trained machine-learning model generates embedded features (i.e., “a characteristic associated with tissue of a patient may include at least one of the following: … one or more of the pharmacokinetic parameters associated with contrast movement through the tissue spaces of a voxel”, Paragraph [0182]) from the at least one inference image in the further machine-learning model (i.e., “image analysis and computation system 614 may determine, based on the measurement information including the parameter of the voxel at the two or more time points, the one or more characteristics … using at least one of the following techniques: … a feature extraction technique using CAD or AI, … 2D AI or machine learning based analysis of a single image, a 2D AI or machine learning based analysis technique of time sequence for a voxel, a 3D AI or machine learning based analysis technique of a time sequence of images technique, an AI or machine learning based denoising or sharpening technique, an AI or machine learning based contrast amplification technique, or any combination thereof”, Paragraph [0183]), and
wherein the semantic context information for the at least one inference image is determined based on the embedded features (i.e., “computation system 614 may analyse the one or more characteristics associated with the voxel of the tissue of the patient and/or the one or more images the one or more characteristics associated with the tissue of the patient to derive diagnosis information from the one or more images. For example, image analysis and computation system 614 may feed the one more characteristics (and/or the one or more images including or showing the one or more characteristics), to a classification model, the classification model having been trained by means of supervised learning to classify, on the basis of the one more characteristics (and/or the one or more images including or showing the one or more characteristics, the tissue as associated with one or more diagnoses”, Paragraph [0187]).
Lenga et al., Gordon et al. and Rohrer et al. are analogous art because they are from the field of digital image processing for processing contrast enhanced images.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to further modify Lenga et al. and Gordon et al. by incorporating the at least part of the pre-trained machine-learning model generates embedded features from the at least one inference image in the further machine-learning model, and the semantic context information for the at least one inference image is determined based on the embedded features, as taught by Rohrer et al.
The suggestion/motivation for doing so would have been to utilize pharmacokinetic characteristics to classify a tissue associated with one or more diagnoses.
Therefore, it would have been obvious to combine Rohrer et al. with Lenga et al. and Gordon et al. to obtain the invention as specified in claim 10.
As to claim 24, Lenga et al. and Gordon et al. do not explicitly disclose wherein the time-related information is resolved in the time domain.
Rohrer et al. teaches the time-related information is resolved in the time domain (See for example, “In the hepatobiliary phase depicted in FIG. 4(f), the liver cells (P) are depicted with signal enhancement; the blood vessels and the tumour no longer have contrast agent and are accordingly depicted darkly”, Paragraph [0161]; and “three MRI images (1), (2) and (3) showing a liver in a first time span are fed to a prediction model (PM). The prediction model calculates from the three MRI images (1), (2) and (3) an MRI image (4) showing the liver in a second time span”, Paragraph [0162]).
Therefore, in view of Rohrer et al., it would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify Lenga et al. and Gordon et al. by incorporating the time-related information is resolved in the time domain, as taught by Rohrer et al., in order to better depict the hepatobiliary phase without movement artefacts.
Claim 13 is rejected under 35 U.S.C. 103 as being unpatentable over Gordon et al. (U.S. Pub. No. 2022/0405915) in view of Lenga et al. (U.S. Pub. No. 2024/0050054).
As to claim 13, Gordon et al. teaches a computer-implemented method for predicting semantic context information from an image acquired by a contrast-enhanced medical imaging system (i.e., “providing a machine learning model that identifies calcification and/or computes calcification parameter(s) for the calcifications in anatomical structures (e.g., arteries) depicted in contrast enhanced images”, Paragraph [0032]) using a further machine-learning model (i.e., “ML model”, Paragraph [0104]) built according to the computer-implemented method of:
building a further machine-learning model using at least part of a pre-trained machine-learning model (See for example, “The ML model is then trained by further training the baseline ML model, using the transfer learning approach”, Paragraph [0104]; see in addition “The ML model is trained on a training dataset of records, each record including a respective contrast enhanced medical image(s) and optional additional parameter(s), labelled with a ground truth label of calcification parameter(s), for example, as described with reference to FIG. 2”, Paragraph [0108]; see also, “At 206, images with a specific contrast phase may be selected … respective contrast enhanced medical image(s) are analyze to determine a respective contrast phase. The contrast phase depicted in the respective contrast enhanced medical image may be determined, for example, by extracting the contrast phase from metadata associated with the scan, and/or by a trained ML model that outputs an indication of the contrast phase. Optionally, contrast enhanced medical images designed one or more specific phases (e.g., arterial, portal venous, delayed) may be selected for including in a training dataset, along with the corresponding non-contrast enhanced image”, Paragraphs [0075]-[0076]), the further machine-learning model being for predicting semantic context information (See for example, “a target calcification parameter(s) is obtained for a target calcification depicted in the target anatomical structure as an outcome of the ML model … a fine mask delineating the target calcification depicted in the target anatomical structure is obtained as an outcome of the ML model”, Paragraphs [0109]-[0110]) from at least one inference image (i.e., “a target contrast enhanced image of a target subject is accessed”, Paragraph [0105]) acquired by the contrast-enhanced medical imaging system (i.e., “Imaging device 112 and/or data repository 114 that store images acquired by imaging device 112. The acquired images may be fed into trained ML model(s) 122A for inference thereof”, Paragraph [0061]) using the contrast-enhanced measurement or a further contrast-enhanced measurement.
However, Gordon et al. does not explicitly disclose generating a pre-trained machine-learning model by unsupervised pre-training of a machine-learning model for predicting time-related information from at least one pre-training image, the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period, and the time-related information being associated with one or more points of time during the observation period.
Lenga et al. teaches generating a pre-trained machine-learning model (i.e., “radiological machine learning system 102 may analyze the training data to generate the predictive machine learning model”, Paragraph [0145]) by unsupervised pre-training of a machine-learning model (i.e., “The machine learning techniques may include, for example, supervised and/or unsupervised techniques”, Paragraph [0145]) for predicting time-related information from at least one pre-training image (i.e., “prediction of a representation in the frequency space of the examination region, F3*, with the third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0159]), the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period (i.e., “at three different time points t1, t2, and t3, radiological machine learning system 102 may generate three representations of an examination region of an examination object”, Paragraph [0153]), and the time-related information being associated with one or more points of time during the observation period (See for example, Paragraph [0153]; “The third radiological image, O3, may include an image of the examination region that was generated with a third amount of the contrast agent administered during the medical imaging technique”, Paragraph [0154]).
Gordon et al. and Lenga et al. are analogous art because they are from the field of digital image processing for processing contrast enhanced images.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to modify Gordon et al. by incorporating the generating of a pre-trained machine-learning model by unsupervised pre-training of a machine-learning model for predicting time-related information from at least one pre-training image, the at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period, and the time-related information being associated with one or more points of time during the observation period, as taught by Lenga et al.
The suggestion/motivation for doing so would have been to provide a prediction of a representation of an examination region that was generated using a medical image technique involving a contrast agent.
Therefore, it would have been obvious to combine Lenga et al. with Gordon et al. to obtain the invention as specified in claim 13.
Claim 23 is rejected under 35 U.S.C. 103 as being unpatentable over Lenga et al. in view of Gordon et al. as applied to claim 1 above, and further in view of Tamir et al. (U.S. Pub. No. 2022/0334208). The teachings of Lenga et al. and Gordon et al. have been discussed above.
As to claim 23, Lenga et al. and Gordon et al. do not explicitly disclose wherein the at least one pre-training image includes a plurality of pre-training images obtained from a plurality of perspectives of a region of interest.
Tamir et al. teaches the at least one pre-training image includes a plurality of pre-training images obtained from a plurality of perspectives of a region of interest (See for example, “Using the multiplanar reconstruction (MPR) technique, the deep learning model may be trained with volumetric images (e.g., augmented 2.5D images) such as from the multiple orientations (e.g., three principal axes)”, Paragraph [0055]; and Paragraphs [0058]-[0059]).
Lenga et al., Gordon et al. and Tamir et al. are analogous art because they are from the field of digital image processing for processing contrast enhanced images.
Before the effective filing date of the claimed invention, it would have been obvious to a person of ordinary skill in the art to further modify Lenga et al. and Gordon et al. by incorporating the at least one pre-training image includes a plurality of pre-training images obtained from a plurality of perspectives of a region of interest, as taught by Tamir et al.
The suggestion/motivation for doing so would have been to better predict contrast-enhanced images with contrast dose reduction across different sites and scanners.
Therefore, it would have been obvious to combine Tamir et al. with Lenga et al. and Gordon et al. to obtain the invention as specified in claim 23.
Response to Arguments
Claim Objections
With respect to claim 18, Applicant has amended the claim in order to correct for minor informalities. Therefore, the objection has been withdrawn.
Claim Rejections - 35 USC § 103
With respect to claims 1-22, Applicant’s arguments (Remarks dated February 27, 2026, pages 10-13) have been fully considered, but they are not persuasive. With respect to claim 1, Applicant respectfully submits that “Paragraph [0104] of Gordon discloses that “[a] baseline ML model is first trained on a baseline training dataset that includes the non-contrast medical images." The disclosed model “is then trained by further training the baseline ML model on the training dataset that includes the contrast enhanced medical images.” This fails to disclose the “building a further machine-learning model using at least part of the pre-trained machine-learning model,” as recited by claim 1, at least because the “pre-trained machine-learning model,” recited by claim 1, is trained on “at least one pre-training image acquired by a contrast-enhanced medical imaging system using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period.” (Remarks dated February 27, 2026, pages 10-11). Examiner respectfully disagrees.
While Gordon et al. does teach that the baseline ML model is first trained on a baseline training dataset that includes the non-contrast medical images (See for example, Paragraph [0104]), upon further consideration Gorgon et al. also teaches that the ML model is trained on a training dataset of records, each record including a respective contrast enhanced medical image(s) and optional additional parameter(s), labelled with a ground truth label of calcification parameter(s), for example, as described with reference to FIG. 2 (i.e., Paragraph [0108]). The training dataset, as taught by Gordon et al., include selected images with a specific contrast phase. The respective contrast enhanced medical images are analyzed to determine a respective contrast phase. The contrast phase depicted in the respective contrast enhanced medical image may be determined, for example, by extracting the contrast phase from metadata associated with the scan, and/or by a trained ML model that outputs an indication of the contrast phase. Optionally, contrast enhanced medical images designed one or more specific phases (e.g., arterial, portal venous, delayed) may be selected for including in a training dataset, along with the corresponding non-contrast enhanced image (See Gordon et al. at Paragraphs [0075] and [0076]). Thus, the ML model (e.g., further machine-learning model) is trained by further training a baseline ML model (e.g., at least part of the pre-trained machine-learning model), and from Figure 2, it can be shown that the baseline ML model is trained using a contrast-enhanced measurement of a patient with multiple contrast agent distribution phases during an observation period, as claimed.
Therefore, the rejections are maintained.
With respect to claims 2-22, Applicant’s arguments (Remarks dated February 27, 2026, pages 11-13) have been fully considered, and are no different from those previously presented with respect to claim 1 and already addressed above.
Therefore, the rejections are maintained.
Conclusion
THIS ACTION IS MADE FINAL. Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to JOSE M TORRES whose telephone number is (571)270-1356. The examiner can normally be reached Monday thru Friday; 10:00 AM to 6:00 PM EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Jennifer Mehmood can be reached at 571-272-2976. 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.
/JOSE M TORRES/Examiner, Art Unit 2664 05/29/2026
/JENNIFER MEHMOOD/Supervisory Patent Examiner, Art Unit 2664
Prosecution Insights