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 .
DETAILED ACTION
Status of Claims
This action is in reply to the amendment filed on 09/03/2026.
Claims 1 and 24 have been amended.
Claims 3, 6-10, 13-14, 16, 18-23, 26-29, 31-39, 42, 44-46 and 51-52 have been cancelled.
Claims 53-56 have been newly added.
Claims 1-2, 4-5, 11-12, 15, 17, 24-25, 30, 40-41, 43, 47-50 and 53-56 are currently pending and have been examined.
Claim Objections
The numbering of claims is not in accordance with 37 CFR 1.126 which requires the original numbering of the claims to be preserved throughout the prosecution. When claims are canceled, the remaining claims must not be renumbered. When new claims are presented, they must be numbered consecutively beginning with the number next following the highest numbered claims previously presented (whether entered or not).
Misnumbered claims 6-10 are objected to because of the following informalities: Status of claims 6-10 indicated “Cancelled” with Request for Continued Examination was entered on 11/12/2025, but renumbered claims 6-7 have been entered in “Original” form on 09/03/2026. Cancelled claims 6-10 are not examined.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1-2, 4-5, 11-12, 15, 17, 24-25, 30, 40-41, 43, 47-50 and 53-56 are rejected under 35 U.S.C. §101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1:
Claims 1-2, 4-7, 11-12 and 14-17 are directed to a method (i.e., a process) and claims 24-25, 30, 40-41, 43 and 49-52 are directed to a system (i.e., a machine). Accordingly, claims 1-2, 4-5, 11-12, 15, 17, 24-25, 30, 40-41, 43, 47-50 and 53-56 are all within at least one of the four statutory categories.
Step 2A - Prong One:
An “abstract idea” judicial exception is subject matter that falls within at least one of the following groupings: a) mathematical concepts, b) certain methods of organizing human activity, and/or c) mental processes.
Representative independent claim 24 includes limitations that recite an abstract idea. Note that independent claim 24 is the system claim and claim 1 covers a method claim.
Specifically, independent claim 24 recites:
A system for generating a medical concordance score from a test subject medical data set using an electronic neural network, the system comprising:
a processor; and
a memory communicatively coupled to the processor, the memory storing instructions which, when executed on the processor, perform operations comprising:
receiving a test subject medical data set comprising an image of a histopathology slide, wherein a size of the image exceeds a volatile memory capacity of the system;
segmenting the image of the histopathology slide into a plurality of;
quality-assuring the plurality of tiles by removing tiles that contain pen ink and tiles that are out of focus, thereby producing quality-assured tiles;
embedding the quality-assured tiles to produce embedded tiles
passing the embedded tile of a test subject medical data set through the electronic neural network, wherein the electronic neural network has been trained on a set of training data that comprises a plurality of reference subject medical data sets that are each labeled with a medical determination and are each assigned a ground truth concordance score generated by a plurality of experts, wherein a value of a given ground truth concordance score comprises a fraction of the plurality of experts, if any, that are in accord with the medical determination label of a given reference subject medical data set in the plurality of reference subject medical data sets, wherein the medical determination comprises an atypical status, a benign status, or a malignant status, wherein the reference subject medical data sets comprise images of histopathology slides, wherein each embedded tile is treated as an instance of a bag containing all quality-assured tiles of an image of a given histopathology slide, wherein the electronic neural network aggregates the instances of the bag using attention mechanism, and wherein a final layer of the electronic neural network performs linear regression instead of a classification to output a continuous concordance score in a range of 0 to 1;
outputting from the electronic neural network the continuous concordance score of the medical determination being indicated by the test subject medical data set; and
ordering one or more medical tests for, recommending administering one or more therapies to, and/or recommending discontinuing administering one or more therapies to, the test subject when the continuous concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value.
The Examiner submits that the foregoing underlined limitations constitute: (a) “certain methods of organizing human activity” because sorting images for storage, determining a medical condition of a test subject by a group of experts, when the experts are in an concordance, determining an atypical status, benign status, or malignant status, recommending administering therapies and having agreement or consensus are providing healthcare services and diagnosing, which is managing human behavior/interactions between people. Furthermore, these limitations constitute (b) “a mathematical concept” because generating a medical concordance score, assigning and outputting a concordance score of the medical determination based on a predetermined threshold value and the test subject medical data set and outputting a continuous concordance score in a range of 0 to 1 are done mathematically. The foregoing underlined limitations also relate to claim 1 (similarly to claim 24).
Accordingly, the claim describes at least one abstract idea.
In relation to claims 2, 5, 11-12, 15, 25, 30, 41, 43, 47-50 and 56, these claims merely recite specific kinds of data, such as: claims 2 and 25 - the plurality of experts comprises human experts, machine experts, or a combination of human and machine experts, claim 5 - the medical determination comprises a diagnosis of a disease, condition, or disorder, claim 11 – the medical determination comprises a survival quantification, claim 12 - the medical determination comprises a therapy response, claim 15 - the one or more medical tests comprise at least one histological stain of a sample obtained from the test subject, claim 30 - the medical determination comprises a recommended treatment plan for a diagnosed disease, condition, or disorder, claim 41 - the test and reference subject medical data sets comprise images of histopathology slides, claim 43 - the test and reference subject medical data sets comprise images selected from the group consisting of: a magnetic resonance (MR) image, a computed tomography (CT) image, a single photon emission computed tomography (SPECT) image, a positron emission tomography (PET) image, and a microscopy image, claim 47 and 49 -the medical determination comprises an in situ melanoma status or an invasive melanoma status and claims 48 and 50 - the images comprise whole slide images (WSJ) and claim 56 - activated attention heads, and wherein the electronic neural network is trained using a root- mean-squared-error (RMSE) regression objective rather than a classification objective.
In relation to claims 4, 17, 40 and 51-55, these claims merely recite determining steps such as: claim 4 - labeling the test subject medical data set with the medical determination prior to or when passing the test subject medical data set through the electronic neural network, claims 17 and 40 - generating or updating at least a portion of a medical report for the test subject when the concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value, claim 51 -the electronic neural network has been trained using a multiple instance learning (MIL) model, claim 52 - size of the images of histopathology slides exceeds a storage capacity of memory operably connected to a system that comprises the electronic neural network, claim 53 - the plurality of tiles comprises 128 pixel by 128 pixel tiles extracted at an objective power of 10x, claim 54 – removing tiles that are out of focus comprises setting a threshold on a variance of a Laplacian computed over each tile, and wherein tiles containing pen ink are identified by a model trained to detect inked slides and are removed from training, validation, and inference, and claim 55 - embedding the quality-assured tiles comprises propagating each quality-assured tile through an embedder trained in a SimCLR framework with a ResNet50 backbone to produce a 2048-dimension feature vector for each quality- assured tile.
Step 2A - Prong Two:
Regarding Prong Two of Step 2A, it must be determined whether the claim as a whole integrates the abstract idea into a practical application. As noted, it must be determined whether any additional elements in the claim beyond the abstract idea integrate the exception into a practical application in a manner that imposes a meaningful limit on the judicial exception. The courts have indicated that additional elements merely using a computer to implement an abstract idea, adding insignificant extra solution activity, or generally linking use of a judicial exception to a particular technological environment or field of use do not integrate a judicial exception into a “practical application.”
The limitations of claims 1 and 24, as drafted is a process that, under its broadest reasonable interpretation, covers performance of the limitations by humans using math but for the recitation of generic computer components. That is, other than reciting a system, a processor, and a memory communicatively coupled to the processor, the memory storing instructions to perform the limitations, nothing in the claim elements precludes the steps from practically being performed in the mind. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation within a health care environment done by humans using math but for the recitation of generic computer components, then it falls within the “certain methods of organizing human activity” and a “Mathematical Concept” grouping of abstract ideas. Accordingly, the claims recite an abstract idea.
The judicial exception is not integrated into a practical application. In particular, the system, processor, and memory communicatively coupled to the processor, the memory storing instructions are recited at high levels of generality (i.e., as generic computer components performing generic computer functions of receiving data/inputs, determining and providing data) such that it amounts no more than mere instructions to apply the exception using the generic computer components.
Regarding the additional limitations “using an electronic neural network”, “the electronic neural network has been trained on a set of training data”, and “machine experts” the Examiner submits that this additional limitation amount to merely using a computer to perform the at least one abstract idea (see MPEP § 2106.05(f)).
Thus, taken alone, the additional elements do not amount to significantly more than the above identified judicial exception (the abstract idea). Looking at the limitations as an ordered combination add nothing that is not already present when looking at the elements taken individually. For instance, there is no indication that the additional elements, when considered as a whole, reflect an improvements in the functioning of a computer or an improvement to another technology or technical field, apply or us the above-noted implement/use to above-noted judicial exception with a particular machine or manufacture that is integral to the claim, effect a transformation or reduction of a particular article to a different state or thing, or apply or use the judicial exception in some meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is not more than a drafting effort designed to monopolize the exception (see MPEP §2106.05). Their collective functions merely provide conventional computer implementation.
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed above with respect to the integration of the abstract idea into practical application, the additional elements amount to no more than mere instructions to apply the exception using generic computer components. Mere instructions to apply an exception using generic computer component provide an inventive concept. The claims are not patent eligible.
Step 2B:
Regarding Step 2B, in representative independent claim 24, regarding the additional limitations of the system, processor, and memory communicatively coupled to the processor, the memory storing instructions, the Examiner submits that these limitations amount to merely using a computer to perform the at least one abstract idea (see MPEP § 2106.05(f)).
Thus, representative independent claim 24 and analogous independent claim 1 do not include additional elements (considered both individually and as an ordered combination) that are sufficient to amount to significantly more than the judicial exception for the same reasons to those discussed above with respect to determining that the claim does not integrate the abstract idea into a practical application.
The dependent claims no not include additional elements (considered both individually and as an ordered combination) that are sufficient to amount to significantly more than the judicial exception for the same reason discussed above with respect to determining that the dependent claims do not integrate the at least abstract idea into a practical application.
Therefore, claims 1-2, 4-5, 11-12, 15, 17, 24-25, 30, 40-41, 43, 47-50 and 53-56 are ineligible under 35 USC §101.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-2, 4-5, 12, 24-25, 30, 53-54 and 30 are rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1).
Claim 1:
Cohen discloses A Computer-implemented method of ordering medical tests for, administering therapy to, and/or discontinuing administering therapy to, a test subject (See [P0043] The machine learning system yields a numerical risk score for each patient tested, which can be used by physicians to make treatment decisions concerning the therapy of cancer patients or, importantly, to further inform screening procedures to better predict and diagnose early stage cancer in patients. Also, see P0174, P0256-P0257 and P0391.), the method comprising:
passing the embedded tiles of a test subject medical data set through the electronic neural network, wherein the electronic neural network has been trained using a multiple instance learning (MIL) model on a set of training data that comprises a plurality of reference subject medical data sets that are each labeled with a medical determination (See Fig. 1A-2B, neural networks mentioned in P0274-P0276, exemplary biomarker (P0173-P0174, P0176) and other detectable labels (P0179-P0180) serve as labelling with a medical determination. Also, see training and validation set mentioned in P0150-P0151 and P0241-P0242.) and are each assigned a ground truth concordance score generated by a plurality of experts, wherein a value of a given ground truth concordance score comprises a fraction of the plurality of experts (See composite score and risk score mentioned in P0014, P0032 shown in Fig. 10.), if any, that are in accord with the medical determination label of a given reference subject medical data set in the plurality of reference subject medical data sets, wherein the medical determination comprises an atypical status, a benign status, or a malignant status (See [P0014] classify the risk score into risk categories for advising the clinician the likelihood that the nodule is or is not malignant, wherein the risk categories are derived from a same cohort population as the patient and wherein each risk category is associated with a benign or malignant grouping, to determine a likelihood of the patient having benign nodules or malignant nodules.), wherein the reference subject medical data sets comprise images of histopathology slides, and wherein a final layer of the electronic neural network performs linear regression instead of a classification (See raw images to imaging studies as unstructured data (P0069-P0070, P0072), AI image processing data (P0227, P0235) when training data sets in [P0241-P0242] the neural net can be trained to learn to identify specific medical data (e.g., images, unstructured, structured, etc.). Neural net NN3 may classify the data into different data types, e.g., raw images, numeric/structured data, BM velocity, unstructured data, etc. Also, see using a multivariate logistic regression (MLR) model (P0078, P0141) and normalization data include linear transformation (P0114, P0208).);
outputting from the electronic neural network a concordance score of the medical determination being indicated by the test subject medical data (See P0289-P0290 where exemplary neural net NN1 is necessary to compute a risk score shown in Fig. 3.); and
ordering one or more medical tests for, administering one or more therapies to, and/or discontinuing administering one or more therapies to, the test subject when the concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value (See P0043, P0226-P0228 where machine learning and the AI perform reasoning and analysis in order to make decisions, automatically advising the patient and the provider to be tested more frequently or to take other actions, predicted with risk score in P0230-P0231.).
Although Cohen discloses a method and system of ordering medical tests for, administering therapy to, or discontinuing administering therapy to a test subject based on passing a test subject medical data set through an electronic neural network, performing a final layer of the electronic neural network performs linear regression instead of a classification, instead of a classification for outputting a concordance score, resulting in ordering medical tests for, administering therapies or discontinuing administering therapies, and being indicated by the test subject medical data set varies from a predetermined threshold value mentioned above, Cohen does not explicitly teach initiating steps that include receiving an image of a histopathology slide, where a size of the image exceeds a volatile memory capacity of a computer system that comprises an electronic neural network, segmenting the image of the histopathology slide into tiles, quality-assuring the tiles by removing tiles that contain pen ink and tiles that are out of focus, by producing quality-assured tiles and embedding the quality-assured tiles to produce embedded tiles. Madabhushi teaches:
receiving a test subject medical data set comprising an image of a histopathology slide, wherein a size of the image exceeds a volatile memory capacity of a computer system that comprises an electronic neural network (Besides artifacts and batch effects including 30,000 digitized tissue slides hosted by The Cancer Genome Atlas (TCGA) mentioned P0019-P0021, see digital pathology (DP) slides split smaller and resized into tiles for reducing memory fit mentioned in P0041. See interface for the digital pathology (DP) slide quality control in [P0076-P0077] the machine learning classifier is configured to detect pen markings on a DP slide. The machine learning classifier may be a quadratic discriminant analysis (QDA) classifier, a linear discriminant analysis (LDA) classifier, a support vector machine (SVM) classifier, or may include a convolutional neural network or other deep-learning classifier. Also, see Fig. 8, Fig. 10 volatile memory in P0060-P0061, P0094, P0107.):
segmenting the image of the histopathology slide into a plurality of tiles (See Fig. 4, P0021 where optimal parameters include segmentation, training a machine learning classifier. Also, see [P0041] in this case, split a member of the plurality of DP images into smaller images according to the base magnification and base number of levels, sequentially loading smaller tiles, resizing the tiles, and then merging the tiles together.);
quality-assuring the plurality of tiles by removing tiles that contain pen ink and tiles that are out of focus, thereby producing quality-assured tiles (See Fig. 3 pen markings 312 serve as pen ink (P0027, P0030) when removing the artifacts. Also, see Fig. 5-6, P0022, P0025-P0029 and interface for the digital pathology (DP) slide quality control in P0076-P0077.);
embedding the quality-assured tiles to produce embedded tiles (See P0022-P0023 where embedding additional metrics include color histograms, brightness, contrast, edge detectors, smoothness detectors and supervised classifiers such as pen detection classifier to identify the DP slides with presentation differences.); and
passing the embedded tiles of a test subject (Taught in P0022-P0023.)
wherein each embedded tile is treated as an instance of a bag containing all quality- assured tiles of an image of a given histopathology slide, wherein the electronic neural network aggregates the instances of the bag using an attention mechanism, (See Fig. 4, P0021-P0023 where slide clusters serve as a bag and embedding additional metrics.).
Therefore, it would have been obvious to one of ordinary skill in the art of digital pathology slides before the effective filing date of the claimed invention to modify the method and system of Cohen to include an image of a histopathology slide, where a size of the image exceeds a volatile memory capacity of a computer system that comprises an electronic neural network, segmenting the image of the histopathology slide into tiles, quality-assuring the tiles by removing tiles that contain pen ink and tiles that are out of focus, by producing quality-assured tiles and embedding the quality-assured tiles to produce embedded tiles as taught by Madabhushi to identify DP slides which need to be reproduced or repaired, or regions in the DP slide which should be avoided during computational analysis, as an advantage in clinical diagnostic workflows and computational pathology mentioned on Madabhushi’s P0003.
Claim 24:
Cohen discloses a system for ordering medical tests for, recommending administering therapy to, and/or recommending discontinuing administering therapy to, a test subject using an electronic neural network (Fig. 1A-2B, neural networks, [P0043] The machine learning system yields a numerical risk score for each patient tested, which can be used by physicians to make treatment decisions concerning the therapy of cancer patients or, importantly, to further inform screening procedures to better predict and diagnose early stage cancer in patients. Also, see P0174, P0256-P0257 and P0391.), the system comprising:
a processor (See processor in P0194, P0220-P0221.); and
a memory communicatively coupled to the processor, the memory storing instructions which, when executed on the processor (See coupled memory in P0155, P0221-P0222.) perform operations comprising:
passing a test subject medical data set through the electronic neural network, wherein the electronic neural network has been trained on a set of training data that comprises a plurality of reference subject medical data sets that are each labeled with a medical determination (See Fig. 1A-2B, neural networks mentioned in P0274-P0276, exemplary biomarker (P0173-P0174, P0176) and other detectable labels (P0179-P0180) serve as labelling with a medical determination. Also, see training and validation set mentioned in P0150-P0151 and P0241-P0242.) and are each assigned a ground truth concordance score generated by a plurality of experts, wherein a value of a given ground truth concordance score comprises a fraction of the plurality of experts (See composite score and risk score mentioned in P0014, P0032 shown in Fig. 10.), if any, that are in accord with the medical determination label of a given reference subject medical data set in the plurality of reference subject medical data sets, wherein the medical determination comprises an atypical status, a benign status, or a malignant status (See [P0014] classify the risk score into risk categories for advising the clinician the likelihood that the nodule is or is not malignant, wherein the risk categories are derived from a same cohort population as the patient and wherein each risk category is associated with a benign or malignant grouping, to determine a likelihood of the patient having benign nodules or malignant nodules.) wherein the reference subject medical data sets comprise images of histopathology slides, and wherein a final layer of the electronic neural network performs linear regression instead of a classification (See raw images to imaging studies as unstructured data (P0069-P0070, P0072), AI image processing data (P0227, P0235) when training data sets in [P0241-P0242] the neural net can be trained to learn to identify specific medical data (e.g., images, unstructured, structured, etc.). Neural net NN3 may classify the data into different data types, e.g., raw images, numeric/structured data, BM velocity, unstructured data, etc. Also, see using a multivariate logistic regression (MLR) model (P0078, P0141) and normalization data include linear transformation (P0114, P0208).);
outputting from the electronic neural network the continuous concordance score of the medical determination being indicated by the test subject medical data set (See P0289-P0290 where exemplary neural net NN1 is necessary to compute a risk score shown in Fig. 3.); and
ordering one or more medical tests for, recommending administering one or more therapies to, and/or recommending discontinuing administering one or more therapies to, the test subject when the continuous concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value (See P0043, P0226-P0228 where machine learning and the AI perform reasoning and analysis in order to make decisions, automatically advising the patient and the provider to be tested more frequently or to take other actions, predicted with risk score in P0230-P0231.).
Although Cohen discloses a method and system of ordering medical tests for, administering therapy to, or discontinuing administering therapy to a test subject based on passing a test subject medical data set through an electronic neural network, performing linear regression instead of a classification for outputting a concordance score, resulting in ordering medical tests for, administering therapies or discontinuing administering therapies, and being indicated by the test subject medical data set varies from a predetermined threshold value mentioned above, Cohen does not explicitly teach initiating steps that include receiving an image of a histopathology slide, where a size of the image exceeds a volatile memory capacity of a computer system that comprises an electronic neural network, segmenting the image of the histopathology slide into tiles, quality-assuring the tiles by removing tiles that contain pen ink and tiles that are out of focus, by producing quality-assured tiles and embedding the quality-assured tiles to produce embedded tiles. Madabhushi teaches:
receiving a test subject medical data set comprising an image of a histopathology slide, wherein a size of the image exceeds a volatile memory capacity of the system (Besides artifacts and batch effects including 30,000 digitized tissue slides hosted by The Cancer Genome Atlas (TCGA) mentioned P0019-P0021, see digital pathology (DP) slides split smaller and resized into tiles for reducing memory fit mentioned in P0041. See interface for the digital pathology (DP) slide quality control in [P0076-P0077] the machine learning classifier is configured to detect pen markings on a DP slide. The machine learning classifier may be a quadratic discriminant analysis (QDA) classifier, a linear discriminant analysis (LDA) classifier, a support vector machine (SVM) classifier, or may include a convolutional neural network or other deep-learning classifier. Also, see Fig. 8, Fig. 10 volatile memory in P0060-P0061, P0094, P0107.):
segmenting the image of the histopathology slide into a plurality of tiles (See Fig. 4, P0021 where optimal parameters include segmentation, training a machine learning classifier. Also, see [P0041] in this case, split a member of the plurality of DP images into smaller images according to the base magnification and base number of levels, sequentially loading smaller tiles, resizing the tiles, and then merging the tiles together.);
quality-assuring the plurality of tiles by removing tiles that contain pen ink and tiles that are out of focus, thereby producing quality-assured tiles (See Fig. 3 pen markings 312 serve as pen ink (P0027, P0030) when removing the artifacts. Also, see Fig. 5-6, P0022, P0025-P0029 and interface for the digital pathology (DP) slide quality control in P0076-P0077.);
embedding the quality-assured tiles to produce embedded tiles (See P0022-P0023 where embedding additional metrics include color histograms, brightness, contrast, edge detectors, smoothness detectors and supervised classifiers such as pen detection classifier to identify the DP slides with presentation differences.);
passing the embedded tiles of a test subject (Taught in P0022-P0023.),
wherein each embedded tile is treated as an instance of a bag containing all quality- assured tiles of an image of a given histopathology slide, wherein the electronic neural network aggregates the instances of the bag using an attention mechanism (See Fig. 4, P0021-P0023 where slide clusters serve as a bag and embedding additional metrics.), and
to output a continuous concordance score in a range of 0 to 1 (See Taught as binary values in P0042-P0044, P0049, P0081, P0091 and P0076.).
Therefore, it would have been obvious to one of ordinary skill in the art of digital pathology slides before the effective filing date of the claimed invention to modify the method and system of Cohen to include initiating steps that include receiving an image of a histopathology slide, where a size of the image exceeds a volatile memory capacity of a computer system that comprises an electronic neural network, segmenting the image of the histopathology slide into tiles, quality-assuring the tiles by removing tiles that contain pen ink and tiles that are out of focus, by producing quality-assured tiles, embedding the quality-assured tiles to produce embedded tiles and 0 to 1 range concordance score as taught by Madabhushi to identify DP slides which need to be reproduced or repaired, or regions in the DP slide which should be avoided during computational analysis, as an advantage in clinical diagnostic workflows and computational pathology mentioned on Madabhushi’s P0003.
Regarding claims 2 and 25, Cohen discloses wherein the plurality of experts comprises human experts, machine experts, or a combination of human and machine experts (See P0007 P0043 and P0226-P0228 assisting human experts, machine learning and the AI perform reasoning and analysis in order to make decisions.).
Regarding claim 4, Cohen discloses the method of claim 1, comprising labeling the test subject medical data set with the medical determination prior to or when passing the test subject medical data set through the electronic neural network (See P0130, P0179-P0181 label giving context when screening. Also, see [P0268-P0269] NN12 is trained to produce the following outputs, as shown at block 180, including patient risk scores (e.g., an individual patient's % risk in a given cohort, margin of error, size of cohort, labels of cohort, etc.).).
Regarding claims 5 and 30, Cohen discloses the method of claim 1 and the system of claim 24, wherein the medical determination comprises a diagnosis or a prognosis of, or a recommended treatment plan for, a disease, condition, or disorder (See cancer diagnosis, recommended diagnosis (DX), treatment success factor, success rates and outcomes mentioned in P0267, P0269-P0270.).
Regarding claim 12, Cohen discloses the method of claim 1, wherein the medical determination comprises a therapy response (Besides treatment success factors in P0269-P0270, see known monitoring response to therapy in P0011.).
Regarding claim 53, although Cohen discloses the method of claim 1 mentioned above, Cohen does not explicitly teach the plurality of tiles comprises 128 pixel by 128 pixel tiles extracted at an objective power of 10x. Madabhushi teaches wherein the plurality of tiles comprises 128 pixel by 128 pixel tiles extracted at an objective power of 10x (Taught as pixel-level metrics in P0021, P0040-P0042 and extracting according to window-level approaches in P0066-P0067.).
Therefore, it would have been obvious to one of ordinary skill in the art of digital pathology slides before the effective filing date of the claimed invention to modify the method and system of Cohen to include the plurality of tiles comprises 128 pixel by 128 pixel tiles extracted at an objective power of 10x as taught by Madabhushi to identify DP slides which need to be reproduced or repaired, or regions in the DP slide which should be avoided during computational analysis, as an advantage in clinical diagnostic workflows and computational pathology mentioned on Madabhushi’s P0003.
Regarding claim 54, although Cohen discloses the method of claim 1 mentioned above, Cohen does not explicitly teach setting a threshold on a variance of a Laplacian computed over each tile. Madabhushi teaches wherein removing tiles that are out of focus comprises setting a threshold on a variance of a Laplacian computed over each tile, and wherein tiles containing pen ink are identified by a model trained to detect inked slides and are removed from training, validation, and inference (See PenMarking module and pixel bases in [P0052-P0053] The identifyBlurryRegions operation uses a Laplace matrix to determine which regions in the DP image are likely blurry. See Fig. 3 pen markings 312 serve as pen ink (P0027, P0030) when removing the artifacts. Also, see Fig. 5-6, P0022, P0025-P0029 and interface for the digital pathology (DP) slide quality control in P0076-P0077.).
Therefore, it would have been obvious to one of ordinary skill in the art of digital pathology slides before the effective filing date of the claimed invention to modify the method and system of Cohen to include setting a threshold on a variance of a Laplacian computed over each tile as taught by Madabhushi to identify DP slides which need to be reproduced or repaired, or regions in the DP slide which should be avoided during computational analysis, as an advantage in clinical diagnostic workflows and computational pathology mentioned on Madabhushi’s P0003.
Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1) further in view of McClelland (US 2011/0236903 A1).
Regarding claim 11, although Cohen and Madabhushi teach the method of claim 1 mentioned above, Cohen and Madabhushi do not explicitly teach determining a survival quantification. McClelland teaches wherein the medical determination comprises a survival quantification (See post-diagnosis survival of the subject, length of survival, recognizing aggressiveness in P0086-P0087, prostate associated genes (P0206) and disease-free survival time (P0309) as quantifying data.).
Therefore, it would have been obvious to one of ordinary skill in the art of diagnosing and prognosis of prostate cancer before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi to include determining a survival quantification as taught by McClelland to help identify patients having prostate cancer by analyzing gene cell expression levels mentioned on McClelland’s P0009.
Claims 15, 41 and 43 are rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1) further in view of Hartman (WO 2023/052367 A1).
Regarding claim 15, although Cohen and Madabhushi teach the method of claim 1 mentioned above, Cohen and Madabhushi do not explicitly teach when a histological stain of a sample is obtained from the test subject. Hartman teaches wherein the one or more medical tests comprise at least one histological stain of a sample obtained from the test subject (See histological stain of a sample on page 7, line 25 to page 8, line 5.).
Therefore, it would have been obvious to one of ordinary skill in the art of neural network training to predict cancer before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi for when the histological stain of a sample is obtained from the test subject as taught by Hartman for classifying patient that belongs to the intermediate cancer progress risk group mentioned on Hartman’s page 2, lines 1-3.
Regarding claim 41, although Cohen and Madabhushi teach the system of claim 24 mentioned above, Cohen and Madabhushi do not explicitly teach images of histopathology slides. Hartman teaches wherein the test and reference subject medical data sets comprise images of histopathology slides (See histological stain of a sample on page 7, line 25 to page 8, line 5.).
Therefore, it would have been obvious to one of ordinary skill in the art of neural network training to predict cancer before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi to include images of histopathology slides as taught by Hartman for classifying patient that belongs to the intermediate cancer progress risk group mentioned on Hartman’s page 2, lines 1-3.
Regarding claim 43, although Cohen and Madabhushi teach the system of claim 24 mentioned above, Cohen and Madabhushi do not explicitly teach a microscopy image. Hartman teaches wherein the test and reference subject medical data sets comprise images selected from the group consisting of: a magnetic resonance (MR) image, a computed tomography (CT) image, a single photon emission computed tomography (SPECT) image, a positron emission tomography (PET) image, and a microscopy image (See glass slides in microscope on page 7, line 25 to page 8, line 5, page 10, lines 4-14.).
Therefore, it would have been obvious to one of ordinary skill in the art of neural network training to predict cancer before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi to include a microscopy image as taught by Hartman for classifying patient that belongs to the intermediate cancer progress risk group mentioned on Hartman’s page 2, lines 1-3.
Claims 17 and 40 are rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1) further in view of Giftakis (US 2020/0139129 A1).
Regarding claims 17 and 40, although Cohen and Madabhushi teach the method of claim 1 and system of claim 24 mentioned above, Cohen and Madabhushi Cohen do not explicitly teach generating or updating at a medical report for the test subject when the concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value. Giftakis teaches further comprising generating or updating at least a portion of a medical report for the test subject when the concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value (See Fig. 17B where seizure severity score is reported as mentioned in P0219-P0220. Also, see Fig. 5 event summery page as reporting, Fig. 7, P0157-P0158 where the event score is greater than or equal to the threshold before adjusting therapy.).
Therefore, it would have been obvious to one of ordinary skill in the art of therapy adjustment before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi to include generating or updating at a medical report for the test subject when the concordance score of the medical determination being indicated by the test subject medical data set varies from a predetermined threshold value as taught by Giftakis in order to find the most effective therapy parameters to treat a condition mentioned in Giftakis’ P0005.
Claims 47-50 are rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1) further in view of Aneja (US 2015/0346191 A1).
Regarding claims 47 and 49, although Cohen and Madabhushi teach the method of claim 1 and the system of claim 24 mentioned above, Cohen and Madabhushi Cohen do not explicitly teach when the medical determination comprises an in situ melanoma status or an invasive melanoma status. Aneja teaches wherein the medical determination comprises an in situ melanoma status or an invasive melanoma status (See kinds of melanomas in P0043 and P0048.).
Therefore, it would have been obvious to one of ordinary skill in the art of prognosis and treatment of neoplasm before the effective filing date of the claimed invention to modify the system, software and method of Cohen and Madabhushi when the medical determination comprises an in situ melanoma status or an invasive melanoma status as taught by Aneja for predicting clinical outcomes, selecting cancer therapies, and assessing a cancer patient's response to a cancer therapy mentioned in Aneja’s P0003.
Regarding claims 48 and 50, although Cohen and Madabhushi teach the method of claim 1 and the system of claim 24 mentioned above, Cohen and Madabhushi Cohen do not explicitly teach whole slide images (WSJ). Aneja teaches wherein the images comprise whole slide images (WSJ) (See whole slide images P0062 and [P0107] Images could either be obtained from 10-15 microscopic fields of view for each sample or by whole-slide imaging as long as optical sections are acquired for 3D volume rendering. For slides stained immunofluorescently for centrosomes, imaging is carried out in areas determined to be “tumor areas” based on comparison with a serial section stained with hematoxylin eosin (wherein tumor areas are pre-marked).).
Therefore, it would have been obvious to one of ordinary skill in the art of prognosis and treatment of neoplasm before the effective filing date of the claimed invention to modify the system, software and method of Cohen and Madabhushi the whole slide images (WSJ) as taught by Aneja for predicting clinical outcomes, selecting cancer therapies, and assessing a cancer patient's response to a cancer therapy mentioned in Aneja’s P0003.
Claim 55 is rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1) further in view of Pre-trained vs. Random Weights for Calculating Fréchet Inception Distance in Medical Imaging by Jamie A. O'Reilly and Fawad Asadi at the 2021 Biomedical Engineering International Conference (BMEiCON-2021) and Azizi (US 2023/02600652 A1).
Regarding claim 55, although Cohen and Madabhushi teach the method of claim 1 mentioned above, Cohen does not explicitly teach propagating each quality-assured tile through an embedder trained in a SimCLR framework with a ResNet50 backbone to produce a 2048-dimension feature vector.
Combined BMEiCON-2021 and Azizi teach:
wherein embedding the quality-assured tiles comprises propagating each quality-assured tile through an embedder trained in a SimCLR framework with a ResNet50 backbone to produce a 2048-dimension feature vector for each quality- assured tile (See Azizi’s SimCLR framework with a ResNet50 backbone in P0034-P0035. See BMEiCON-2021’s produced 2048-dimension feature vector in Abstract, page 1, col. 2, lines 3-12, page 2, col. 1, lines 28-48 and page 4, col. 1, lines 1-11.).
Therefore, it would have been obvious to one of ordinary skill in the art of assessing medical imaging before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi to include propagating each quality-assured tile through an embedder trained in a SimCLR framework with a ResNet50 backbone to produce a 2048-dimension feature vector as taught by combined BMEiCON-2021 and Azizi when determining preferrable evaluations and analysis of medical images.
Claim 56 is rejected under 35 U.S.C. 103 as being unpatentable over Cohen (US 2021/0256323 A1) in view of Madabhushi (US 2019/0266726 A1) further in view of Tran (US 2023/0089026 A1).
Regarding claim 56, although Cohen and Madabhushi teach the method of claim 1 mentioned above, and Cohen teaches wherein the attention mechanism comprises sigmoid- activated attention heads (See sigmoidal normalization in P0064.), Cohen and Madabhushi do not explicitly teach using a root- mean-squared-error (RMSE) regression. Tran teaches:
wherein the electronic neural network is trained using a root- mean-squared-error (RMSE) regression objective rather than a classification objective (See root mean squared error (RMSE) in P0376-P0378, P0462-P0463.).
Therefore, it would have been obvious to one of ordinary skill in the art of medical image analysis before the effective filing date of the claimed invention to modify the method of Cohen and Madabhushi to using a root- mean-squared-error (RMSE) regression as taught by Tran when predicting medical findings when using machine learning techniques mentioned in Tran’s P0002.
Response to Arguments
Applicant’s arguments have been fully considered, but are now moot in view of the new grounds of rejection. The amendments to claims 1, 24 and newly added claims 53-56 necessitated further consideration and prior art search. The Examiner has entered a new rejection under 35 USC § 103 by applying new art and art already of record.
Applicant argues that the claims are neither directed to organizing human activity nor to a disembodied mathematical concept. See pgs. 10-11 of Remarks – Examiner disagrees.
Ordering medical tests for, administering therapy to, discontinuing administering therapy to a subject, determining a typical status, a benign status or malignant status from viewing an image of a histopathology slide are ways of providing medical services to a patient, a part of a healthcare workflow, diagnosing, treating, which are all human interactions between people, in a healthcare environment. Outputting a continuous concordance score in a range of 0 to 1 is a mathematical concept. Furthermore, the Laplacian variance of an image, which is assigning a singular numerical value according to deviating intensity of an image’s sharpness or blurriness, embodies a mathematical concept and could be observations by humans.
Applicant argues that the claims integrate into a practical application, improves computer technology McRO-style/digital pathology image processing, treatment and diagnostic application. See pgs. 12-14 of Remarks – Examiner disagrees.
The limitations as a whole “wherein a size of the image exceeds a volatile memory capacity of a computer system that comprises an electronic neural network: segmenting the image of the histopathology slide into a plurality of tiles; quality-assuring the plurality of tiles by removing tiles that contain pen ink and tiles that are out of focus, thereby producing quality-assured tiles”, is sorting and storing images, which is still abstract steps, and performed using generic hardware, see paragraph 50 in the Applicant’s specification.
McRO-style specific rules that improve a technological process. In McRO the claims recited a specific set of rules (evaluating subsequences of phonemes against morph-weight sets) that produced more realistic computer-generated lip synchronization.
The claimed invention in Enfish creates a specific type of “self-referential table” that provided an improvement to the way computers operate, thus providing a “practical application.” Here there is no specific type of data structure described. The claim is merely assigning concordance scores and linking neural network outputs to clinical actions that could be achieved with the generic computer manipulating observed data.
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 TERESA S WILLIAMS whose telephone number is (571)270-5509. The examiner can normally be reached Mon-Fri, 8:30 am -6:30 pm.
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, Mamon Obeid can be reached at (571) 270-1813. 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.
/T.S.W./Examiner, Art Unit 3687 09/26/2026
/Anita Y Coupe/Supervisory Patent Examiner, Art Unit 3619
