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 .
Continued Examination Under 37 CFR 1.114
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 12/8/2025 has been entered.
Status of the Claims
The status of the claims as of the response filed 12/8/2025 is as follows: Claims 1-7 are currently amended and have been considered below.
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 9/22/2025 is in compliance with the provisions of 37 CFR 1.97 and is being considered by the examiner.
Response to Amendment
Rejection Under 35 USC 101
The claims have been amended but the 35 USC 101 rejections for claims 1-7 are upheld.
Rejection Under 35 USC 103
The amendments made to the claims introduce limitations that are not fully addressed in the previous Office action, and thus the corresponding 35 USC 103 rejections are withdrawn. However, Examiner will consider the amended claims in light of an updated prior art search and address their patentability with respect to prior art below.
Response to Arguments
Rejection Under 35 USC 101
On pages 9-10 of the response filed 12/8/2025 Applicant argues that the amended claims do not fall into the certain methods of organizing human activity grouping of abstract idea, at least because they do not “recite any of the types of activities recognized in the MPEP as managing personal behavior, relationships, or interactions between people,” such as “storing pre-set limits on spending, filtering content, considering historical usage information while inputting data, testing a patient for nervous system malfunctions, voting, providing information to a person while avoiding interruption of their current activity, playing a dice game, assigning hair designs to balance head shape, or hedging risk.” Applicant further asserts that because the claims now specify that the model is learned by splitting a learning data set into a teacher data set and a test data set, training the model based on the teacher data set, and determining a receiver operating characteristic curve and an area under the curve with the test data set, the claims are not abstract.
Applicant’s arguments are fully considered, but are not persuasive. Examiner notes that the examples in MPEP 2106.04(a)(2)(II)(C) are not exhaustive, and are provided to give guidance about the types of human activities that are considered abstract. Though the instant claims do not match the exact fact patterns of any of the listed examples, Examiner maintains that the recited steps of inputting various parameters to a model that has been learned via a teacher data set and test data set with evaluation of performance metrics including a receiver operating characteristic curve and an area under the curve, and outputting from the model a determination of a probability of a given outcome, can be achieved by a human actor managing their personal behavior to fit a predictive model to a teacher set of data, test/validate the model on a test data set by calculating mathematical performance metrics of the model, and input new data to the model to obtain target outputs. Examiner notes that such steps can also be considered to fit into the “mathematical concepts” category of abstract idea because they describe how a predictive model (i.e. a mathematical representation) is fitted to a portion of a training dataset, validated by calculating performance metrics using another portion of a training dataset, and executed by providing inputs and calculating outputs using the learned mathematical relationships between input parameters. Though the claims specify that the models are machine learning models trained via machine learning processes, the underlying functions of fitting, validating, and executing a predictive model are abstract, and the machine learning nature of the models and training processes are addressed as an additional element under steps 2A – Prong 2 and 2B of the eligibility analysis. Accordingly, Examiner maintains that the claims as presently drafted recite an abstract idea under Step 2A – Prong 1.
On pages 11-13 of the response Applicant argues that the claims integrate any recited abstract idea into a practical application, “at least because the claims include specific recitations that place meaningful limits on the alleged judicial exception,” pointing to the newly-added limitations describing how the machine learning model is trained by splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve. Applicant also contends that the previous Office action “appears to have ignored the 2019 Guidance […] which requires that even conventional elements be considered in determining whether an abstract idea has been integrated into a practical application” (emphasis original).
Applicant’s arguments are fully considered, but are not persuasive. Examiner notes that every additional element beyond the abstract idea itself, including those that may be considered ‘conventional’ like an information processing device comprising a determination unit and use of a machine learning process to train the model, were considered in light of the 2019 Guidance in the Step 2A – Prong 2 eligibility analysis, and were found not to provide integration into a practical application (see para. 15 of the final rejection mailed 8/6/2025). Examiner notes that the more specific details now recited in the claims describing how the model is trained are not additional elements beyond the abstract idea, because they are part of the abstract idea itself as explained above. Because these steps are part of the abstract idea itself, they cannot provide “significantly more” than the abstract idea and thus do not confer eligibility (see MPEP 2106.05(a): “It is important to note, the judicial exception alone cannot provide the improvement. The improvement can be provided by one or more additional elements.” See also 2106.05(a)(II): “it is important to keep in mind that an improvement in the abstract idea itself… is not an improvement in technology.”). Examiner notes that the only additional elements beyond the abstract idea itself are an information processing device comprising a determination unit, the fact that the model is a machine learning model, as well as use of a machine learning process to train the machine learning model. These additional elements, when considered in the context of each claim as a whole, amount to instructions to “apply” the abstract idea on a computer because the otherwise-abstract steps of fitting, validating, and executing a predictive model are merely specified as occurring on a generic computer processing device via high-level “machine learning” methods such that the otherwise-abstract steps are digitized and/or automated within a computing environment. See MPEP 2106.05(f). Such additional elements do not impose meaningful limits on the judicial exception, instead merely instructing the judicial exception to occur via machine learning in a computer rather than via mathematical processes that a human actor could otherwise achieve.
On pages 13-14 of the response Applicant argues that “each of [the] claims includes one or more ‘additional features’ that are not well-understood, routine, or conventional” (emphasis original) and appears to argue that the prior art fails to teach the technical features of claims 1-7 and thus they are not well-understood, routine, and conventional. Applicant’s arguments are fully considered, but are not persuasive. Examiner notes that issues of patentability over the prior art are a separate consideration to the question of eligibility under 35 USC 101; MPEP 2106.05(I) states that:
Although the courts often evaluate considerations such as the conventionality of an additional element in the eligibility analysis, the search for an inventive concept should not be confused with a novelty or non-obviousness determination. See Mayo, 566 U.S. at 91, 101 USPQ2d at 1973 (rejecting "the Government’s invitation to substitute §§ 102, 103, and 112 inquiries for the better established inquiry under § 101 "). As made clear by the courts, the "‘novelty’ of any element or steps in a process, or even of the process itself, is of no relevance in determining whether the subject matter of a claim falls within the § 101 categories of possibly patentable subject matter." Intellectual Ventures I v. Symantec Corp., 838 F.3d 1307, 1315, 120 USPQ2d 1353, 1358 (Fed. Cir. 2016) (quoting Diamond v. Diehr, 450 U.S. at 188–89, 209 USPQ at 9).
Accordingly, whether the claims are found to be novel and/or non-obvious over the prior art has no bearing on analysis of patent eligibility under 35 USC 101.
Further, Applicant has not provided any evidence that the combination of additional elements recited in claims 1-7 is unconventional. The combination of an information processing device comprising a determination unit used to perform the inputting and outputting steps of the invention as well as use of a machine learning process to train the machine learning model amount to mere instructions to “apply” the exception using generic computer components. As evidence of the generic nature of the above recited additional elements, Examiner notes Fig. 1 and paras. [0034]-[0045] of Applicant’s specification, which describe the information processing device as including many typical computer components like one or more processors, a communication interface, an input/output interface, memory/storage, etc. with the ability to execute instructions to achieve the functions of the invention. These disclosures do not indicate that the elements of the invention are particular machines, and instead provide generic examples of computer hardware, such that one of ordinary skill in the art would understand that any generic information processing device could be used to implement the invention. Examiner further notes para. [0050] of Applicant’s disclosure, which provides examples of the machine learning processing being “supervised learning and semi-supervised learning,” but provides no additional detail about what type of models are used, what specific supervised or semi-supervised learning techniques are utilized, or any other details about the specific machine learning training techniques or machine learning model architectures, leaving one of ordinary skill in the art to understand that any known methods of supervised or semi-supervised machine learning may be utilized to achieve the training of any compatible type of machine learning model. Further, Examiner notes that it is well-understood, routine, and conventional to utilize an information processing device with machine learning capabilities to train, validate, and execute diagnostic models to predict patient health conditions, as evidenced by at least Van Hooser et al. (US 20240257973 A1) [0004], [0092]-[0099], & [0104]; McGovern et al. (US 20200342958 A1) [0113], [0182], [0206], & [0251]; Taylor (Reference V on the accompanying PTO-892) Pgs 5977, 5979, & 5981; and Rehman (Reference W on the accompanying PTO-892) Pgs 9444-9445.
On page 15 of the response Applicant discusses Ex Parte Desjardins as avoiding evaluating claims at a high level of generality and “respectfully requests that such directive be followed” in evaluating eligibility of the instant claims. Applicant’s request is fully considered, and Examiner submits that the instant claims have been evaluated for eligibility in line with applicable patent rules, guidance, and precedential cases. While Applicant has not made any specific analogies between the instant claims and those found eligible in Desjardins (beyond characterizing the eligible claims of Desjardins as being “directed to training a machine learning model”), Examiner notes that the fact patterns of each case are different. In Desjardins, the claims reflected an improvement to how a machine learning model itself is trained and operates to address the technical problem of ‘catastrophic forgetting’ encountered in continual learning systems, which was identified and explained as a technical problem in the specification. In contrast, the instant specification does not outline a specific technical problem in machine learning technology whose solution is reflected in the claims. As noted in paras. [0002]-[0006] of the specification, the invention appears to seek to solve a problem rooted in the drawbacks of conventional business practices of thyroid condition diagnosis (i.e. a lack of consideration of certain types of data in the diagnostic process) rather than a technical problem rooted in computer technology, machine learning, or another technical field. The use of high-level machine learning training and modelling techniques to digitize and/or automate the consideration of different types of parameters for performing thyroid-related diagnostic processes does not provide a technical improvement to a technical problem as in Desjardins, and instead amount to instructions to “apply” the exception with a computer as explained above.
For the reasons outlined above, the 35 USC 101 rejections are upheld for claims 1-7.
Rejection Under 35 USC 103
On pages 16-20 of the response Applicant argues that the cited art of record fails to teach the newly-added subject matter directed to how each model is trained. Applicant’s arguments are fully considered but are moot because the new grounds of rejection do not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
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 1-7 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1
In the instant case, claims 1-7 are directed to information processing devices (i.e. machines), such that each of the claims falls within one of the four statutory categories. Nevertheless, the claims fall within the judicial exception of an abstract idea.
Step 2A – Prong 1
Independent claims 1-7 each recite steps that, under their broadest reasonable interpretations, cover certain methods of organizing human activity, e.g. managing personal behavior, relationships, or interactions between people, as well as mathematical concepts. Specifically, claim 1 (as representative) recites: An information processing device comprising a determination unit that inputs parameters that represent total protein (TP), cholinesterase (ChE), total cholesterol (TC), creatinine (CREA), and creatine phosphokinase (CPK) from a hematological examination of a subject to a learning-completed machine learning model that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve, executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that thyroid stimulation hormone (TSH) of the subject is in a range for which medical treatment is needed.
But for the recitation of generic computer components like an information processing device comprising a determination unit, and use of a high-level machine learning process, and specifying that the model is a machine learning model, the italicized functions, when considered as a whole, describe an operation for fitting, validating, and executing a clinical diagnostic that a clinician or other medical professional could undertake by managing their personal behavior. For example, a clinician could obtain a training dataset with various input parameters and split it into a teacher data set representing changes in the input parameters over time as well as a test data set, train/fit a diagnostic model (e.g. a decision tree, regression equation, list of questions with a scoring ruleset, etc.) using the teacher dataset, calculate performance metrics like ROC and AUC using the test dataset, and new input information related to different blood examination measurements of a patient into the fitted model to receive a probability of a given diagnostic output, e.g. whether a subject’s TSH levels are in a certain range. Examiner notes that such steps can also be considered to fit into the “mathematical concepts” category of abstract idea because they describe how a predictive model (i.e. a mathematical representation) is fitted to a portion of a training dataset, validated by calculating performance metrics using another portion of a training dataset, and executed by providing inputs and calculating outputs using the learned mathematical relationships between input parameters. Accordingly, claim 1 recites an abstract idea in the form of a certain method of organizing human activity and/or mathematical concepts. Claims 2-7 recite substantially similar subject matter and functions as claim 1, while merely specifying different information types to be input to the model and different types of determinations to be output from the model (each of which are types of data that a clinician would be capable of evaluating and determining via mathematical processes) and are thus found to recite an abstract idea under the same analysis.
However, recitation of an abstract idea is not the end of the analysis. Each of the claims must be analyzed for additional elements that indicate the abstract idea is integrated into a practical application to determine whether the claim is considered to be “directed to” an abstract idea.
Step 2A – Prong 2
The judicial exception is not integrated into a practical application. In particular, independent claims 1-7 do not include additional elements that integrate the abstract idea into a practical application. The additional elements of claims 1-7 include an information processing device comprising a determination unit, specifying that the learning-completed model is a machine learning model, as well as use of a machine learning process to train the machine learning model. These additional elements, when considered in the context of each claim as a whole, amount to instructions to “apply” the abstract idea on a computer because the otherwise-abstract steps of fitting, validating, and executing a predictive model are merely specified as occurring on a generic computer processing device via high-level “machine learning” methods such that the otherwise-abstract steps are digitized and/or automated within a computing environment. See MPEP 2106.05(f). Such additional elements do not impose meaningful limits on the judicial exception, instead merely instructing the judicial exception to occur via machine learning in a computer rather than via mathematical processes that a human actor could otherwise achieve. Accordingly, the additional elements of claims 1-7 do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Claims 1-7 are directed to an abstract idea.
Step 2B
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 integration of the abstract idea into a practical application, the additional elements of an information processing device comprising a determination unit used to perform the inputting and outputting steps of the invention as well as use of a machine learning process to train the machine learning model amount to mere instructions to “apply” the exception using generic computer components. As evidence of the generic nature of the above recited additional elements, Examiner notes Fig. 1 and paras. [0034]-[0045] of Applicant’s specification, which describe the information processing device as including many typical computer components like one or more processors, a communication interface, an input/output interface, memory/storage, etc. with the ability to execute instructions to achieve the functions of the invention. These disclosures do not indicate that the elements of the invention are particular machines, and instead provide generic examples of computer hardware, such that one of ordinary skill in the art would understand that any generic information processing device could be used to implement the invention. Examiner further notes para. [0050] of Applicant’s disclosure, which provides examples of the machine learning processing being “supervised learning and semi-supervised learning,” but provides no additional detail about what type of models are used, what specific supervised or semi-supervised learning techniques are utilized, or any other details about the machine learning process or machine learning model architectures, leaving one of ordinary skill in the art to understand that any known methods of supervised or semi-supervised machine learning may be utilized to achieve the training of any compatible type of machine learning model.
Further, Examiner notes that it is well-understood, routine, and conventional to utilize an information processing device with machine learning capabilities to train, validate, and execute diagnostic models to predict patient health conditions, as evidenced by at least Van Hooser et al. (US 20240257973 A1) [0004], [0092]-[0099], & [0104]; McGovern et al. (US 20200342958 A1) [0113], [0182], [0206], & [0251]; Taylor (Reference V on the accompanying PTO-892) Pgs 5977, 5979, & 5981; and Rehman (Reference W on the accompanying PTO-892) Pgs 9444-9445.
Analyzing these additional elements as an ordered combination adds nothing that is not already present when considering the elements individually; the overall effect of the computer implementation and high-level machine learning training and modeling is to digitize and/or automate an operation for fitting, validating, and executing a clinical diagnostic model that could otherwise be achieved as a certain method of organizing human activity and/or mathematical concept. Thus, when considered as a whole and in combination, claims 1-7 are not patent eligible.
Claim Rejections - 35 USC § 103
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
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 and 6 are rejected under 35 U.S.C. 103 as being unpatentable over Van Hooser et al. (US 20240257973 A1) in view of Katsuta et al. (JP 2007147494 A).
Claim 1
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent (Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine and cholesterol, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that thyroid stimulation hormone (TSH) of the subject is in a range for which medical treatment is needed (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism (i.e. a condition in which a subject’s level of TSH is typically higher than normal, indicating that medical treatment is likely needed), Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism (i.e. a condition in which a subject’s level of TSH is typically lower than normal, indicating that medical treatment is likely needed), subacute thyroiditis, etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as hormone imbalance or hyper/hypothyroidism (i.e. conditions that indicate a subject’s level of TSH is in a lower than normal or higher than normal range for which medical treatment is likely needed), it fails to explicitly disclose the specific parameters of total protein, cholinesterase, and creatinine phosphokinase as the inputs. However, Katsuta teaches that metabolic or endocrine conditions can be evaluated with a variety of disease-related biomarkers, including substances measured in general medical examinations such as total protein, cholinesterase, total cholesterol, creatinine, and CPK (Katsuta Pg 11 paragraph beginning “Any substance can be used…”, Pg 12 paragraph beginning “More specifically, in addition to…”). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood specimen inputs of Van Hooser to specifically include TP, ChE, TC, CREA, and CPK as in Katsuta because Katsuta shows that such test items are obtained during general, routine medical examinations and may be used and evaluated as biomarkers of patient diseases, including metabolic or endocrine conditions (i.e. including thyroid conditions), such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Claim 6
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent total cholesterol (TC), (Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine and cholesterol, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that the subject is affected with painless thyroiditis (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism, Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism, subacute thyroiditis (i.e. painless thyroiditis), etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as subacute thyroiditis (i.e. painless thyroiditis), it fails to explicitly disclose the specific parameters of cholinesterase and creatinine phosphokinase as the inputs. However, Katsuta teaches that diseases like metabolic or endocrine conditions can be evaluated with a variety of disease-related biomarkers, including substances measured in general medical examinations such as total cholesterol, cholinesterase, creatinine, and CPK (Katsuta Pg 11 paragraph beginning “Any substance can be used…”, Pg 12 paragraph beginning “More specifically, in addition to…”). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood specimen inputs of Van Hooser to specifically include TC, ChE, CREA, and CPK as in Katsuta because Katsuta shows that such test items are obtained during general, routine medical examinations and may be used and evaluated as biomarkers of patient diseases, including metabolic or endocrine conditions (i.e. including thyroid conditions), such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Claims 2-3 and 5 are rejected under 35 U.S.C. 103 as being unpatentable over Van Hooser in view of Karlov et al. (US 7392199 B2).
Claim 2
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent (Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine and bilirubin, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that thyroid stimulation hormone (TSH) of the subject is in a normal value range (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism (i.e. a condition in which a subject’s level of TSH is typically higher than normal, indicating that medical treatment is likely needed), Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism (i.e. a condition in which a subject’s level of TSH is typically lower than normal, indicating that medical treatment is likely needed), subacute thyroiditis, etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as hormone imbalance or hyper/hypothyroidism (i.e. conditions that indicate a subject’s level of TSH is in a lower than normal or higher than normal range for which medical treatment is likely needed), it fails to explicitly disclose the specific parameters of total protein and neutrophil as the inputs. However, Karlov teaches that thyroid health conditions can be evaluated with a variety of disease-related biomarkers, including substances measured in standard blood analyses such as total protein, creatinine, neutrophils, and total bilirubin (Karlov Col6 L59 – Col7 L66). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood specimen inputs of Van Hooser to specifically include TP, CREA, Neu, and T-bill as in Karlov because Karlov shows that such parameters are obtained during standard blood analyses and may be used and evaluated as biomarkers of patient diseases, including thyroid conditions, such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Claim 3
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent creatinine (CREA), total cholesterol (TC), (Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine and cholesterol, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that the subject is affected with Graves' disease (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism, Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism, subacute thyroiditis, etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as Grave’s disease, it fails to explicitly disclose the specific parameters of alkaline phosphatase and total protein as the inputs. However, Karlov teaches that thyroid health conditions can be evaluated with a variety of disease-related biomarkers, including substances measured in standard blood analyses such as creatinine, total cholesterol, alkaline phosphatase, and total protein (Karlov Col6 L59 – Col7 L66). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood specimen inputs of Van Hooser to specifically include CREA, TC, ALP, and TP as in Karlov because Karlov shows that such parameters are obtained during standard blood analyses and may be used and evaluated as biomarkers of patient diseases, including thyroid conditions, such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Claim 5
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent (Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that the subject is affected with Graves' disease or painless thyroiditis (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism, Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism, subacute thyroiditis (i.e. painless thyroiditis), etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as Grave’s disease or subacute thyroiditis (i.e. painless thyroiditis), it fails to explicitly disclose the specific parameters of alkaline phosphatase, total protein, y glutamyl transpeptidase, and white blood cells as the inputs. However, Karlov teaches that thyroid health conditions can be evaluated with a variety of disease-related biomarkers, including substances measured in standard blood analyses such as alkaline phosphatase, creatinine, total protein, gamma glutamyl tran (i.e. y glutamyl transpeptidase), and leukocytes (i.e. white blood cells) (Karlov Col6 L59 – Col7 L66). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood test measurement inputs of Van Hooser to specifically include ALP, CREA, TP, yGPT, and WBC as in Karlov because Karlov shows that such parameters are obtained during standard blood analyses and may be used and evaluated as biomarkers of patient diseases, including thyroid conditions, such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Claim 4 is rejected under 35 U.S.C. 103 as being unpatentable over Van Hooser in view of Dodds (US 20110093298 A1).
Claim 4
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine and cholesterol, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that the subject is affected with Graves' disease or painless thyroiditis (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism, Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism, subacute thyroiditis (i.e. painless thyroiditis), etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as Grave’s disease or subacute thyroiditis (i.e. painless thyroiditis), it fails to explicitly disclose the specific parameters of thyroid hormone FT3, thyroid hormone FT4, FT3/FT4, and alkaline phosphatase as the inputs. However, Dodds teaches that thyroid health conditions can be evaluated with a variety of disease-related biomarkers, including substances measured in comprehensive diagnostic test panel such as a complete thyroid profile indicating FT3 and FT4 (and thus also representing a ratio of FT3/FT4), alkaline phosphatase, and creatinine (Dodds [0081]-[0085]). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood specimen inputs of Van Hooser to specifically include FT3, FT4, ALP and CREA as in Dodds because Dodds shows that such parameters are obtained during comprehensive diagnostic testing and may be used and evaluated as biomarkers of patient diseases, including thyroid conditions, such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Van Hooser in view of Saito (US 20190139440 A1).
Claim 7
Van Hooser teaches an information processing device comprising a determination unit (Van Hooser [0113], [0117], noting a computing device that implements software modules to perform the functions of the invention) that inputs parameters that represent creatinine (CREA), total cholesterol (TC), (Van Hooser, [0004], [0061]-[0065], noting a wide variety of biomarker data of a subject obtained from a biological specimen (e.g. a blood or plasma sample) is input to a trained machine learning model; non-limiting examples of such biomarkers include metabolites like creatinine and cholesterol, immune response protein panels, inflammation protein panels, and others as listed in [0065] and Table 2 between paras. [0063] & [0064]) that is trained through a machine learning process including splitting a learning data set into a teacher data set and a test data set, training the machine learning model based on the teacher data set in which at least a portion of the teacher data set indicates a change in one or more of the parameters from the hematological examination over time, and determining, with the test data set, a receiver operating characteristic curve and an area under the curve (Van Hooser [0097]-[0099], [0104], noting a training set of samples (i.e. a learning set) is used to train (i.e. teach) and test the machine learning models via cross-validation with different splits of the data, for example by calculating ROC and AUC metrics with a test set of data. See also [0098], noting that in some instances specific input metabolites (i.e. parameters from the hematological examination) that vary (i.e. change) over time are excluded from the training and/or testing of the models, indicating that in other embodiments such inputs representing changes in one or more of the inputs over time are left in for the training and/or testing of the model, such that the teacher subset of the training dataset is considered to include changes in one or more of the parameters from the hematological examination over time in at least some embodiments of the invention), executes the learning-completed machine learning model based on the parameters, and outputs a determination, by the learning-completed machine learning model, of a probability that the subject is affected with thyrotoxicosis (Van Hooser [0004], [0046]-[0048], noting the trained machine learning model is executed on the input data to output a classification or other indication of a disease or condition, such as a probability or likelihood that the user has a given health condition or is outside of a “normal” score range; many diseases and conditions are contemplated for prediction as exemplified in the non-limiting examples of Table 1 between paras. [0060] & [0061], including thyroid-related conditions like hypothyroidism, Grave’s disease, Hashimoto’s thyroiditis, hormonal imbalance, hyperthyroidism (i.e. a cause of thyrotoxicosis), subacute thyroiditis, etc.).
Though Van Hooser contemplates using a non-limited variety of biomarkers (e.g. those obtained from a blood specimen) as input to a trained machine learning model that outputs a probability of a given disease such as hormone imbalance or hyperthyroidism (i.e. conditions considered equivalent to thyrotoxicosis which occurs when there is too much thyroid hormone in a subject’s body), it fails to explicitly disclose the specific parameters of cholinesterase, creatine phosphokinase, and basophil as the inputs. However, Saito teaches that health conditions can be evaluated with a variety of lifestyle biomarkers, including substances measured in blood tests such as creatinine, total cholesterol, cholinesterase, creatine kinase, and basophil (Saito [0092], [0165]). It would have been obvious to one of ordinary skill in the art, before the effective filing date of the claimed invention, to modify the non-limited blood specimen inputs of Van Hooser to specifically include CREA, TC, ChE, CPK, and Ba as in Saito because Saito shows that such test items are commonly obtained and evaluated as biomarkers of patient health conditions, such that inclusion of these specific inputs to the model of Van Hooser would allow for expanded consideration of these known biomarkers, thereby improving the diagnostic accuracy of the model with respect to the contemplated thyroid conditions.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Zou et al. (Reference U on the accompanying PTO-892) describes ROC and AUC analysis of diagnostic test and predictive models in a medical context.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to KAREN A HRANEK whose telephone number is (571)272-1679. The examiner can normally be reached M-F 8:00-4:00 ET.
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, Shahid Merchant can be reached on 571-270-1360. 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.
/KAREN A HRANEK/ Primary Examiner, Art Unit 3684