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 .
Response to Arguments
Applicant’s arguments filed 06/22/2026 have been fully considered but are not persuasive or are moot in view of new grounds of rejection.
Applicant argues, “The claim further requires the system to output a patient- specific predicted conversion time from MCI to Alzheimer's Disease, track that prediction during drug administration, and transmit a threshold-triggered notification to a medical professional, thereby integrating the machine learning steps into a concrete method of managing treatment for patients with MCI and/or Alzheimer's Disease. Applicant respectfully asserts the amended claim recites steps that cannot be performed in the human mind and is directed to a practical application.”
Examiner respectfully disagrees. Claim 1 recites the following abstract ideas:
“determine a plurality of interaction indications of the patient, each interaction indication of the plurality of interaction indications being an indication of interaction between at least two risk factor indications of the plurality of risk factor indications;
each interaction indication of the plurality of interaction indications being an indication of interaction between at least two risk factor indications of the plurality of risk factor indications”,
which are directed to mental processes, since a person could receive a plurality of risk factor indications and determine a plurality of interaction indications, which may between at least two risk factor indications.
These limitations, under their broadest reasonable interpretation, cover concepts that can be practically performed in the human mind, i.e., using pen and paper (i.e. mental processes).
Applicant argues, “Amended claim 1 recites precisely such an improvement. The claim specifies a multivariate linear regression model trained by a three-step computational pipeline: (i) applying an ordered quantile normalization transformation to risk factor indications and interaction indications to generate normalized training data; (ii) applying a backward feature selection process to the normalized training data to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time; and (iii) iteratively adjusting a plurality of coefficients to minimize root mean square error (RMSE) between predicted and ground truth conversion times. Applicant respectfully submits that this is not generic application of machine learning to data. Rather, it is a specific, ordered training methodology that improves how the model itself operates…. Accordingly, the claimed training pipeline constitutes an improvement to the machine learning technology itself under MPEP § 2106.05(a), Enfish, and Desjardins.”
Examiner respectfully states that although amended claim 1 may provide an improvement, there is nothing in the claims which show how transmitting a notification to a medical professional associated with the patient integrates the judicial exception into a practical application. For instance, under broadest reasonable interpretation, a notification that prompts an appointment between the medical professional and the patient does not require a particular prophylaxis to be provided.
Applicant argues, “First (a), the claims are amended to require administering a particular course of treatment for patients with Mild Cognitive Impairment (MCI) who are undergoing drug therapy for Alzheimer's Disease. Amended claim 1 affirmatively requires that the system: (i) track a plurality
of predicted conversion times from MCI to Alzheimer's Disease "while the patient is administered a drug for Alzheimer's Disease treatment," including "determining whether the drug is increasing or decreasing the predicted conversion time in the patient"; and (ii) "transmit a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate, wherein the notification is configured to prompt an appointment between the medical professional and the patient." These limitations affirmatively recite the action that affects the treatment-tracking the patient's disease trajectory under active drug administration and triggering clinical intervention when that trajectory deteriorates beyond a defined threshold.”
Examiner respectfully disagrees. Although the claims are amended to require a drug to be administered to a patient, there is nothing in the claims which show how transmitting a notification to a medical professional associated with the patient integrates the judicial exception into a practical application.
Applicant argues, “the claimed treatment includes limitations that impose meaningful limits on the judicial exceptions. The step of tracking a plurality of predicted conversion times requires that the machine learning model's output be applied repeatedly over time during actual drug administration. The notification step further requires that the system's output be compared against two independent clinical thresholds-specifically, (1) whether the predicted conversion time falls below a predetermined period of time, and (2) whether the rate of change of the predicted conversion time exceeds a threshold change rate-and, upon satisfaction of either condition, actively communicate with the treating medical professional to prompt a clinical appointment. Taken together, these limitations reflect a practical method of treatment: the system continuously tracks a patient's disease trajectory during drug administration and, when a clinically defined threshold is crossed, directs the responsible physician to take action-actively participating in the patient's ongoing care through timely, threshold-conditioned alerts.”
Examiner respectfully disagrees. As stated above, although the claims are amended to require a drug to be administered to a patient, there is nothing in the claims which show how transmitting a notification to a medical professional associated with the patient integrates the judicial exception into a practical application. Further, Examiner asserts that under broadest reasonable interpretation, the notification that prompts an appointment between the medical professional and the patient does not require the responsible physician to actively participate in the patient's ongoing care through timely, threshold-conditioned alerts. Specifically, the broadest reasonable interpretation of “prompt” is to bring about or encourage an action, instead of requiring the action to occur.
Applicant argues, “Wang does not disclose ordered quantile normalization, backward feature selection, iterative RMSE-based coefficient adjustment, drug-efficacy tracking, or threshold-triggered notification to a medical professional. The Office's burden also extends to the ordered combination of elements, not just the individual elements - which, again, Wang does not teach. See BASCOM Global Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341, 1350 (Fed. Cir. 2016) (unconventional ordered combination of individually conventional elements may amount to significantly more).”
Examiner respectfully states that although Wang does not disclose the above recited limitations, the recited “ordered quantile normalization, backward feature selection, iterative RMSE-based coefficient adjustment, drug-efficacy tracking, or threshold-triggered notification to a medical profession” are done by the trained machine learning model and processor, which is disclosed by Wang, as evidenced below.
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 1-8, 10-11, and 13 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception, specifically an abstract idea without significantly more.
Step 1:
Independent claim 1 is directed to a system for cognitive disease prediction. Thus, it is directed to statutory categories of invention.
Step 2A, Prong 1:
Claim 1 recites the following claim limitations:
“determine a plurality of interaction indications of the patient, each interaction indication of the plurality of interaction indications being an indication of interaction between at least two risk factor indications of the plurality of risk factor indications;
each interaction indication of the plurality of interaction indications being an indication of interaction between at least two risk factor indications of the plurality of risk factor indications”,
which are directed to mental processes, since a person could receive a plurality of risk factor indications and determine a plurality of interaction indications, which may between at least two risk factor indications.
These limitations, under their broadest reasonable interpretation, cover concepts that can be practically performed in the human mind, i.e., using pen and paper (i.e. mental processes).
Step 2A, Prong 2:
Claim 1 recites the following additional elements:
a memory; and
a processor communicatively coupled to the memory;
wherein the memory stores a set of instructions which, when executed by the processor, cause the processor to:
receive a plurality of risk factor indications for a given patient;
transform, using an ordered quantile transformer, a first probability distribution of the plurality of risk factor indications and the plurality of interaction indications to a second probability distribution such that the plurality of risk factor indications and the plurality of interaction indications follow anormal distribution:
obtain a trained machine learning model,
wherein the trained machine learning model is a multivariate linear regression model trained by:
(i) applying an ordered quantile normalization transformation to risk factor indications and interaction indications of a plurality of training patients to generate normalized training data:
(ii) applying a backward feature selection process to the normalized training data to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time from a first cognitive disease state to a second cognitive disease state: and
(iii) iteratively adjusting a plurality of coefficients corresponding to the identified subset of risk factor indications and interaction indications to minimize a root mean square error (RMSE) between predicted conversion times and ground truth conversion time indications of the plurality of training patients;
apply the transformed plurality of risk factor indications and the transformed plurality of interaction indications to the trained machine learning model; and
output a predicted conversion time from Mild Cognitive Impairment to Alzheimer's Disease in the patient based on the trained machine learning model;
track a plurality of predicted conversion times from Mild Cognitive Impairment to Alzheimer's Disease in the patient while the patient is administered a drug for Alzheimer's Disease treatment, wherein tracking the plurality of predicted conversion times comprises determining whether the drug is increasing or decreasing the predicted conversion time in the patient; and
transmit a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate, wherein the notification is configured to prompt an appointment between the medical professional and the patient.
The following limitations: “wherein the memory stores a set of instructions which, when executed by the processor, cause the processor to:
receive a plurality of risk factor indications for a given patient;
transform, using an ordered quantile transformer, a first probability distribution of the plurality of risk factor indications and the plurality of interaction indications to a second probability distribution such that the plurality of risk factor indications and the plurality of interaction indications follow anormal distribution”,
are interpreted as insignificant extra solution activities. Specifically, the above recited limitations are directed towards pre-solution activity (see MPEP §2106.05(g)) since they’re used to obtain information about the user to provide an output regarding a prediction of conversion to the cognitive disease for the patient (i.e. mere data gathering).
The following limitations:
“obtain a trained machine learning model” and
wherein the trained machine learning model is a multivariate linear regression model trained by:
(i) applying an ordered quantile normalization transformation to risk factor indications and interaction indications of a plurality of training patients to generate normalized training data:
(ii) applying a backward feature selection process to the normalized training data to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time from a first cognitive disease state to a second cognitive disease state: and
(iii) iteratively adjusting a plurality of coefficients corresponding to the identified subset of risk factor indications and interaction indications to minimize a root mean square error (RMSE) between predicted conversion times and ground truth conversion time indications of the plurality of training patients;
apply the transformed plurality of risk factor indications and the transformed plurality of interaction indications to the trained machine learning model; and
output a predicted conversion time from Mild Cognitive Impairment to Alzheimer's Disease in the patient based on the trained machine learning model;
track a plurality of predicted conversion times from Mild Cognitive Impairment to Alzheimer's Disease in the patient while the patient is administered a drug for Alzheimer's Disease treatment, wherein tracking the plurality of predicted conversion times comprises determining whether the drug is increasing or decreasing the predicted conversion time in the patient; and
transmit a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate, wherein the notification is configured to prompt an appointment between the medical professional and the patient,”
are directed to additional elements, specifically insignificant post solution activity (see MPEP 2106.05(g)).
32. The above recited limitations merely process information and then output the results of the above identified abstract ideas. Additionally, the recited outputting is neither particular enough to meaningfully limit the recited exception nor does it have more than a nominal relationship to the exception. In other words, the breadth of the recited “output” is such that it substantially encompasses all applications of the recited exception (such as moving information around to output a prediction of conversion to the cognitive disease for the patient).
Further, there is nothing in the claims which show how transmitting a notification to a medical professional associated with the patient integrates the judicial exception into a practical application.
34. Further, there is no evidence of record that would support the assertion that this step is an improvement to a computer or a technological solution to a technological problem.
In other words, these claims are merely directed to an abstract idea with additional generic computer elements which do not add a meaningful limitation to the abstract idea because they amount to simply implementing the abstract idea on a computer.
Further, the limitations:
a memory; and
a processor communicatively coupled to the memory;
wherein the memory stores a set of instructions…
are directed toward generically recited computer elements which do not improve the functioning of a computer, or any other technology or technical field.
Accordingly, the combination of these additional elements is no more than insignificant extra solution activity. Thus, the abstract ideas are not integrated into a practical application.
Step 2B:
The claim does not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additionally, regarding the detecting device, Applicant’s specification discloses: “process 1100 can obtain a trained machine learning model. In some examples, the trained machine learning model can include a multivariate linear regression machine learning model. However, it should be appreciated that the machine learning model is not limited to the multivariate linear regression machine learning model. As one example, a machine learning model can be configured as a feedforward network, in which the connections between nodes do not form any loops in the network. As another example, a machine learning algorithm can be configured as a recurrent neural network ("RNN"), in which connections between nodes are configured to allow for previous outputs to be used as inputs while having one or more hidden states, which in some instances may be referred to as a memory of the RNN...” [0038].
Regarding the limitations directed to the trained machine learning mode, the memory, and the processor, see Wang et al. (US 2021/0110926), which discloses predicting outcomes using prediction models [0002] comprising of
a trained machine learning mode [0032] which may be a linear regression model having multiple variables [0032, 0043],
a memory [0061], and
a processor [0061].
Wang further teaches that traditionally, statistical correlations that generate predictions based on multiple variables use techniques such as linear regression [0004], and that a prediction model can be a linear regression model [0032].
Thus, the limitations directed to the trained machine learning mode, the memory, and the processor are well-understood, routine, and conventional, as evidenced by the reference above.
As discussed with respect to Step 2A, Prong 2 above, the additional elements in the claim amount to no more than insignificant extra solution activity and applying the exception in a general way, as well as establishing an environment for which data is gathered.
Moreover, implementing an abstract idea on a generic computer, does not add significantly more, similar to how the recitation of the computer in the claim in Alice amounted to mere instructions to apply the abstract idea of intermediated settlement on a generic computer.
Thus, none of the claims 1-8, 10-11, and 13 amount to significantly more than the abstract idea itself. Accordingly, claims 1-8, 10-11, and 13 are not patent eligible and rejected under 35 U.S.C. 101 as being directed to abstract ideas in view of the Supreme Court Decision in Alice Corporation Pty. Ltd. v. CLS Bank International, et al., MPEP §2106.04(a)(2), MPEP §2106.04(d)(2),and MPEP §2106.05(g).
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.
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.
Claim 1-8 are rejected under 35 U.S.C. 103 as being unpatentable over Reitermann et al. (US 2024/0003918) in view of NPL “Ordered quantile normalization: a semiparametric transformation built for the cross-validation era” (hereinafter referred to as “Peterson”) in view of NPL “Alzheimer’s Disease: The Relative Importance Diagnostic” (hereinafter referred to as “Habadi”) in view of Newman et al. (US 2016/0217253) in view of Gotcher et al. (US 2017/0132379).
In re claim 1, Reitermann discloses a system for cognitive disease prediction [0007], comprising:
a memory [0009]; and
a processor [0009] communicatively coupled to the memory [0009];
wherein the memory stores a set of instructions [0009] which, when executed by the processor [0009], cause the processor to:
receive a plurality of risk factor ([0070-0072]: factors such as genetic markers, age, gender, education, etc.) indications for a given patient;
determine a plurality of interaction indications of the patient ([0070]: plurality of interaction indications are correlations between protein markers and other factors),
each interaction indication of the plurality of interaction indications being an indication of interaction between at least two risk factor indications of the plurality of risk factor indications ([0070]: risk score algorithm may determine correlation between protein markers and other factors on subjects’ likelihood to develop AD and may identity top components for the risk score algorithm; [0072, 0177-0178]);
obtain a trained machine learning model ([0177-0178]: machine learning model is trained using patient records that correlates protein markers to AD surrogate variables; [0163]: AI-based algorithm),
apply the plurality of risk factor indications and the plurality of interaction indications to the trained machine learning model [0177-0178]; and
output a predicted conversion time ([0053]: AD risk score predicts brain neurodegeneration disease risk) based on the trained machine learning model ([0163]: AI algorithm used to generate AD risk score).
Reitermann fails to disclose
transform, using an ordered quantile transformer, a first probability distribution of the plurality of risk factor indications and the plurality of interaction indications to a second probability distribution such that the plurality of risk factor indications and the plurality of interaction indications follow a normal distribution;
wherein the trained machine learning model is a multivariate linear regression model trained by:
(i) applying an ordered quantile normalization transformation to risk factor indications and interaction indications of a plurality of training patients to generate normalized training data:
(ii) applying a backward feature selection process to the normalized training data to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time from a first cognitive disease state to a second cognitive disease state: and
(iii) iteratively adjusting a plurality of coefficients corresponding to the identified subset of risk factor indications and interaction indications to minimize a root mean square error (RMSE) between predicted conversion times and ground truth conversion time indications of the plurality of training patients;
apply the transformed plurality of risk factor indications and the transformed plurality of interaction indications to the trained machine learning model; and
output a predicted conversion time from Mild Cognitive Impairment to Alzheimer's Disease in the patient based on the trained machine learning model;
track a plurality of predicted conversion times from Mild Cognitive Impairment to Alzheimer's Disease in the patient while the patient is administered a drug for Alzheimer's Disease treatment, wherein tracking the plurality of predicted conversion times comprises determining whether the drug is increasing or decreasing the predicted conversion time in the patient; and
transmit a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate, wherein the notification is configured to prompt an appointment between the medical professional and the patient.
Regarding the limitations,
“transform, using an ordered quantile transformer, a first probability distribution of the plurality of risk factor indications and the plurality of interaction indications to a second probability distribution such that the plurality of risk factor indications and the plurality of interaction indications follow a normal distribution;
wherein a trained machine learning model is a multivariate linear regression model trained by: (i) applying an ordered quantile normalization transformation to risk factor indications and interaction indications of a plurality of training patients to generate normalized training data;
apply the transformed plurality of risk factor indications and the transformed plurality of interaction indications to the trained machine learning model,”
Peterson teaches normally distributing transformed data (pg. 2, lines 13-23) to analyze data (pg. 2, lines 18-20) and teaches
transforming, using an ordered quantile transformer (pg. 2, lines 15-23), a first probability distribution (pg. 2, lines 15-23) of a plurality indications such that the plurality of indications follow a normal distribution (pg. 3, “2. The ordered quantile normalization technique” lines 1-8)
wherein training occurs by: (i) applying an ordered quantile normalization transformation to indications and interaction indications of a plurality of training data to generate normalized training data (pg. 4, lines 11-24).
apply the transformed plurality of indications and the transformed plurality of interaction indications to the trained machine learning model (pg. 15, lines, 1-18.
Peterson further teaches that OQT is an effective normalization technique (pg. 15, lines 1-3) and that also obtains superior results compared to those arising from generalized additive and ordinary least squares fits (pg. 4, lines 1-7).
It would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the system for cognitive disease prediction taught by Reitermann, to cause the processor to transform, using an ordered quantile transformer, a first probability distribution of the plurality of risk factor indications and the plurality of interaction indications to a second probability distribution such that the plurality of risk factor indications and the plurality of interaction indications follow a normal distribution; wherein a trained machine learning model is a multivariate linear regression model trained by: (i) applying an ordered quantile normalization transformation to risk factor indications and interaction indications of a plurality of training patients to generate normalized training data; apply the transformed plurality of risk factor indications and the transformed plurality of interaction indications to the trained machine learning model, as taught by the transformed plurality of indications and the transformed plurality of interaction indications of Peterson, because OQT is an effective normalization technique that obtains superior results compared to those arising from generalized additive and ordinary least squares fits.
Regarding the limitations,
“wherein the trained machine learning model is a multivariate linear regression model trained by: (ii) applying a backward feature selection process to the normalized training data to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time from a first cognitive disease state to a second cognitive disease state”,
Habadi teaches an analogous method for predicting a status of Alzheimer’s disease in a patient (pg. 80, 3. Statistical Method, lines 1-6) and wherein a trained machine learning model (abstract: statistical prediction model with multiple logistic regression is used to identify Alzheimer’s disease patients) comprises a multivariate linear regression machine learning model (pg. 81, 4. Implementation of the Multiple Logistic Model: equation 2 provides a logistic regression module with multiple variables; pg. 80: 3. Statistical Method, lines 7-14: equation 1 is a linear form of the logistic regression model) learned trained b: (ii) a backward feature selection process to a normalized training data (pg. 81: , 4. Implementation of the Multiple Logistic Model: lines 8-16: backward elimination algorithm) to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time from a first cognitive disease state to a second cognitive disease state (pg. 81: , 4. Implementation of the Multiple Logistic Model: lines 8-27: risk factors and interaction terms are used after backward elimination algorithm to predict patient’s condition, which is interpreted as going from a first cognitive disease state to a second cognitive disease state).
Habadi further teaches that the full logistic regression model predicts the probability of Alzheimer’s disease in a subject based on all predictors and all possible interactions (pg. 81, 4. Implementation of the Multiple Logistic Model: lines 1-8) to determine which interactions are statistically signification (pg. 80, 3. Statistical Method, lines 1-5; pg. 81, 4. Implementation of the Multiple Logistic Model: lines 1-8, lines 22-26) for predicting Alzheimer’s disease (pg. 80, 3. Statistical Method, lines 1-5).
Habadi also teaches wherein a first coefficient (pg. 81, 4. Implementation of the Multiple Logistic Model: lines 1-8: B1) and a second coefficient (pg. 81, 4. Implementation of the Multiple Logistic Model: lines 1-8: B2) were determined during a training phase of the multivariate linear regression machine learning model (pg. 81, 4. Implementation of the Multiple Logistic Model: lines 1-8: data set is divided into training and testing and the coefficients must be determined).
It would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the system for cognitive disease prediction yielded by the proposed combination, to provide wherein the trained machine learning model is a multivariate linear regression model trained by: (ii) applying a backward feature selection process to the normalized training data to identify a subset of risk factor indications and interaction indications that significantly contribute to a conversion time from a first cognitive disease state to a second cognitive disease state, as taught by Habadi, because the linear form of the logistic regression model predicts the probability of Alzheimer’s disease in a subject based on all predictors and all possible interactions to determine which interactions are statistically signification for predicting Alzheimer’s disease.
Regarding the limitation, “(iii) iteratively adjusting a plurality of coefficients corresponding to the identified subset of risk factor indications and interaction indications to minimize a root mean square error (RMSE) between predicted conversion times and ground truth conversion time indications of the plurality of training patients”, Newman teaches a method of evaluating a disease in an individual [025], wherein the disease may be Alzheimer’s disease [0189], and teaches iii) iteratively adjusting a plurality of coefficients [0143] corresponding to an identified subset of indications ([0143]: feature profile of a physical sample) and interaction indications [0143] to minimize a root mean square error (RMSE) [0143] between predicted conversion times and ground truth conversion time indications of a plurality of training patients ([0143]: RMSE calculated between feature profile and product of the result and would include biological sample from patients).
Newman further teaches that iterating through different values to obtain different estimates of a relative proportions of distinct components during training [0143].
It would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the system for cognitive disease prediction yielded by the proposed combination, to provide (iii) iteratively adjusting a plurality of coefficients corresponding to the identified subset of risk factor indications and interaction indications to minimize a root mean square error (RMSE) between predicted conversion times and ground truth conversion time indications of the plurality of training patients, as taught by Newman, because iterating through different values to obtain different estimates of a relative proportions of distinct components during training.
Regarding the limitations,
“output a predicted conversion time from Mild Cognitive Impairment to Alzheimer's Disease in the patient based on the trained machine learning model;
track a plurality of predicted conversion times from Mild Cognitive Impairment to Alzheimer's Disease in the patient while the patient is administered a drug for Alzheimer's Disease treatment, wherein tracking the plurality of predicted conversion times comprises determining whether the drug is increasing or decreasing the predicted conversion time in the patient; and
transmit a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate”,
Reiman teaches a method of measuring activity in a human brain [0003] determine efficacy of treatments for brain-related disorders [0003], and teaches wherein a result comprises a predicted conversion time of Mild Cognitive Impairment (MCI) to Alzheimer's Disease (AD) in a patient ([0050]: rates of conversion of MCI to AD; [0046-0048]) and teaches
outputting a predicted conversion time from Mild Cognitive Impairment to Alzheimer's Disease in the patient based on the trained machine learning model [0046-0048];
tracking a plurality of predicted conversion times from Mild Cognitive Impairment to Alzheimer's Disease in the patient while the patient is administered a drug for Alzheimer's Disease treatment [0046-0048],
wherein tracking the plurality of predicted conversion times comprises determining whether the drug is increasing or decreasing the predicted conversion time in the patient [0050].
Reiman further teaches that a treatment administered to a patient may be used to predict decreased rates of conversion of MCI to AD [0050], and that prevention therapy provides extraordinary public health benefit [0008] such that delaying AD onset reduces a number of cases [0008].
It would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the system for cognitive disease prediction yielded by the proposed combination, to provide outputting a predicted conversion time from Mild Cognitive Impairment to Alzheimer's Disease in the patient based on the trained machine learning model; tracking a plurality of predicted conversion times from Mild Cognitive Impairment to Alzheimer's Disease in the patient while the patient is administered a drug for Alzheimer's Disease treatment, wherein tracking the plurality of predicted conversion times comprises determining whether the drug is increasing or decreasing the predicted conversion time in the patient, as taught by Reiman, because a treatment administered to a patient may be used to predict decreased rates of conversion of MCI to AD, which provides extraordinary public health benefit such as reducing a number of cases.
Regarding the limitations, “transmit a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate, wherein the notification is configured to prompt an appointment between the medical professional and the patient”, Gotcher teaches improving care for asthma sufferers [0012] and teaches outputting a notification [0097] to prompt an appointment between a medical professional and a patient [0097] when a user’s score [0097] is below a certain threshold [097].
Gotcher further teaches that healthcare provider is only notified when the user’s score is below the threshold [0097] and allows them to automatically call or email the patient to schedule an appointment [0097].
It would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the system for cognitive disease prediction yielded by the proposed combination, to provide transmitting a notification to a medical professional associated with the patient when the predicted conversion time is less than a predetermined threshold period or when a change in the tracked predicted conversion time exceeds a threshold change rate, wherein the notification is configured to prompt an appointment between the medical professional and the patient, as taught by the user’s score of Gotcher being below a certain threshold, because the healthcare provider is only notified when the user’s score is below the threshold and allows them to automatically call or email the patient to schedule an appointment.
The proposed combination would be for the trained machine learning model of Reitermann to apply the ordered quantile transformer of Peterson to the risk factor indications and interaction indications of a plurality of training patients of Reitermann. The proposed combination would further yield applying a backward feature selection process to the normalized training data of the proposed combination of Reitermann and Peterson, to identify a subset of risk factor indications and interaction indications, as well as iteratively adjust a plurality of coefficients corresponding to the identified subset of risk factor indications and interaction indications to minimize a root mean square error (RMSE) between predicted conversion times and ground truth conversion time indications of the plurality of training patients, as taught by Newman. The proposed combination would be further modified in view of Reiman to predict whether the drug is increasing or decreasing the predicted conversion time in the patient, which would then be used to provide a health provider with a notification to schedule an appointment with the patient when the predicted conversion time is less than a predetermined threshold period, as taught by Gotcher.
In re claim 2, Reitermann discloses wherein the plurality of risk factor indications includes at least one selected from a group of:
clinical data ([0173]: cognitive assessment, subject’s gender, and subject’s education may be used),
genetic data [0173],
biospecimen data ([0052]: at least 4 protein markers are selected; [0074, 0077]), and
medical image data [0188, 0254].
In re claim 3, Reitermann discloses wherein the clinical data includes at least one selected from a group of:
a cognitive assessment score [0173]: subject’s performance on a cognitive assessment,
an omega-3 score, or
demographic statistics ([0188]: age and gender; [0173]).
In re claim 4, Reitermann discloses wherein the genetic data includes: Apolipoprotein E (APOE) genotyping ([0214-0217]: APOE status may be associated with each subject).
In re claim 5, Reitermann discloses wherein the biospecimen data includes at least one selected from a group of:
a quantity of P-tau protein,
tau protein [0074], or
Beta-amyloid [0077].
In re claim 6, Reitermann discloses wherein the medical image data includes at least one selected from a group of:
Hippocampus ([0243-0245]: variables can be categorized from images such as from TAU images; [0083]: Tau accumulation seen in the hippocampus),
ventricles,
entorhinal ([0243-0245]: variables can be categorized from images such as from TAU images; [0083]: Tau accumulation seen in the entorhinal cortex),
intracranial volume (ICV), and
fusiform medical image data.
In re claim 7, Reitermann discloses wherein the plurality of risk factor indications comprises at least one selected from a group of:
an age indication [0178],
an education indication [0178],
a ventricles indication,
a hippocampus indication,
an entorhinal indication,
a fusiform indication,
an amyloid- beta indication,
a tau indication,
a pTau indication, and
an Alzheimer Disease Assessment Scale (ADAS) indication for the patient.
In re claim 8, Reitermann discloses wherein the plurality of interaction indications comprises at least one selected from a group of:
a first interaction indication between amyloid-beta ([0074]: amyloid peptide can be used, for instance amyloid beta [0077]) and tau ([0074]: tau peptide marker; [0052]: AD risk score comprises at least 4 protein markers),
a second interaction indication between hippocampus and pTau,
a third interaction indication between ventricles and pTau, and
a fourth interaction indication between hippocampus and education.
Claims 10-11 are rejected under 35 U.S.C. 103 as being unpatentable over Reitermann et al. (US 2024/0003918) in view of NPL “Ordered quantile normalization: a semiparametric transformation built for the cross-validation era” (hereinafter referred to as “Peterson”) in view of NPL “Alzheimer’s Disease: The Relative Importance Diagnostic” (hereinafter referred to as “Habadi”) in view of Newman et al. (US 2016/0217253) in view of Gotcher et al. (US 2017/0132379) in view of Huiku (US 2010/0081942).
In re claim 10, the proposed combination fails to yield wherein the multivariate linear regression machine learning model is defined as:
a result = β0 + ∑iaixi + ∑jYjkj + £i,
wherein β0 is an intercept of the multivariate linear regression machine learning model,
ai is a first coefficient of ith individual risk factor indication xi of the plurality of risk factor indications,
Yj is a second coefficient of jth interaction indication kj of the plurality of interaction indications, and
£i is a residual error of the multivariate linear regression machine learning model.
Habadi teaches wherein the multivariate linear regression machine learning model is defined as:
a result = β0 + ∑iaixi + ∑jYjkj (pg. 81, equation 3)
wherein β0 is an intercept of the multivariate linear regression machine learning model (equation 3: inherent that β0 is an intercept),
ai is a first coefficient of ith individual risk factor indication xi of the plurality of risk factor indications (pg. 81, 4. Implementation of the Multiple Logistic Model, lines 1-8: β1 is a first coefficient of the risk factors X’s),
Yj is a second coefficient of jth interaction indication kj of the plurality of interaction indications (pg. 81, 4. Implementation of the Multiple Logistic Model, lines 1-8: β2 is a first coefficient of the interaction indications X’s which are also for the risk factors).
For substantially the same reasons as discussed in re claim 10 above, it would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the system for cognitive disease prediction yielded by the proposed combination, to provide wherein the multivariate linear regression machine learning model is defined as: a result = β0 + ∑iaixi + ∑jYjkj, as taught by Habadi.
Regarding the limitation, “£i is a residual error of the multivariate linear regression machine learning model”, Huiku teaches estimating blood plasma volume using a linear regression model [0069], wherein the linear regression module includes adding a residual error [0069].
Huiku further teaches that the residual error should be minimized [0072], and the residual error is known for representing unexplained variation after a model has been made.
It would have been obvious to someone of ordinary skill in the art at the time the instant invention was filed to modify the result yielded by the proposed combination, to provide wherein a result includes adding the residual error, as taught by Huiku, because the residual error should be minimized, and because the residual error is known for representing unexplained variation after a model has been made.
In re claim 11, regarding the limitation, “wherein the first coefficient and the second coefficient were determined during a training phase of the multivariate linear regression machine learning model”, see in re claim 10 above.
In re claim 13, regarding the limitation, “further comprising obtaining a result indicating effectiveness of a drug in increasing or decreasing the predicted conversion time in the patient”, see the proposed combination yielded in re claim 1 above.
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). 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.
Contact
Any inquiry concerning this communication or earlier communications from the examiner should be directed to RUMAISA R BAIG whose telephone number is (571)270-0175. The examiner can normally be reached Mon-Fri: 8am- 5pm.
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, David Hamaoui can be reached at (571) 270-5625. 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.
/RUMAISA RASHID BAIG/Examiner, Art Unit 3796
/DAVID HAMAOUI/SPE, Art Unit 3796
Prosecution Insights