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 .
The response was received on 7/13/2026. Claims 1-19 and 21 are pending where claims 1-19 were previously presented; claim 20 is cancelled; and claim 21 is newly added.
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 1-19 and 21 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
With regard to claim 1:
Step 2A, Prong One:
The claim recites the following limitations which are drawn towards an abstract idea:
A method, comprising: …
accessing a machine learning model configured to generate risk score output indicating a chronic kidney disease (CKD) progression, … applying an input dataset associated with a new patient patient (recites certain methods of organizing human activity and mental process step of making a prediction/judgment based on evaluations/observations similar to health care professionals when evaluating patient information where the mental process steps can perform the respective analysis steps with the aid of a computer).
As seen from above, the identified limitations recite concepts associated with an abstract idea and thus the respective claim recites a judicial exception (see 2106.04(a)) and thus requires further analysis as discussed below.
Step 2A, Prong Two:
The following limitations have been identified as being additional elements as discussed below.
accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression, wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk score output (recites merely apply it limitation using generic computer element including machine learning to perform the judicial exception including the initial training of the machine learning model at a high-level of generality, see MPEP 2106.05(f)),
wherein the training dataset comprises (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients, and (iv) CKD clinical outcomes associated with the plurality of patients, the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase, serum phosphate, serum bicarbonate, serum magnesium, serum calcium, aspartate aminotransferase (AST), alanine transaminase (ALT), bilirubin, gamma-glutamyl transferase (GGT), hematocrit, and platelet count (recites field of use limitations describing the particular/preferred data and its respective meaning in order to perform the judicial exception, see MPEP 2106.05(h));
to the machine learning model to cause the machine learning model to process the input dataset via the model parameters of the machine learning model
the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient: eGFR, urine ACR, urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase (ALKP), serum phosphate, serum bicarbonate, serum magnesium, serum calcium, AST, ALT, bilirubin, GGT, hematocrit, and platelet count (recites field of use limitations describing the particular data and its respective meaning, see MPEP 2106.05(h)).
This judicial exception is not integrated into a practical application because, as seen from the above discussion, the identified limitations did not integrate the judicial exception into a practical application (see MPEP 2106.04(d)). The additional elements recite field of use limitations describing the particular data and its respective meaning as well as the usage of a generic computer machine learning model to implement the abstract idea.
Step 2B:
Below is the analysis of the claims:
accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression, wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk score output (recites merely apply it limitation using generic computer element including machine learning to perform the judicial exception including the initial training of the machine learning model at a high-level of generality, see MPEP 2106.05(f)),
wherein the training dataset comprises (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients, and (iv) CKD clinical outcomes associated with the plurality of patients, the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase, serum phosphate, serum bicarbonate, serum magnesium, serum calcium, aspartate aminotransferase (AST), alanine transaminase (ALT), bilirubin, gamma-glutamyl transferase (GGT), hematocrit, and platelet count (recites field of use limitations describing the particular/preferred data and its respective meaning in order to perform the judicial exception, see MPEP 2106.05(h));
to the machine learning model to cause the machine learning model to process the input dataset via the model parameters of the machine learning model
the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient: eGFR, urine ACR, urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase (ALKP), serum phosphate, serum bicarbonate, serum magnesium, serum calcium, AST, ALT, bilirubin, GGT, hematocrit, and platelet count (recites field of use limitations describing the particular data and its respective meaning, see MPEP 2106.05(h)).
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, as seen from above, the respective claim elements taken individually do not amount to significantly more than the judicial exception. When taken as a whole (in combination), the claim also does not amount to significantly more than the abstract idea because the additional elements recite field of use limitations describing the particular data and its respective meaning as well as the usage of a generic computer machine learning model to implement the abstract idea.
With regard to claim 2, this claim recites wherein the new patient is not associated with a CKD stage of G3 or later (recites field of use limitations describing particular preferred conditions of a user, see MPEP 2106.05(h)).
With regard to claim 3, this claim recites wherein the machine learning model comprises a random survival forest model (recites merely using generic machine learning models as a computer tool to perform the abstract idea, see MPEP 2106.05(f)).
With regard to claim 4, this claim recites wherein the risk score output indicating CKD progression for the new patient indicates a risk of experiencing CKD progression within a particular amount of time from a time period associated with the input dataset for the new patient (recites field of use limitations describing what the respective prediction/output value is meant to represent, see MPEP 2106.05(h)).
With regard to claim 5, this claim recites wherein the particular amount of time is provided as input to the machine learning model for generating the risk score output indicating CKD progression for the new patient (recites insignificant extrasolution activity of transmitting/receiving information over a network which amounts to well-understood, routine, and conventional activity of transmitting/receiving information, see MPEP 2106.05(d)).
With regard to claim 6, this claim recites wherein the particular amount of time comprises 2 years or 5 years (recites field of use limitations describing preferred data values to be used, see MPEP 2106.05(f)).
With regard to claim 7, this claim recites wherein the urine ACR for one or more of the plurality of patients or the new patient is converted from a urine protein-to-creatinine test or a urine dipstick test (recites field of use limitations describing particular manipulation/usage of particular tests).
With regard to claim 8, this claim recites wherein the risk score output indicating CKD progression for the new patient indicates a risk of the new patient experiencing kidney failure or about a 40% or greater decline of the eGFR for the new patient (recites mental process steps of evaluating the health of the employee and the respective meaning of the prediction).
With regard to claim 9, this claim recites wherein the risk of the new patient experiencing kidney failure comprises an indication that the new patient is at risk of (i) requiring chronic dialysis, (ii) requiring a kidney transplant, or (iii) experiencing a glomerular filtration rate of less than 10 ml/min/1.73m2 (recites field of use limitations describing a potential outcome associated with the prediction, see MPEP 2106.05(h)).
With regard to claim 10, this claim recites determining that the risk score output indicating CKD progression for the new patient indicates a predicted risk of the new patient experiencing CKD within a particular time period that satisfies one or more predicted risk threshold values (recites mental process steps of evaluating patient/user information);
and (i) generating a notification that the new patient may need an interventive kidney treatment; (ii) generating a recommendation of an interventive kidney treatment for the new patient based on the risk score indicating CKD progression for the new patient; (iii) generating a recommendation of a frequency of monitoring of CKD progression for the new patient based on the risk score output indicating CKD progression for the new patient; or (iv) administering an interventive kidney treatment to the new patient (recites mental process steps of generating/deciding information for the user/patient including recommendations of best options/treatments).
With regard to claim 11, this claim recites wherein the one or more predicted risk threshold values are based upon the particular time period associated with the risk score output indicating CKD progression for the new patient (recites mental process steps of using particular information as means to perform evaluation and judgments of the input data).
With regard to claim 12, this claim recites wherein the recommendation of the interventive kidney treatment or the recommendation of the frequency of monitoring of CKD progression is further based upon at least some of the second set of medical laboratory data associated with the new patient (recites mental process steps of evaluating particular information in order to make mental decisions/determinations/judgments).
With regard to claim 13, this claim recites wherein the interventive kidney treatment comprises one or more of: renin-angiotensin-aldosterone system (RAAS) inhibition, blood pressure control, sodium-glucose cotransporter-2 (SGLT2) inhibitor medication, mineralocorticoid receptor antagonists (MRAs) therapy, or preparation for nephrology consultation, home dialysis, dialysis access, or kidney transplant (recites field of use limitations describing particular treatment options, see MPEP 2106.05(h)).
With regard to claim 14, this claim recites wherein the first set of medical laboratory data comprises one or more imputed values in place of missing values (recites mental process steps of assuming/inferring a value).
With regard to claim 15, this claim recites wherein the first set of medical laboratory data indicates, with a degree of value imputation of 30% or less, eGFR, urine ACR, urea, potassium, hemoglobin, platelet count, albumin, calcium, glucose, bilirubin, sodium, bicarbonate, and GGT (recites field of use limitations describing the particular data that is being used, see MPEP 2106.05(h)).
With regard to claim 16:
Step 2A, Prong One:
The claim recites the following limitations which are drawn towards an abstract idea:
generate one or more imputed values corresponding to one or more missing values in the first set of medical laboratory data to construct an imputed training dataset (recites mental process steps of interpolation or other statistical analysis steps to infer/estimate a value);
to generate risk score output indicating chronic kidney disease (CKD) progression for the new patient by applying an input dataset associated with the new patient to the machine learning model (recites certain methods of organizing human activity and mental process step of making a prediction/judgment based on evaluations/observations similar to health care professionals when evaluating patient information where the mental process steps can perform the respective analysis steps with the aid of a computer).
As seen from above, the identified limitations recite concepts associated with an abstract idea and thus the respective claim recites a judicial exception (see 2106.04(a)) and thus requires further analysis as discussed below.
Step 2A, Prong Two:
The following limitations have been identified as being additional elements as discussed below.
A system, comprising: one or more processors; and one or more hardware storage devices storing instructions that are executable by the one or more processors to configure the systems (recites generic computer hardware at a high-level of generality to implement generic computer functionality to implement the judicial exception, see MPEP 2106.05(f)) to:
accessing a training dataset (recites insignificant extrasolution activity of retrieving information from memory, see MPEP 2106.05(g))
comprising (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients, and (iv) CKD clinical outcomes associated with the plurality of patients, the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea, serum hemoglobin, glucose, platelet count, and hematocrit (recites field of use limitations describing the particular/preferred data and its respective meaning in order to perform the judicial exception, see MPEP 2106.05(h));
generate a machine learning model by applying the imputed training dataset to an untrained machine learning model, wherein applying the imputed dataset to the untrained machine learning model configures model parameters of the machine learning model for processing input data
the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient: eGFR, urine ACR, urea, serum hemoglobin, glucose, platelet count, and hematocrit (recites field of use limitations describing the particular data and its respective meaning, see MPEP 2106.05(h)).
This judicial exception is not integrated into a practical application because, as seen from the above discussion, the identified limitations did not integrate the judicial exception into a practical application (see MPEP 2106.04(d)). The additional elements recite field of use limitations describing the particular data and its respective meaning as well as the usage of a generic computer machine learning model to be able to implement the abstract idea associated with medical evaluation of a patient.
Step 2B:
Below is the analysis of the claims:
A system, comprising: one or more processors; and one or more hardware storage devices storing instructions that are executable by the one or more processors to configure the systems (recites generic computer hardware at a high-level of generality to implement generic computer functionality to implement the judicial exception, see MPEP 2106.05(f)) to:
accessing a training dataset (recites well-understood, routine, and conventional activity of retrieving information from memory, see MPEP 2106.05(d))
comprising (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients, and (iv) CKD clinical outcomes associated with the plurality of patients, the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea, serum hemoglobin, glucose, platelet count, and hematocrit (recites field of use limitations describing the particular/preferred data and its respective meaning in order to perform the judicial exception, see MPEP 2106.05(h));
generate a machine learning model by applying the imputed training dataset to an untrained machine learning model, wherein applying the imputed dataset to the untrained machine learning model configures model parameters of the machine learning model for processing input datathe computer as a tool, such as using a machine learning model, to implement the abstract idea, see MPEP 2106.05(f)),
the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient: eGFR, urine ACR, urea, serum hemoglobin, glucose, platelet count, and hematocrit (recites field of use limitations describing the particular data and its respective meaning, see MPEP 2106.05(h)).
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, as seen from above, the respective claim elements taken individually do not amount to significantly more than the judicial exception. When taken as a whole (in combination), the claim also does not amount to significantly more than the abstract idea because the additional elements recite field of use limitations describing the particular data and its respective meaning as well as the usage of a generic computer machine learning model to implement the abstract idea.
With regard to claim 17, this claim recites wherein the machine learning model comprises a random survival forest model (recites merely using generic machine learning models as a computer tool to perform the abstract idea, see MPEP 2106.05(f)), and wherein the one or more imputed values are generated using adaptive tree imputation (recites merely using a particular machine learning algorithm/model as a computer tool to perform the abstract idea, see MPEP 2106.05(f)).
With regard to claim 18:
Step 2A, Prong One:
The claim recites the following limitations which are drawn towards an abstract idea:
A method, comprising: …
accessing a machine learning model configured to generate risk score output indicating a chronic kidney disease (CKD) progression, … applying (i) the time period input and (ii) an input dataset associated with a new patient
As seen from above, the identified limitations recite concepts associated with an abstract idea and thus the respective claim recites a judicial exception (see 2106.04(a)) and thus requires further analysis as discussed below.
Step 2A, Prong Two:
The following limitations have been identified as being additional elements as discussed below.
accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression, wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk score output (recites merely apply it limitation using generic computer element including machine learning to perform the judicial exception including the initial training of the machine learning model at a high-level of generality, see MPEP 2106.05(f)),
wherein the training dataset comprises a first set of medical laboratory data associated with a plurality of patients, the first of medical laboratory data including an age and sex of each patient included in the plurality of patients, the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: urine albumin-to-creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), urea, glucose, hematocrit, platelet count, and hemoglobin (recites field of use limitations describing the particular/preferred data and its respective meaning in order to perform the judicial exception, see MPEP 2106.05(h));
receiving user input indicating a time period input from a plurality of available time period inputs (recites insignificant extrasolution activity of receiving information, see MPEP 2106.05(g));
to the machine learning model such that the machine learning model processes the time period input together with the input dataset via the model parameters of the machine learning model
the input dataset comprising patient information including at least an age and sex of the new patient, and a second set of medical laboratory data comprising: urine ACR, eGFR, urea, glucose, hematocrit, platelet count, and hemoglobin for the new patient (recites field of use limitations describing the particular data and its respective meaning, see MPEP 2106.05(h)).
This judicial exception is not integrated into a practical application because, as seen from the above discussion, the identified limitations did not integrate the judicial exception into a practical application (see MPEP 2106.04(d)). The additional elements recite field of use limitations describing the particular data and its respective meaning as well as receiving input for when to make a prediction via usage of a generic computer machine learning model to implement the abstract idea.
Step 2B:
Below is the analysis of the claims:
accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression, wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk score output (recites merely apply it limitation using generic computer element including machine learning to perform the judicial exception including the initial training of the machine learning model at a high-level of generality, see MPEP 2106.05(f)),
wherein the training dataset comprises a first set of medical laboratory data associated with a plurality of patients, the first of medical laboratory data including an age and sex of each patient included in the plurality of patients, the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: urine albumin-to-creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), urea, glucose, hematocrit, platelet count, and hemoglobin (recites field of use limitations describing the particular/preferred data and its respective meaning in order to perform the judicial exception, see MPEP 2106.05(h));
receiving user input indicating a time period input from a plurality of available time period inputs (recites well-understood, routine, and conventional activity of receiving information, see MPEP 2106.05(d));
to the machine learning model such that the machine learning model processes the time period input together with the input dataset via the model parameters of the machine learning model tool, such as using a machine learning model, to implement the abstract idea, see MPEP 2106.05(f)),
the input dataset comprising patient information including at least an age and sex of the new patient, and a second set of medical laboratory data comprising: urine ACR, eGFR, urea, glucose, hematocrit, platelet count, and hemoglobin for the new patient (recites field of use limitations describing the particular data and its respective meaning, see MPEP 2106.05(h)).
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, as seen from above, the respective claim elements taken individually do not amount to significantly more than the judicial exception. When taken as a whole (in combination), the claim also does not amount to significantly more than the abstract idea because the additional elements recite field of use limitations describing the particular data and its respective meaning as well as receiving input for when to make a prediction via usage of a generic computer machine learning model to implement the abstract idea.
With regard to claim 19, this is substantially similar to claim 3 and is rejected for similar reasons as discussed above.
With regard to claim 21, this claim recites wherein applying the imputed training dataset to the untrained machine learning model causes the machine learning model to exhibit a variable importance ranking in which urine ACR, eGFR, urea, and serum hemoglobin are ranked higher than hematocrit and glucose (recites mental process steps and method of organizing human activity of evaluation and judgement of what attributes to give greater weight or importance too when forming an overall medical judgement/decision/evaluation).
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1-6, 8-13, and 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over Cha et al [WO 2020/006571 A1] (from IDS) in view of Naylor et al [US 2010/0076787 A1].
With regard to claim 1, Cha teaches a method, comprising: accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression (see page 6, lines 26-28; page 8, lines 8-13; the system can employ/access machine learning models to make predictions associated with CKD including calculation of risk scores),
wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk output (see page 8, lines 8-13; page 21, lines 13-19; page 52, lines 4-6 & 13-15; training data based on features from patient information can be used to train the machine learning model and generate a CKD progression prediction/risk),
wherein the training data set comprises: (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients (see page 17, lines 9-13; page 18, line 22 through col 19, line 2; demographic information and lab test information can be used as training data set),
and (iv) CKD clinical outcomes associated with the plurality of patients (see page 47, lines 8-12; page 21, lines 13 through page 24, line 25; and page 31, lines 1-18; the system has means to utilize CKD clinical outcomes as means part of the training/validation process for the machine learning model),
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea, serum sodium, serum chloride,
applying an input dataset associated with a new patient to the machine learning model to cause the machine learning model to process the input dataset via the model parameters of the machine learning model to generate risk score output indicating CKD progression for the new patient, the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient (see page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used including patient demographics and various other diagnostic tests).
Cha teaches various laboratory data but does not appear to explicitly teach:
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: serum hemoglobin, glucose, alkaline phosphatase, aspartate aminotransferase (AST), alanine transaminase (ALT), bilirubin, gamma-glutamyl transferase (GGT), hematocrit, and platelet count;
the input dataset comprising: eGFR, urine ACR, urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase (ALKP), serum phosphate, serum bicarbonate, serum magnesium, serum calcium, AST, ALT, bilirubin, GGT, hematocrit, and platelet count.
Naylor teaches the first set of medical laboratory data indicating (see paragraph [0012]; the medical data can have various individual readings; see paragraphs [0194]-[0365] for entire list of examples and individual readings that can be measured and used), for at least a combination of patients included in the plurality of patients: serum hemoglobin (paragraphs [0231]-[0233]), glucose (see paragraphs [0336]-[0338]), alkaline phosphatase (see paragraphs [0204]-[0206]), aspartate aminotransferase (AST) (see paragraphs [0213]-[0215]), alanine transaminase (ALT) (see paragraphs [0195]-[0197]), bilirubin (see paragraphs [0216]-[0218]), gamma-glutamyl transferase (GGT) (see paragraphs [0330]-[0332]), hematocrit (see paragraphs [0234]-[0236]), and platelet count (see paragraphs [0243]-[0245]).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the feature data that is used for training and making predictions as taught by Cha by including other feature data from various medical lab and test results as taught by Naylor in order to provide more relevant data so that the models can have more information that help identify amount of risk of the disease so that the trained model can be more accurate when being used since accuracy of medical diagnoses is a high-priority for both doctors (and other medical personnel) and the respective patients.
Cha in view of Naylor teach the input dataset comprising: eGFR, urine ACR, urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase (ALKP), serum phosphate, serum bicarbonate, serum magnesium, serum calcium, AST, ALT, bilirubin, GGT, hematocrit, and platelet count (see Cha, page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; page 36, lines 1-16; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used from the patient’s records to determine their risk score; see Cha page 4, lines 1-7; page 18, line 22 through col 19, line 2; for examples of the input data that can be utilized to make a prediction or determine a risk score and also see Naylor, paragraphs [0194]-[0365] for entire list of examples and individual readings that can be measured and used including serum hemoglobin (paragraphs [0231]-[0233]), glucose (see paragraphs [0336]-[0338]), alkaline phosphatase (see paragraphs [0204]-[0206]), aspartate aminotransferase (AST) (see paragraphs [0213]-[0215]), alanine transaminase (ALT) (see paragraphs [0195]-[0197]), bilirubin (see paragraphs [0216]-[0218]), gamma-glutamyl transferase (GGT) (see paragraphs [0330]-[0332]), hematocrit (see paragraphs [0234]-[0236]), and platelet count (see paragraphs [0243]-[0245]).
With regard to claim 2, Cha in view of Naylor teach wherein the new patient is not associated with a CKD stage of G3 or later (see Cha, page 51, lines 7-12; the patients did not have any previous kidney disease diagnoses).
With regard to claim 3, Cha in view of Naylor teach wherein the machine learning model comprises a random survival forest model (see Cha, page 51, lines 7-12; random forest ML models were employed to predict risk).
With regard to claim 4, Cha in view of Naylor teach wherein the risk score output indicating CKD progression for the new patient indicates a risk of experiencing CKD progression within a particular amount of time from a time period associated with the input dataset for the new patient (see Cha, page 51, lines 17-18; the system can make a prediction of patients who would experience, i.e. have a risk of experiencing, a decline over a particular time period).
With regard to claim 5, Cha in view of Naylor teach wherein the particular amount of time is provided as input to the machine learning model for generating the risk score output indicating CKD progression for the new patient (see Cha, page 51, line 17-18; page 21, lines 21-31; the system can make predictions based on a desired time period).
With regard to claim 6, Cha in view of Naylor teach wherein the particular amount of time comprises 2 years or 5 years (see Cha, page 51, line 17-18; the system can make predictions based on a desired time period).
With regard to claim 8, Cha in view of Naylor teach wherein the risk score output indicating CKD progression for the new patient indicates a risk of the new patient experiencing kidney failure or a 40% or greater decline of the eGFR for the new patient (see Cha, page 31, lines 9-15; the system can determine CKD progression based on a 40% decline).
With regard to claim 9, Cha in view of Naylor teach wherein the risk of the new patient experiencing kidney failure comprises an indication that the new patient is at risk of (i) requiring chronic dialysis, (ii) requiring a kidney transplant, or (iii) experiencing a glomerular filtration rate of less than 10 ml/min/1.73m2 (see Cha, page 8, lines 19-23; page 49, lines 2-9; the system can evaluate and predict risk of a patient of various events including kidney failure).
With regard to claim 10, Cha in view of Naylor teach determining that the risk score output indicating CKD progression for the new patient indicates a predicted risk of the new patient experiencing CKD within a particular time period that satisfies one or more predicted risk threshold values; and (i) generating a notification that the new patient may need an interventive kidney treatment; (ii) generating a recommendation of an interventive kidney treatment for the new patient based on the risk score output indicating CKD progression for the new patient; (iii) generating a recommendation of a frequency of monitoring of CKD progression for the new patient based on the risk score output indicating CKD progression for the new patient; or (iv) administering an interventive kidney treatment to the new patient (see Cha, page 4, lines 8-24; see page 6, lines 26-28; the system can make a risk score for the patient over a time period and be able to determine a treatment recommendation and transmit a notification of the treatment recommendation).
With regard to claim 11, Cha in view of Naylor teach wherein the one or more predicted risk threshold values are based upon the particular time period associated with the risk score output indicating CKD progression for the new patient (see Cha, page 25, line 30 through page 26, line 2; the threshold values can be based or associated with a time period).
With regard to claim 12, Cha in view of Naylor teach wherein the recommendation of the interventive kidney treatment or the recommendation of the frequency of monitoring of CKD progression is further based upon at least some of the second set of medical laboratory data associated with the new patient (see Cha, page 25, lines 3-12; the patient information or second data records are utilized to determine the risk score for the patient and what workflow actions should be performed including providing any notifications of recommendations).
With regard to claim 13, Cha in view of Naylor teach wherein the interventive kidney treatment comprises one or more of: renin-angiotensin-aldosterone system (RAAS) inhibition, blood pressure control, sodium-glucose cotransporter-2 (SGLT2) inhibitor medication, mineralocorticoid receptor antagonists (MRAs) therapy, or preparation for nephrology consultation, home dialysis, dialysis access, or kidney transplant (see Cha, page 26, lines 18-21; Figure 2; various treatment options are available).
With regard to claim 18, Cha teaches a method, comprising: accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression (see page 6, lines 26-28; page 8, lines 8-13; the system can employ/access machine learning models to make predictions associated with CKD including calculation of risk scores),
wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk output (see page 8, lines 8-13; page 21, lines 13-19; page 52, lines 4-6 & 13-15; training data based on features from patient information can be used to train the machine learning model and generate a CKD progression prediction/risk),
wherein the training data set comprises a first set of medical laboratory data associated with a plurality of patients, the first set of medical laboratory data including an age and sex of each patient included in the plurality of patients (see page 17, lines 9-13; page 18, line 22 through col 19, line 2; demographic information and lab test information can be used as training data set),
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: urine albumin-to-creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), urea,
receiving user input indicating a time period input from a plurality of available time period inputs (see first paragraph on page 54 and Figure 6 and page 44, lines 18-27 and page 32, lines 5-8; the system allows for prediction time periods to be used to form predictions including prediction start date and prediction end date to from a prediction time period);
applying (i) the time period input and (ii) an input dataset associated with a new patient to the machine learning model such that the machine learning model processes the time period input together with the input dataset via the model parameters of the machine learning model to generate risk score output indicating CKD progression for the new patient based on both (i) the time period input and (ii) the input dataset associated with the new patient, the input dataset comprising patient information including at least an age and sex of the new patient, and a second set of medical laboratory data indicating for the new patient (see page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; page 32, lines 5-8; see Figure 1, box 145; see Figure 6; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used including patient demographics and various other diagnostic tests including being able to have an analysis for a particular prediction time period).
Cha teaches various laboratory data but does not appear to explicitly teach:
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: glucose, hematocrit, platelet count and hemoglobin;
the input dataset comprising…a second set of medical laboratory data indicating for the new patient: urine ACR, eGFR, urea, glucose, hematocrit, platelet count and hemoglobin for the new patient.
Naylor teaches the first set of medical laboratory data indicating (see paragraph [0012]; the medical data can have various individual readings; see paragraphs [0194]-[0365] for entire list of examples and individual readings that can be measured and used), for at least a combination of patients included in the plurality of patients: glucose, hematocrit, platelet count and hemoglobin (serum hemoglobin (paragraphs [0231]-[0233]), glucose (see paragraphs [0336]-[0338]), alkaline phosphatase (see paragraphs [0204]-[0206]), aspartate aminotransferase (AST) (see paragraphs [0213]-[0215]), alanine transaminase (ALT) (see paragraphs [0195]-[0197]), bilirubin (see paragraphs [0216]-[0218]), gamma-glutamyl transferase (GGT) (see paragraphs [0330]-[0332]), hematocrit (see paragraphs [0234]-[0236]), and platelet count (see paragraphs [0243]-[0245])).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the feature data that is used for training and making predictions as taught by Cha by including other feature data from various medical lab and test results as taught by Naylor in order to provide more relevant data so that the models can have more information that help identify amount of risk of the disease so that the trained model can be more accurate when being used since accuracy of medical diagnoses is a high-priority for both doctors (and other medical personnel) and the respective patients.
Cha in view of Naylor teach the input dataset comprising…a second set of medical laboratory data indicating for the new patient: urine ACR, eGFR, urea, glucose, hematocrit, platelet count and hemoglobin for the new patient (see Cha, page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; page 36, lines 1-16; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used from the patient’s records to determine their risk score; see Cha page 4, lines 1-7; page 18, line 22 through col 19, line 2; for examples of the input data that can be utilized to make a prediction or determine a risk score and also see Naylor, paragraphs [0194]-[0365] for entire list of examples and individual readings that can be measured and used including serum hemoglobin (paragraphs [0231]-[0233]), glucose (see paragraphs [0336]-[0338]), alkaline phosphatase (see paragraphs [0204]-[0206]), aspartate aminotransferase (AST) (see paragraphs [0213]-[0215]), alanine transaminase (ALT) (see paragraphs [0195]-[0197]), bilirubin (see paragraphs [0216]-[0218]), gamma-glutamyl transferase (GGT) (see paragraphs [0330]-[0332]), hematocrit (see paragraphs [0234]-[0236]), and platelet count (see paragraphs [0243]-[0245]).
With regard to claim 19, this claim is substantially similar to claim 3 and is rejected for similar reasons as discussed above.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Cha et al [WO 2020/006571 A1] (from IDS) in view of Naylor et al [US 2010/0076787 A1] in further view of Barasch et al [US 2010/0233740 A1].
With regard to claim 7, Cha in view of Naylor teach all the claim limitations of claim 1.
Cha in view of Naylor do not appear to explicitly teach wherein the urine ACR for one or more of the plurality of patients or the new patient is converted from a urine protein-to-creatinine test or a urine dipstick test.
Barasch teaches wherein the urine ACR for one or more of the plurality of patients or the new patient is converted from a urine protein-to-creatinine test or a urine dipstick test (see paragraph [0074]; standard testing methods including usage of a dipstick can be used to determine particular feature’s values such as urine albumin-to-creatinine ratio).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the testing and data collection process as taught by Cha in view of Naylor by utilizing standard laboratory tests and urine dipstick methods to measure/acquire feature values as taught by Barasch in order to utilize standard and widely-used methods of data acquisition that ensues that the acquired samples are retrieved for analysis.
Claims 14 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Cha et al [WO 2020/006571 A1] (from IDS) in view of Naylor et al [US 2010/0076787 A1] in further view of Hsich et al, Identifying Important Risk Factors for Survival in Systolic Heart Failure Patients Using Random Survival Forests (from IDS).
With regard to claim 14, Cha in view of Naylor teach all the claim limitations of claim 1.
Cha in view of Naylor do not appear to explicitly teach wherein the first set of medical laboratory data comprises one or more imputed values in place of missing values.
Hsich teaches wherein the first set of medical laboratory data comprises one or more imputed values in place of missing values (see page 2, last paragraph through the top paragraph on page 3; the system could use informed imputation to fill in missing values).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the testing and data collection process as taught by Cha in view of Naylor by incorporating imputation as taught by Hsich in order to ensure that data sets are complete and provide reasonable data values so that missing data from datasets doesn’t skew the analysis via having too few datasets with values to consider for initial training and potentially giving more undue weight to the non-missing features thus helping the prediction process of the machine learning model to be as informed and accurate as possible so that the most accurate predictions can be given to a patient.
With regard to claim 15, Cha in view of Naylor in further view of Hsich teach wherein the first set of medical laboratory data indicates, with a degree of value imputation of 30% or less, eGFR, urine ACR, urea, potassium, hemoglobin, platelet count, albumin, calcium, glucose, bilirubin, sodium, bicarbonate, and GGT (see Hsich, see page 2, last paragraph through the top paragraph on page 3; the system could use informed imputation to fill in missing values, e.g. 10%; see Naylor, paragraphs [0194]-[0365] for entire list of examples and individual readings that can be measured and used, or have their respective values be imputed).
Claims 1-13 and 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over Cha et al [WO 2020/006571 A1] (from IDS) in view of Jain et al [US 2018/0055885 A1].
With regard to claim 1, Cha teaches a method, comprising: accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression (see page 6, lines 26-28; page 8, lines 8-13; the system can employ/access machine learning models to make predictions associated with CKD including calculation of risk scores),
wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing put data to generate the risk output (see page 8, lines 8-13; page 21, lines 13-19; page 52, lines 4-6 & 13-15; training data based on features from patient information can be used to train the machine learning model and generate a CKD progression prediction/risk),
wherein the training data set comprises: (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients (see page 17, lines 9-13; page 18, line 22 through col 19, line 2; demographic information and lab test information can be used as training data set),
and (iv) CKD clinical outcomes associated with the plurality of patients (see page 47, lines 8-12; page 21, lines 13 through page 24, line 25; and page 31, lines 1-18; the system has means to utilize CKD clinical outcomes as means part of the training/validation process for the machine learning model),
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea, serum sodium, serum chloride,
applying an input dataset associated with a new patient to the machine learning model to cause the machine learning model to process the input dataset via the model parameters of the machine learning model to generate risk score output indicating CKD progression for the new patient, the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient (see page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used including patient demographics and various other diagnostic tests).
Cha teaches various laboratory data but does not appear to explicitly teach:
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: serum hemoglobin, glucose, alkaline phosphatase, aspartate aminotransferase (AST), alanine transaminase (ALT), bilirubin, gamma-glutamyl transferase (GGT), hematocrit, and platelet count;
the input dataset comprising: eGFR, urine ACR, urea, serum sodium, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase (ALKP), serum phosphate, serum bicarbonate, serum magnesium, serum calcium, AST, ALT, bilirubin, GGT, hematocrit, and platelet count.
Jain teaches the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: serum hemoglobin, glucose, alkaline phosphatase, aspartate aminotransferase (AST), alanine transaminase (ALT), bilirubin, gamma-glutamyl transferase (GGT), hematocrit, and platelet count (see Table 8 in paragraphs [0420]-[0427] and [0134]; various parameters/variables that are considered and collected/measured from laboratory assessments/tests).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the feature data that is used for training and making predictions as taught by Cha by including other feature data from various medical lab and test results as taught by Jain in order to provide more relevant data so that the models can have more information that help identify amount of risk of the disease so that the trained model can be more accurate when being used since accuracy of medical diagnoses is a high-priority for both doctors (and other medical personnel) and the respective patients.
Cha in view of Jain teach the input dataset comprising: eGFR, urine ACR, urea, serum sodium, serum chloride, serum hemoglobin, serum potassium, glucose, serum albumin, alkaline phosphatase (ALKP), serum phosphate, serum bicarbonate, serum magnesium, serum calcium, AST, ALT, bilirubin, GGT, hematocrit, and platelet count (see Cha, page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; page 36, lines 1-16; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used from the patient’s records to determine their risk score; see Cha page 4, lines 1-7; page 18, line 22 through col 19, line 2; for examples of the input data that can be utilized to make a prediction or determine a risk score and also see Jain, paragraphs [0420]-[0427] and [0134]).
With regard to claim 2, Cha in view of Jain teach wherein the new patient is not associated with a CKD stage of G3 or later (see Cha, page 51, lines 7-12; the patients did not have any previous kidney disease diagnoses).
With regard to claim 3, Cha in view of Jain teach wherein the machine learning model comprises a random survival forest model (see Cha, page 51, lines 7-12; random forest ML models were employed to predict risk).
With regard to claim 4, Cha in view of Jain teach wherein the risk score output indicating CKD progression for the new patient indicates a risk of experiencing CKD progression within a particular amount of time from a time period associated with the input dataset for the new patient (see Cha, page 51, lines 17-18; the system can make a prediction of patients who would experience, i.e. have a risk of experiencing, a decline over a particular time period).
With regard to claim 5, Cha in view of Jain teach wherein the particular amount of time is provided as input to the machine learning model for generating the risk score output indicating CKD progression for the new patient (see Cha, page 51, line 17-18; page 21, lines 21-31; the system can make predictions based on a desired time period).
With regard to claim 6, Cha in view of Jain teach wherein the particular amount of time comprises 2 years or 5 years (see Cha, page 51, line 17-18; the system can make predictions based on a desired time period).
With regard to claim 7, Cha in view of Jain teach wherein the urine ACR for one or more of the plurality of patients or the new patient is converted from a urine protein-to -creatinine test or a urine dipstick test (see Jain, paragraph [0244]; urine dipstick test can be used as means to collect the measurements).
With regard to claim 8, Cha in view of Jain teach wherein the risk score output indicating CKD progression for the new patient indicates a risk of the new patient experiencing kidney failure or a 40% or greater decline of the eGFR for the new patient (see Cha, page 31, lines 9-15; the system can determine CKD progression based on a 40% decline).
With regard to claim 9, Cha in view of Jain teach wherein the risk of the new patient experiencing kidney failure comprises an indication that the new patient is at risk of (i) requiring chronic dialysis, (ii) requiring a kidney transplant, or (iii) experiencing a glomerular filtration rate of less than 10 ml/min/1.73m2 (see Cha, page 8, lines 19-23; page 49, lines 2-9; the system can evaluate and predict risk of a patient of various events including kidney failure).
With regard to claim 10, Cha in view of Jain teach determining that the risk score output indicating CKD progression for the new patient indicates a predicted risk of the new patient experiencing CKD within a particular time period that satisfies one or more predicted risk threshold values; and (i) generating a notification that the new patient may need an interventive kidney treatment; (ii) generating a recommendation of an interventive kidney treatment for the new patient based on the risk score output indicating CKD progression for the new patient; (iii) generating a recommendation of a frequency of monitoring of CKD progression for the new patient based on the risk score output indicating CKD progression for the new patient; or (iv) administering an interventive kidney treatment to the new patient (see Cha, page 4, lines 8-24; see page 6, lines 26-28; the system can make a risk score for the patient over a time period and be able to determine a treatment recommendation and transmit a notification of the treatment recommendation).
With regard to claim 11, Cha in view of Jain teach wherein the one or more predicted risk threshold values are based upon the particular time period associated with the risk score output indicating CKD progression for the new patient (see Cha, page 25, line 30 through page 26, line 2; the threshold values can be based or associated with a time period).
With regard to claim 12, Cha in view of Jain teach wherein the recommendation of the interventive kidney treatment or the recommendation of the frequency of monitoring of CKD progression is further based upon at least some of the second set of medical laboratory data associated with the new patient (see Cha, page 25, lines 3-12; the patient information or second data records are utilized to determine the risk score for the patient and what workflow actions should be performed including providing any notifications of recommendations).
With regard to claim 13, Cha in view of Jain teach wherein the interventive kidney treatment comprises one or more of: renin-angiotensin-aldosterone system (RAAS) inhibition, blood pressure control, sodium-glucose cotransporter-2 (SGLT2) inhibitor medication, mineralocorticoid receptor antagonists (MRAs) therapy, or preparation for nephrology consultation, home dialysis, dialysis access, or kidney transplant (see Cha, page 26, lines 18-21; Figure 2; various treatment options are available).
With regard to claim 16, this claim is substantially similar to claim 1 and is rejected for similar reasons as discussed above.
With regard to claim 17, this claim is substantially similar to claim 3 and is rejected for similar reasons as discussed above.
With regard to claim 18, Cha teaches a method, comprising: accessing a machine learning model configured to generate risk score output indicating chronic kidney disease (CKD) progression (see page 6, lines 26-28; page 8, lines 8-13; the system can employ/access machine learning models to make predictions associated with CKD including calculation of risk scores),
wherein the machine learning model is generated by applying a training dataset to an untrained machine learning model to configure model parameters of the machine learning model for processing input data to generate the risk output (see page 8, lines 8-13; page 21, lines 13-19; page 52, lines 4-6 & 13-15; training data based on features from patient information can be used to train the machine learning model and generate a CKD progression prediction/risk),
wherein the training data set comprises a first set of medical laboratory data associated with a plurality of patients, the first set of medical laboratory data including an age and sex of each patient included in the plurality of patients (see page 17, lines 9-13; page 18, line 22 through col 19, line 2; demographic information and lab test information can be used as training data set),
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: urine albumin-to-creatinine ratio (ACR), estimated glomerular filtration rate (eGFR), urea,
receiving user input indicating a time period input from a plurality of available time period inputs (see first paragraph on page 54 and Figure 6 and page 44, lines 18-27 and page 32, lines 5-8; the system allows for prediction time periods to be used to form predictions including prediction start date and prediction end date to from a prediction time period);
applying (i) the time period input and (ii) an input dataset associated with a new patient to the machine learning model such that the machine learning model processes the time period input together with the input dataset via the model parameters of the machine learning model to generate risk score output indicating CKD progression for the new patient based on both (i) the time period input and (ii) the input dataset associated with the new patient, the input dataset comprising patient information including at least an age and sex of the new patient, and a second set of medical laboratory data indicating for the new patient (see page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; page 32, lines 5-8; see Figure 1, box 145; see Figure 6; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used including patient demographics and various other diagnostic tests including being able to have an analysis for a particular prediction time period).
Cha teaches various laboratory data but does not appear to explicitly teach:
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: glucose, hematocrit, platelet count and hemoglobin;
the input dataset comprising…a second set of medical laboratory data indicating for the new patient: urine ACR, eGFR, urea, glucose, hematocrit, platelet count and hemoglobin for the new patient.
Jain teaches the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: glucose, hematocrit, platelet count and hemoglobin (see Table 8 in paragraphs [0420]-[0427] and [0134]; various parameters/variables that are considered and collected/measured from laboratory assessments/tests).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the feature data that is used for training and making predictions as taught by Cha by including other feature data from various medical lab and test results as taught by Jain in order to provide more relevant data so that the models can have more information that help identify amount of risk of the disease so that the trained model can be more accurate when being used since accuracy of medical diagnoses is a high-priority for both doctors (and other medical personnel) and the respective patients.
Cha in view of Jain teach the input dataset comprising…a second set of medical laboratory data indicating for the new patient: urine ACR, eGFR, urea, glucose, hematocrit, platelet count and hemoglobin for the new patient (see Cha, page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; page 36, lines 1-16; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used from the patient’s records to determine their risk score; see Cha page 4, lines 1-7; page 18, line 22 through col 19, line 2; for examples of the input data that can be utilized to make a prediction or determine a risk score and also see Jain, paragraphs [0420]-[0427] and [0134]).
With regard to claim 19, this claim is substantially similar to claim 3 and is rejected for similar reasons as discussed above.
Claims 14 and 15 are rejected under 35 U.S.C. 103 as being unpatentable over Cha et al [WO 2020/006571 A1] (from IDS) in view of Jain et al [US 2018/0055885 A1] in further view of Hsich et al, Identifying Important Risk Factors for Survival in Systolic Heart Failure Patients Using Random Survival Forests (from IDS).
With regard to claim 14, Cha in view of Jain teach all the claim limitations of claim 1.
Cha in view of Jain do not appear to explicitly teach wherein the first set of medical laboratory data comprises one or more imputed values in place of missing values.
Hsich teaches wherein the first set of medical laboratory data comprises one or more imputed values in place of missing values (see page 2, last paragraph through the top paragraph on page 3; the system could use informed imputation to fill in missing values).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the testing and data collection process as taught by Cha in view of Jain by incorporating imputation as taught by Hsich in order to ensure that data sets are complete and provide reasonable data values so that missing data from datasets doesn’t skew the analysis via having too few datasets with values to consider for initial training and potentially giving more undue weight to the non-missing features thus helping the prediction process of the machine learning model to be as informed and accurate as possible so that the most accurate predictions can be given to a patient.
With regard to claim 15, Cha in view of Jain in further view of Hsich teach wherein the first set of medical laboratory data indicates, with a degree of value imputation of 30% or less, eGFR, urine ACR, urea, potassium, hemoglobin, platelet count, albumin, calcium, glucose, bilirubin, sodium, bicarbonate, and GGT (see Hsich, see page 2, last paragraph through the top paragraph on page 3; the system could use informed imputation to fill in missing values, e.g. 10%; see Jain, paragraphs [0420]-[0427] and [0134] for entire list of examples and individual readings that can be measured and used, or have their respective values be imputed).
Claims 16, 17, and 21 are rejected under 35 U.S.C. 103 as being unpatentable over Cha et al [WO 2020/006571 A1] (from IDS) in view of Bradley et al [US 2023/0215575 A1].
With regard to claim 16, Cha teaches a system, comprising: one or more processors; and one or more hardware storage devices storing instructions that are executable by the one or more processors to configure the system to (see second to last paragraph on page 39 through last paragraph on page 40; the system can comprises processors and storage devices):
access a training dataset (see page 6, lines 26-28; page 8, lines 8-13; the system can employ/access machine learning models to make predictions associated with CKD including calculation of risk scores) comprising (i) a first set of medical laboratory data associated with a plurality of patients, (ii) an age of each patient included in the plurality of patients, (iii) a sex of each patient included in the plurality of patients (see page 17, lines 9-13; page 18, line 22 through col 19, line 2; demographic information and lab test information can be used as training data set),
and (iv) CKD clinical outcomes associated with the plurality of patients (see page 47, lines 8-12; page 21, lines 13 through page 24, line 25; and page 31, lines 1-18; the system has means to utilize CKD clinical outcomes as means part of the training/validation process for the machine learning model),
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: estimated glomerular filtration rate (eGFR), urine albumin-to-creatinine ratio (ACR), urea,
and generate one or more imputed values (see last paragraph on page 14 through line 15 on page 15; the system can generate imputed values);
generate a machine learning model by applying the imputed training dataset to an untrained machine learning model, wherein applying the imputed training dataset to the untrained machine learning model configures model parameters of the machine learning model for processing input data to generate the risk output indicating chronic kidney disease (CKD) progression for a new patient by applying an input dataset associated with the new patient to the machine learning model (see page 8, lines 8-13; page 21, lines 13-19; page 52, lines 4-6 & 13-15; training data based on features from patient information can be used to train the machine learning model and generate a CKD progression prediction/risk),
the input dataset comprising an age of the new patient, a sex of the new patient, and a second set of medical laboratory data indicating for the new patient (see page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; page 36, lines 1-16; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used including patient demographics and various other diagnostic tests).
Cha teaches various laboratory data but does not appear to explicitly teach:
the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: serum hemoglobin, glucose, platelet count, and hematocrit;
and generate one or more imputed values corresponding to one or more missing values in the first set of medical laboratory data to construct an imputed training dataset;
and generate a machine learning model by applying the imputed training dataset to an untrained machine learning model,
applying an input dataset associated with the new patient to the machine learning model, the input dataset comprising a second set of medical laboratory data indicating for the new patient: eGFR, urine ACR, urea, serum hemoglobin, glucose, platelet count, and hematocrit.
Bradley teaches the first set of medical laboratory data indicating, for at least a combination of patients included in the plurality of patients: serum hemoglobin, glucose, platelet count, and hematocrit (see paragraph [0047]; various biomarkers or medical laboratory data can be collected and used including hematocrit, hemoglobin, glucose, and platelet count);
and generate one or more imputed values corresponding to one or more missing values in the first set of medical laboratory data to construct an imputed training dataset (see paragraphs [0080], [0090], and [0074]-[0076]; the system can utilize imputation to fill out missing data values for the data set);
and generate a machine learning model by applying the imputed training dataset to an untrained machine learning model (see paragraphs [0007], [0068], [0090] and [0074]-[0076]; the machine learning model can be trained using the imputed training dataset).
It would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to modify the testing and data collection process as taught by Cha by incorporating imputation as taught by Bradley in order to ensure that data sets are complete and provide reasonable data values so that missing data from datasets doesn’t skew the analysis via having too few datasets with values to consider for initial training and potentially giving more undue weight to the non-missing features thus helping the prediction process of the machine learning model to be as informed and accurate as possible so that the most accurate predictions can be given to a patient.
Cha in view of Bradley teach applying an input dataset associated with the new patient to the machine learning model, the input dataset comprising a second set of medical laboratory data indicating for the new patient: eGFR, urine ACR, urea, serum hemoglobin, glucose, platelet count, and hematocrit (see Cha, page 25, lines 3-9; page 6, lines 21-25; page 2, lines 23-31; see Figure 1, box 145; page 36, lines 1-16; the system can utilize the machine learning model with input data for new patients where various pieces of data can be used from the patient’s records to determine their risk score; see Cha page 4, lines 1-7; page 18, line 22 through col 19, line 2; for examples of the input data that can be utilized to make a prediction or determine a risk score and also see Bradley, paragraph [0047] for other laboratory values).
With regard to claim 17, Cha in view of Bradley teach wherein the machine learning model comprises a random survival forest model, and wherein the one or more imputed values are generated using adaptive tree imputation (see Cha, page 51, lines 7-12; see Bradley, paragraph [0076]; random forest ML models were employed to predict risk and MissForest is an adaptive tree imputation technique for determining imputed values).
With regard to claim 21, Cha in view of Bradley teach wherein applying the imputed training dataset to the untrained machine learning model causes the machine learning model to exhibit a variable importance ranking in which urine ACR, eGFR, urea, and serum hemoglobin are ranked higher than hematocrit and glucose (see Cha, Tables 5 and 6 on page 50; and last paragraph on page 14 through line 15 on page 15; eGFR, urea, and urine ACR are given higher weights than other features since they are top model features where the system has feature-specific weights for each of the features where the system will learn the important features and assign greater weight to those features).
Response to Arguments
Applicant's arguments (see the third paragraph on page 8 through the second paragraph on page 10) have been fully considered but they are not persuasive. The applicant argues that the steps don’t recite mental steps but ‘accessing a machine learning model’ where that model is trained to generate the risk output. The Examiner respectfully disagrees. The Examiner notes that evaluating information and forming a decision (i.e. risk output) is a mental process step, albeit current claims indicate the usage of a computer to do the evaluation/decision steps which amounts to apply-it type limitations of using the computer as a tool to perform the judicial exception. Additionally, performing medical diagnosis via evaluation/analysis of variables relates to methods of organizing human activity such as doctors/nurses in evaluating patients. The applicant argues (top of page 9) that the human mind is not equipped to perform these calculations; however, the claims recite at a high-level of receiving the data set and getting the result. Additionally, as noted in MPEP 2106.04(a)(2)(III)(C) indicates that a claim that requires a computer may still recite a mental process step, in particular, the usage of the computer as a tool to perform a mental process. As for applicant’s argument (second paragraph of page 9) about organizing human activity argument, the Examiner notes that a mental process that a neurologist should follow when testing a patient for nervous system malfunctions is recited as certain methods of organizing human activity, where, although different type of doctor and evaluation is conducted, the concept of someone in the medical field evaluating information to determine kidney decline/malfunction are substantially related. With regards to ordered combination (see second to last paragraph on page 9 through second paragraph on page 10), that analysis occurs in step 2A, prong 2; however, as noted above, the claims recite the steps at a high-level of generality including listing the various variables as part of the dataset being used. As indicated before, merely collecting data at a high-level of generality is an insignificant extrasolution activity. There is no discussion of means to collect the data (thus standard biomedical data collection means can be used) nor how that data is utilized except by saying that it is applied and used by a model, which is similar to how a human would interact with respective data including evaluating the data and determining which data points are more important in their evaluation than others. Therefore, in view of the high-level of generality of the claim limitations, applicant’s arguments are not persuasive.
Applicant's arguments (see the second to last paragraph on page 10 through the second to last paragraph on page 11) have been fully considered but they are not persuasive. The applicant argues that the Office Action errs with the Declaration in characterizing applicant’s position as attorney arguments. The Examiner respectfully disagrees. As discussed in the Office Action, the expert opinion was evaluated including strength of evidence and interest of the expert in the outcome of the case as well as the presence or absence of factual support for the expert’s opinion. Although the Examiner cited Jain, the Examiner also indicated that having access to variables and assigning zero or near zero weight to those variables doesn’t preclude those variables from being considered. Additionally, as noted above, Bradley also illustrates various parameters/variables that were used for CKD. Accordingly, applicant’s arguments are not persuasive.
Applicant's arguments (see the last paragraph on page 11) have been fully considered but they are not persuasive. The applicant argues that that the random survival forest model is not a generic machine learning model but a particular machine learning model that is operated in a particular manner. The Examiner respectfully disagrees. Claim 3 recites the usage of a random survival forest model. No other details are present and this constitutes a label for a machine learning algorithm. The Examiner also notes that random survival forest models were not invented by applicant at the time of the invention and claim 3 merely recites that the model is a random survival forest model at a high-level of generality with no other discussions or details. Merely reciting a particular known machine learning model type with no other details adds no meaningful limitation beyond that of the abstract idea as discussed above. The Examiner notes that using generically claimed named algorithms is no different than using a generic processor or memory to run the abstract idea. It does not cause the claim to be significantly more than the abstract idea, and therefore the rejection is maintained as shown above.
Applicant's arguments (see the first whole paragraph on page 12) have been fully considered but they are not persuasive. The applicant argues that claim 4 recites a risk score output that indicates a risk of experiencing CKD progression within a particular amount of time for the new patient and claim 5 indicates that the particular amount of time is provided as input to the ML model for generating the risk output score which applicant asserts is not merely receiving information since the input is provided to the ML model as input to generate the risk score output. The Examiner respectfully disagrees. With regard to claims 4 and 5, as discussed with respect to claim 1, the prediction/judgement is part of the judicial exception based on the evaluation of the patient’s data to form a prediction where such predictions can be based on a future time period, e.g. how a disease progresses over time. This technique is called forecasting. Although claim 5 indicates that the input is provided to the model, as noted above, the judicial exception of using the data to generate a risk score output for the CKD progression for the new patient is recited as a judicial exception where a human can evaluate the data and extrapolate/predict/forecast for various future dates. Accordingly, applicant’s arguments are not persuasive.
Applicant's arguments (see the second to last whole paragraph on page 12 through third paragraph on page 14) have been fully considered but they are not persuasive. The applicant argues similar arguments regarding the other independent claims while also arguing newly amended limitations. Accordingly, the respective 35 USC 101 rejections were updated to address the amended limitations. Overall, as noted above, the recitation of generating a machine learning model is similar to concepts of usage of the computer as a tool (i.e. machine learning model) to perform the judicial exception which were discussed above. With regards to the generated imputed values limitations, as noted in the 35 USC 101 rejections above, this relates to statistical analysis which is a mental process step involving mathematical calculations. Therefore, applicant’s arguments are not persuasive.
Applicant's arguments (see the last paragraph on page 14) have been fully considered but they are not persuasive. The applicant argues that the newly amended limitation should not have a 35 USC 101 rejection for a variety of reasons. The Examiner respectfully disagrees. The claim indicates that certain variables are weighted as more important than others which the Examiner notes relates to mental process steps of analysis/evaluating data where human users can mentally bias particular traits/attributes/variables when forming a judgement/decision. Therefore, applicant’s argument is not persuasive.
Applicant's arguments (see the second whole paragraph on page 15 through the last paragraph on page 16) have been fully considered but they are not persuasive. The applicant argues similar arguments regarding the other independent claims while also arguing newly amended limitations. Accordingly, the respective 35 USC 101 rejections were updated to address the amended limitations. Overall, as noted above, the recitation of generating a machine learning model is similar to concepts of usage of the computer as a tool (i.e. machine learning model) to perform the judicial exception which were discussed above. With regards to the receiving user input about a time period and using that time period as an input to the machine learning model, the Examiner notes that the recitation merely indicates additional data that is considered where the usage of the machine learning model to form a prediction or ‘risk score output’ of CKD progression recites the judicial exception. The time period input being received and used is recited at a high-level of generality that further elaborates on statistical methods such as future forecasting which relates to mental process steps involving mathematical calculations. The respective rejection has been updated to reflect the newly amended limitation. Therefore, applicant’s arguments are not persuasive.
Applicant's arguments (see the first paragraph on page 17 through the last paragraph on page 18) have been fully considered but they are not persuasive. The applicant argues the Naylor reference is not analogous art since it isn’t pertinent to the field of inventors endeavor or reasonably pertinent to the problem. The Examiner respectfully disagrees. In response to applicant's argument that Naylor is nonanalogous art, it has been held that a prior art reference must either be in the field of the inventor’s endeavor or, if not, then be reasonably pertinent to the particular problem with which the inventor was concerned, in order to be relied upon as a basis for rejection of the claimed invention. See In re Oetiker, 977 F.2d 1443, 24 USPQ2d 1443 (Fed. Cir. 1992). In this case, as noted by applicant, the Cha reference relates to the field of endeavor with discussion of non-limiting variables and tests that are considered but rather other data can be considered too ((see Cha, page 2, lines 23-31; page 4, lines 1-13; page 9, lines 15-22; page 18, line 27 through page 19, line 7). Thus, Cha illustrates that "any number of features relating to lab tests information" can be used/employed where Cha provides examples of the features with Naylor illustrating other features that relate to the human body that can be considered including features from lab tests. As noted in the 35 USC 103 rejections, Naylor would be considered analogous art since Naylor relates to the acquisition of particular medical data. As noted in the Office Action, Cha did not limit itself to only a finite set of data to only ever be considered but allowed for consideration of other data, where, someone of ordinary skill in the art when viewing Cha’s teachings, that wanted other medical data to consider, even if for statistical analysis to see if any correlation exists, would look at other medical-related references to determine what variables and respective tests would and/or should be gathered/performed. As noted in the 35 USC 103 rejections, Naylor modifies Cha's teaching to illustrate the usage of additional test data that can be used/considered by Cha's system which seeks to use all relevant test data (see page 52, lines 13-18). Therefore, Naylor is in the field of inventor's endeavor in that Naylor provides various medical data that can be accessed and used for analysis, thus making Naylor reasonably pertinent to the problem of medical data analysis via acquiring medical data variables via various widely used and known medical tests.
Applicant's arguments (see the first paragraph on page 19 through the second paragraph on page 22) have been fully considered but they are not persuasive. The applicant argues that even if the references were analogous, the references fail to establish a prima facie case of obviousness. The Examiner respectfully disagrees. The Examiner indicates that the Cha reference does not recite particular variable/lab test usage in their methodology and thus one cannot add any other collected lab results/tests/variables into Cha’s teachings. However, as discussed above, Cha provides examples of various tests and variables being used and indicates that it is open to other information as well. In other words, Cha doesn’t restrict itself and that the reference is open to a variety of other information to be used by their system (see page 18, last paragraph: “The system may further employ any number of features relating to lab tests information”) where, as noted above, the Bradley reference illustrates that the various variables not mentioned in Cha are used in the prior art for chronic kidney disease. Thus, one of ordinary skill in the art, looking at Cha and wanting other known medical data/tests, would be motivated to look at other teachings of acquiring and collecting the desired information that they want. Therefore, applicant’s arguments and, as discussed previously, applicant’s Declaration, are not persuasive for at least the reasons discussed above and discussed in the prior Office Action.
Applicant's arguments (see the last paragraph on page 22) have been fully considered but they are not persuasive. The applicant argues that claim 18 recites additional limitations associated with the time period input that is not taught by the cited prior art references. The Examiner respectfully disagrees. The applicant points to certain sections of the Cha reference to conclude that the limitation is not taught. Applicant's arguments fail to comply with 37 CFR 1.111(b) because they amount to a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them from the references. As illustrated throughout Cha, various discussions of predictions and time periods are used including that a model can have various model information as input including prediction window information. Therefore, applicant’s arguments are not persuasive.
Applicant's arguments (see the first whole paragraph on page 23) have been fully considered but they are not persuasive. The applicant argues that new claim 21 is not taught by the cited prior art references. The Examiner respectfully disagrees. Applicant's arguments fail to comply with 37 CFR 1.111(b) because they amount to a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them from the references. Since claim 21 is a new claim and was not previously mapped, the 35 USC 103 rejection has been updated to illustrate how the claim is being taught by the cited prior art references.
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.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MARC S SOMERS whose telephone number is (571)270-3567. The examiner can normally be reached M-F 11-8 EST.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Ann Lo can be reached at 5712729767. 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.
/MARC S SOMERS/Primary Examiner, Art Unit 2159 8/21/2026
Prosecution Insights