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 .
Priority
Acknowledgment is made of applicant’s claim for foreign priority under 35 U.S.C. 119 (a)-(d). The certified copy has been filed in parent Application No. EP20211555.6, filed on November 26, 2020.
Information Disclosure Statement
The information disclosure statement(s) (IDS) submitted on May 3, 2023, June 28, 2023, and June 11, 2024 is/are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement(s) is/are being considered by the examiner.
Claim Objections
Claim 3 is objected to because the claim recites "identifying a user who training the anonymized subject". However, "training" seems to be a typographical error. For example, paragraph [0083] of the specification states "a first user (e.g., doctor)", and thus a more appropriate term for this limitation would be treated, treating, or treats. Examiner will construe the limitation as "identifying a user who treated/treats the anonymized subject".
Claim Rejections - 35 USC § 112
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claim 12 is rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding Claim 12, the claim recites “the data leakage prevention protocol”. The term lacks antecedent basis because “a data leakage prevention protocol” is introduced in claim 11, not claim 1. For purposes of examination, claim 12 will be construed as depending from claim 11.
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 1-20 are rejected under 35 U.S.C.101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1: Claims 1-13 are directed to a process. Claims 14-20 are directed to a machine or an article of manufacture.
With respect to claim(s) 1, 14, and 15:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
generating a partial word sequence by combining the subset of the set of features into a sequence of one or more words, each word of the one or more words representing a feature of the subset of features; (Mental process – A person can mentally think of (e.g., generate) a combination of word sequences or use the physical aid of pen and paper to generate the combination of words – see MPEP § 2106.04(a)(2)(III))
transforming the partial word sequence into a numerical representation […]; (Mental process – A person can mentally transform a word into a numerical representation or by using the physical aid of a pen and paper – see MPEP § 2106.04(a)(2)(III))
generating, based on the completion word or phrase outputted by the NLP model, a disease progression representing a predicted progression of one or more SMA phenotypes specific to the subject over a period of time; (Mental process– A person can mentally generate (think of) a disease progression or with the physical aid of a pen and paper – see MPEP § 2106.04(a)(2)(III))
If claim limitations, under their broadest reasonable interpretation, cover performance of the limitations as a mental process, but for the recitation of generic computer components, then the claim limitations fall within the mathematical or mental process grouping of abstract ideas. Accordingly, the claim “recites” an abstract idea.
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
(Claim 1) A computer-implemented method comprising: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
(Claim 14) one or more processors; (Mere recitation of a generic computer component – see § MPEP 2106.05(b)(I))
(Claim 14) a non-transitory, computer-readable storage medium containing instructions which, when executed on the one or more processors, cause the one or more processors to perform a set of actions including: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
(Claim 15) A computer-program product tangibly embodied in a non-transitory, machine-readable storage medium, including instructions configured to cause one or more data processors to perform a set of actions including: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
retrieving a subject record associated with a subject, the subject record including a set of features characterizing the subject, and the subject having been diagnosed with spinal muscular atrophy (SMA); (Mere data gathering – Adding insignificant extra-solution activity of mere data gathering to the judicial exception – see § MPEP2106.05(g).)
extracting a subset of the set of features included in the subject record,each feature of the subset of the set of features being associated with an SMA characteristic; (Mere data gathering – Adding insignificant extra-solution activity of mere data gathering to the judicial exception – see § MPEP2106.05(g).)
[…] using a trained word-to-vector model (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
inputting the numerical representation of the partial word sequence into a natural language processing (NLP) model having been trained to predict a completion word or phrase for completing the partial word sequence; (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
outputting an indication that the subject is predicted to exhibit the one or more SMA phenotypes included in the disease progression. (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
Since the claim as a whole, looking at the additional elements individually and in combination, does not contain any other additional elements that are indicative of integration into a practical application, the claim is directed to an abstract idea.
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
(Claim 1) A computer-implemented method comprising: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
(Claim 14) one or more processors; (Mere recitation of a generic computer component – see § MPEP 2106.05(b)(I))
(Claim 14) a non-transitory, computer-readable storage medium containing instructions which, when executed on the one or more processors, cause the one or more processors to perform a set of actions including: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
(Claim 15) A computer-program product tangibly embodied in a non-transitory, machine-readable storage medium, including instructions configured to cause one or more data processors to perform a set of actions including: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
retrieving a subject record associated with a subject, the subject record including a set of features characterizing the subject, and the subject having been diagnosed with spinal muscular atrophy (SMA); (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
extracting a subset of the set of features included in the subject record, each feature of the subset of the set of features being associated with an SMA characteristic; (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
[…] using a trained word-to-vector model (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
inputting the numerical representation of the partial word sequence into a natural language processing (NLP) model having been trained to predict a completion word or phrase for completing the partial word sequence; (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
outputting an indication that the subject is predicted to exhibit the one or more SMA phenotypes included in the disease progression. (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
Considering the additional elements individually and in combination, and the claim as a whole, the additional elements do not provide significantly more than the abstract idea. Therefore, the claim is not patent eligible.
With respect to claim(s) 2 and 16:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
determining that the predicted progression of the one or more SMA phenotypes specific to the subject satisfies an early treatment condition, (Mental process – A person can mentally determine that an early treatment condition is satisfied – see MPEP § 2106.04(a)(2)(III))
wherein satisfying the early treatment condition is indicative of a recommendation to perform a treatment before the subject exhibits an SMA phenotype of the one or more SMA phenotypes. (Mental process – A person can mentally evaluate an early treatment condition being satisfied to indicate a recommendation to perform treatment – see MPEP § 2106.04(a)(2)(III))
Additionally, the claim(s) do not recite any new additional elements that would amount to an integration of the abstract idea into a practical application (individually or in combination) or significantly more than the judicial exception.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 3 and 17:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
when the predicted progression of the one or more SMA phenotypes satisfies the early treatment condition: (Mental process – A person can mentally determine a condition is satisfied – see MPEP § 2106.04(a)(2)(III))
identifying an existing disease progression associated with an anonymized subject, the existing disease progression matching the predicted progression of the one or more SMA phenotypes specific to the subject, and the anonymized subject having been diagnosed with SMA; (Mental process – A person can mentally identify the disease progression – see MPEP § 2106.04(a)(2)(III))
identifying a user who training (treated?) the anonymized subject associated with the existing disease progression; (Mental process – A person can mentally identify a user – see MPEP § 2106.04(a)(2)(III))
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
transmitting a communication to a user device associated with the user, the communication requesting treatment recommendations for the subject. (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
transmitting a communication to a user device associated with the user, the communication requesting treatment recommendations for the subject. (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 4 and 18:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
when the predicted progression of the one or more SMA phenotypes does not satisfy the early treatment condition: (Mental process – A person can mentally determine a condition is not satisfied – see MPEP § 2106.04(a)(2)(III))
identifying an existing disease progression associated with an anonymized subject, the existing disease progression matching the predicted progression of the one or more SMA phenotypes specific to the subject, and the anonymized subject having been diagnosed with SMA (Mental process – A person can mentally identify and match the disease progression – see MPEP § 2106.04(a)(2)(III))
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
retrieving an anonymized subject record characterizing the anonymized subject; (Mere data gathering – Adding insignificant extra-solution activity of mere data gathering to the judicial exception – see § MPEP2106.05(g).)
extracting a treatment schedule from the anonymized subject record; and (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
transmitting the treatment schedule to a user device. (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
retrieving an anonymized subject record characterizing the anonymized subject; (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
extracting a treatment schedule from the anonymized subject record; and (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
transmitting the treatment schedule to a user device. (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 5 and 19:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
matching the completion word or phrase associated with the subject to another one or more SMA phenotypes associated with another subject having been previously treated for SMA; (Mental process – A person can mentally match the completion word or phrase to SMA phenotypes – see MPEP § 2106.04(a)(2)(I))
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
retrieving an anonymized subject record characterizing the other subject; (Mere data gathering – Adding insignificant extra-solution activity of mere data gathering to the judicial exception – see § MPEP2106.05(g).)
extracting a treatment schedule from the anonymized subject record; (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
transmitting the treatment schedule to a user device. (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
retrieving an anonymized subject record characterizing the other subject; (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
extracting a treatment schedule from the anonymized subject record; (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
transmitting the treatment schedule to a user device. (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 6 and 20:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
wherein the completion word or phrase is predicted as a next word in a complete word sequence including the partial word sequence, and wherein the completion word or phrase represents an SMA phenotype (Mental process – A person can mentally predict a next word – see MPEP § 2106.04(a)(2)(I))
Additionally, the claim(s) do not recite any new additional elements that would amount to an integration of the abstract idea into a practical application (individually or in combination) or significantly more than the judicial exception.
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 7:
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
wherein the disease progression is output at a computing device of the subject using a chatbot. (Adding insignificant extra-solution activity to the judicial exception – see § MPEP2106.05(g).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the disease progression is output at a computing device of the subject using a chatbot. (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 8:
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
wherein the subject record includes data identified in an electronic medical record corresponding to the subject. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the subject record includes data identified in an electronic medical record corresponding to the subject. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 9:
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
wherein the subject record corresponding to the subject includes a diagnosis of SMA Type-I, SMA Type-II, SMA Type III, or SMA Type-IV. (Generally linking the use of a judicial exception to a particular technological environment or field of use – see MPEP § 2106.05(h).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein the subject record corresponding to the subject includes a diagnosis of SMA Type-I, SMA Type-II, SMA Type III, or SMA Type-IV. (Generally linking the use of a judicial exception to a particular technological environment or field of use – see MPEP § 2106.05(h).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 10:
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
wherein training the NLP model further comprises: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
collecting a training data set including a set of subject records, each subject record of the set of subject records corresponding to another subject diagnosed with SMA, and each subject record of the set of subject record including one or more features representing a progression of SMA phenotypes during a time period; (Mere data gathering – Adding insignificant extra-solution activity of mere data gathering to the judicial exception – see § MPEP2106.05(g).)
executing a learning algorithm associated with a generative sequence model using the training data set, wherein the learning algorithm detects patterns associated with the progression of SMA phenotypes exhibited by a set of subjects corresponding to the set of subject records; and (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
generating the NLP model in response to executing the learning algorithm associated with the generative sequence model using the training data set. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein training the NLP model further comprises: (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
collecting a training data set including a set of subject records, each subject record of the set of subject records corresponding to another subject diagnosed with SMA, and each subject record of the set of subject record including one or more features representing a progression of SMA phenotypes during a time period; (Simply appending well-understood, routine, conventional activities previously known to the industry, specified at a high level of generality, to the judicial exception (WURC)- see MPEP § 2106.05(d)(ll)(i) - Receiving or transmitting data over a network, e.g., using the Internet to gather data, Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information).)
executing a learning algorithm associated with a generative sequence model using the training data set, wherein the learning algorithm detects patterns associated with the progression of SMA phenotypes exhibited by a set of subjects corresponding to the set of subject records; and (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
generating the NLP model in response to executing the learning algorithm associated with the generative sequence model using the training data set. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 11:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
detecting data leakage associated with the NLP model, the data leakage exposing a feature of the set of features included in the subject record characterizing the subject; (Mental process – A person can mentally detect data leakage by determining if records were exposed – see MPEP § 2106.04(a)(2)(II))
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
in response to detecting data leakage associated with the NLP model, executing a data leakage prevention protocol that prevents or blocks exposure of the feature of the set of features included in the subject record. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
in response to detecting data leakage associated with the NLP model, executing a data leakage prevention protocol that prevents or blocks exposure of the feature of the set of features included in the subject record. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 12:
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
wherein executing the data leakage prevention protocol includes re-training the NLP model according to a differential privacy model. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
wherein executing the data leakage prevention protocol includes re-training the NLP model according to a differential privacy model. (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
With respect to claim(s) 13:
2A Prong 1: The claim(s) recite(s) an abstract idea. Specifically:
generating […] a reduced-dimensionality subject record characterizing the subject, the reduced-dimensionality subject record removing one or more features from the set of features included in the subject record, the one or more features being characterized as noise. (Mental process – A person can reduce the dimensionality of a subject record by removing features using the physical aid of a pen and paper – see MPEP § 2106.04(a)(2)(I))
2A Prong 2: The additional elements recited in the claim(s) do not integrate the abstract idea into a practical application, individually or in combination.
Additional elements:
[…] using a feature selection model […] (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
2B: The claim(s) do(es) not include additional elements that are sufficient to amount to significantly more than the judicial exception.
Additional elements:
[…] using a feature selection model […] (Mere instructions to apply an abstract idea on a computer, or merely uses a computer as a tool to perform an abstract idea – see MPEP 2106.05(f).)
Since the claim does not recite additional elements that either integrate the judicial exception into a practical application, nor provide significantly more than the judicial exception, the claim is not patent eligible. Therefore, the claim is not patent eligible.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1, 6, 8-10, 13-15, and 20 are rejected under 35 U.S.C. 103 as being unpatentable over CHOI ("Doctor AI: Predicting Clinical Events via Recurrent Neural Networks") in view of SRIVASTAVA ("Machine learning algorithms to classify spinal muscular atrophy subtypes"), hereafter CHOI and SRIVASTAVA respectively.
Regarding Claim 1:
CHOI teaches:
retrieving a subject record associated with a subject, the subject record including a set of features characterizing the subject, and the subject having been diagnosed […] (CHOI [page 4, section 3. Cohort] teaches: "As inputs, we use ICD-9 codes, medication codes, and procedure codes. We extracted ICD-9 codes (i.e., retrieving a subject record associated with a subject) from encounter records, medication orders, problem list records and procedure orders. Generic Product Identifier (GPI) medication codes and CPT procedure codes were extracted from medication orders and procedure orders respectively. All codes were timestamped with the patients visit time. If a patient received multiple codes in a single visit, those codes were given the same timestamp." CHOI [page 4, section 4. Methods] teaches: "For each patient, the observations are drawn from a multilabel point process in the form of
t
i
,
x
i
for
i
=
1
,
…
,
n
. Each pair represents an event, such as an ambulatory care visit, during which multiple medical codes such as ICD-9 diagnosis codes (i.e., and the subject having been diagnosed), procedure codes, or medication codes are documented in the patient record. (i.e., the subject record including a set of features characterizing the subject).")
extracting a subset of the set of features included in the subject record, each feature of the subset of the set of features being associated with an […] characteristic; (CHOI [page 4, section 3. Cohort] teaches: "Therefore, to predict diagnosis and medication order, we grouped codes into higher-order categories to reduce the feature set (i.e., extracting a subset of the set of features included in the subject record) and information overload. For the diagnosis codes, we use the 3-digit ICD-9 codes, yielding 1183 unique codes. For the medication codes, we use the Generic Product Identifier Drug Class, which groups the medication codes into 595 unique groups (i.e., each feature of the subset of the set of features being associated with an […] characteristic).")
generating a partial word sequence by combining the subset of the set of features into a sequence of one or more words, each word of the one or more words representing a feature of the subset of features; (CHOI [ page 15, Appendix B] teaches: "We processed the private dataset so that diagnosis codes, medication codes, procedure codes are laid out in a temporal order. If there are multiple codes at a single visit, they were laid out in a random order. Then using the context window size of 5 to the left and 5 to the right, and applying Skip-gram, we were able to project diagnosis codes, medication codes and procedure codes into the same lower dimensional space, where similar or related codes are embedded close to one another." Examiner's note: CHOI teaches ordering the codes for Skip-gram processing, which is a Word2Vec (word-to-vector) model that transforms text inputs into vector embeddings. Under broadest reasonable interpretation, a partial word sequence by combining the subset of the set of features into a sequence of one or more words can be interpreted as the laid out codes prior to Skip-gram processing. Each of these codes correspond to a specific feature of the private dataset (i.e., each word of the one or more words representing a feature of the subset of features).)
transforming the partial word sequence into a numerical representation using a trained word-to-vector model; (CHOI [ page 15, Appendix B] teaches: "We specifically used Skip-gram Mikolov et al. (2013) to learn real-valued multidimensional vectors to capture the latent representation of medical codes from the EHR." Examiner's note: CHOI teaches ordering the codes for Skip-gram processing, which is a Word2Vec (word-to-vector) model that transforms text inputs into vector embeddings (i.e., numerical representation).)
inputting the numerical representation of the partial word sequence into a natural language processing (NLP) model having been trained to predict a completion word or phrase for completing the partial word sequence; (CHOI [pg. 5, section 4. Methods] teaches: "Using effective patient representations, we are interested in predicting diagnosis and medication categories (i.e., a completion word or phrase) in the next visit
y
i
+
1
and the time duration until the next visit
d
i
+
1
=
t
i
+
1
-
t
i
. Finally, we would like to perform all these steps jointly in a single supervised learning scheme. We use RNN (i.e., inputting the numerical representation of the partial word sequence into a natural language processing (NLP) model) to learn such patient representations, treating the hidden layer as the representation for the patient status and use it for the prediction tasks (i.e., having been trained to predict a completion word or phrase for completing the partial word sequence)."CHOI [page 5, Figure 1] teaches: "Figure 1: This diagram shows how we have applied RNNs to solve the problem of forecasting of next visits' time and the codes assigned during each visit." Examiner's note: Paragraph [00196] describes using a generative sequence model (e.g., a natural language processing (NLP) model) to generate predictions of the next word. CHOI's RNN predicts diagnosis and medication categories, and thus under broadest reasonable interpretation, the NLP model can be interpreted as CHOI's RNN model.)
generating, based on the completion word or phrase outputted by the NLP model, a disease progression representing a predicted progression […] specific to the subject over a period of time; (CHOI [page 1, section 1. Introduction] teaches: "Leveraging large historical data in EHR, we developed Doctor AI, a generic predictive model that covers observed medical conditions and medication uses. Doctor AI is a temporal model using recurrent neural networks (RNN) and was developed and applied to longitudinal time stamped EHR data. […] Applications that accurately forecast could have many uses such as anticipating the patient status at the time of visit and presenting data a physician would want to see at the moment. The primary goal of this study was to use longitudinal patient visit records to predict the physician diagnosis and medication order of the next visit. As a secondary goal we predicted the time to the patients next visit. Predicting the visit time facilitates guidance of whether a patient may be delayed in seeking care.”)
outputting an indication that the subject is predicted to exhibit the one or more […] disease progression. (CHOI [page 1, Abstract] teaches: "Doctor AI assesses the history of patients to make multilabel predictions (one label for each diagnosis or medication category)."
CHOI is not relied upon for teaching, but SRIVASTAVA teaches: […] and the subject having been diagnosed with spinal muscular atrophy (SMA); […] features being associated with an SMA characteristic; […] one or more SMA phenotypes specific to the subject […]; […] one or more SMA phenotypes […] (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype." SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI and SRIVASTAVA before them, to include SRIVASTAVA’s spinal muscular atrophy (SMA) patient data in CHOI’s clinical event prediction method. One would have been motivated to make such a combination in order to identify biomarkers to improve the efficiency of neurologic disease clinical trials and develop stronger, more robust biomarkers for classification of SMA (SRIVASTAVA [page 358, Abstract] and [page 363, Discussion]).
Regarding Claim 6:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI further teaches:
wherein the completion word or phrase is predicted as a next word in a complete word sequence including the partial word sequence, and wherein the completion word or phrase represents an […] phenotype (CHOI [pg. 5, section 4. Methods] teaches: "Using effective patient representations, we are interested in predicting diagnosis and medication categories (i.e., a completion word or phrase) in the next visit
y
i
+
1
and the time duration until the next visit
d
i
+
1
=
t
i
+
1
-
t
i
. Finally, we would like to perform all these steps jointly in a single supervised learning scheme. We use RNN to learn such patient representations, treating the hidden layer as the representation for the patient status and use it for the prediction tasks (i.e., is predicted as a next word in a complete word sequence including the partial word sequence)."CHOI [page 5, Figure 1] teaches: "Figure 1: This diagram shows how we have applied RNNs to solve the problem of forecasting of next visits' time and the codes assigned during each visit." CHOI [page 3, section 2. Related Work] teaches: “Previous studies such as phenotype learning (Lasko et al., 2013; Che et al., 2015; Hammerla et al., 2015) or representation learning (Choi et al., 2016b,a; Miotto et al., 2016), however, have not fully addressed the sequential nature of EHR. Lipton et al. (2016) is especially related to our work in that both studies use RNN for sequence prediction.” CHOI [page 15, Appendix D] teaches: “For all five of the cases shown in Table 3, the set of predicted diseases contain most, if not all of the true diseases. For example, in the first case, the top 3 predicted diseases match the true diseases. A human doctor would likely predict similar diseases to the ones predicted with Doctor AI, since old myocardial infarction and chronic ischemic heart disease can be associated with infections and diabetes (Stevens et al., 1978).” Examiner's note: Paragraph [00196] describes using a generative sequence model (e.g., a natural language processing (NLP) model) to generate predictions of the next word. CHOI's RNN predicts diagnosis and medication categories, and thus under broadest reasonable interpretation, the NLP model can be interpreted as CHOI's RNN model using skip-gram. Furthermore, under broadest reasonable interpretation, phenotype can be interpreted as CHOI’s specific predicted disease, such as those discussed in Table 3.)
SRIVASTAVA further teaches: […] SMA phenotype. (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype.")
Regarding Claim 8:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI further teaches:
wherein the subject record includes data identified in an electronic medical record corresponding to the subject. (CHOI [page 1, Abstract] teaches: "Leveraging large historical data in electronic health record (EHR), we developed Doctor AI," CHOI [page 1, section 1. Introduction] teaches: "Intelligent clinical decision support anticipates the information at the point of care that is specific to the patient and provider needs. Electronic health records (EHR), now commonplace in U.S. healthcare, represent the longitudinal experience of both patients and doctors.”)
Regarding Claim 9:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. SRIVASTAVA further teaches:
wherein the subject record corresponding to the subject includes a diagnosis of SMA Type-I, SMA Type-II, SMA Type III, or SMA Type-IV. (SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
Regarding Claim 10:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI further teaches:
wherein training the NLP model further comprises: collecting a training data set including a set of subject records, each subject record of the set of subject records corresponding to another subject diagnosed [...], and each subject record of the set of subject record including one or more features representing a progression [...] during a time period; (CHOI [page 4, section 3. Cohort] teaches: "As inputs, we use ICD-9 codes, medication codes, and procedure codes. We extracted ICD-9 codes from encounter records, medication orders, problem list records and procedure orders. Generic Product Identifier (GPI) medication codes and CPT procedure codes were extracted from medication orders and procedure orders respectively. All codes were timestamped with the patients visit time. If a patient received multiple codes in a single visit, those codes were given the same timestamp (i.e., and each subject record of the set of subject record including one or more features representing a progression [...] during a time period). The resulting dataset consists of 263,706 patients who made on average 54.61 visits per person. [...] For example, pulmonary tuberculosis (ICD-9 code 011) is divided into 70 subcategories (ICD-9 code 011.01, 011.02, ..., 011.95, 011.96)." CHOI [page 6, section 5.1 Experiment Setup] teaches: “For training all models including the baselines, we used 85% of the patients as the training set and 15% as the test set.”)
executing a learning algorithm associated with a generative sequence model using the training data set, wherein the learning algorithm detects patterns associated with the progression [...] exhibited by a set of subjects corresponding to the set of subject records; (CHOI [page 5, section 4. Methods] teaches: "The proposed neural network architecture (Figure 1) receives input at each timestamp
t
i
as the concatenation of the multi-hot input vector
x
i
of the multilabel categories and the duration
d
i
since the last event. In our datasets, the input dimension is as large as 40,000. Thus, the next layer projects the input to a lower dimensional space. Then, we pass the lower dimensional vector through RNN (implemented with GRU in our study). We can also stack multiple layers of RNN to increase the representative power of the network. Finally, we use a Softmax layer to predict the diagnosis codes and the medication codes, and a rectified linear unit (ReLU) to predict the time duration until next visit." Examiner’s note: Under broadest reasonable interpretation, detects patterns associated with the progression can be interpreted as the network predicting diagnosis and medication codes based on the multi-hot input vectors.)
generating the NLP model in response to executing the learning algorithm associated with the generative sequence model using the training data set. (CHOI [page 6, section 5.1 Experiment Setup] teaches: "We trained the RNN models for 20 epochs (i.e., 20 iterations over the entire training data) and then evaluated the final performance against the test set.")
SRIVASTAVA further teaches: […] records corresponding to another subject diagnosed with SMA, and each subject record of the set of subject record including one or more features representing […] SMA phenotypes […]; […] SMA phenotypes exhibited by a set of subjects corresponding to the set of subject records; (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype." SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
Regarding Claim 13:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. SRIVASTAVA further teaches:
generating, using a feature selection model, a reduced-dimensionality subject record characterizing the subject, the reduced-dimensionality subject record removing one or more features from the set of features included in the subject record, the one or more features being characterized as noise. (SRIVASTAVA [page 363, Discussion] teaches: "Other feature selection methods could also be used to rank the features and identify the optimal set of features for the classification. One well-known technique is principal component analysis (PCA), which is a dimensionality reduction method that combines features to transform a set of observations of possibly intercorrelated variables into a set of values of uncorrelated variables that maximally account for the variance in the original data set." Examiner’s note: Under broadest reasonable interpretation, the one or more features being characterized as noise can be interpreted as the features that were not considered optimal and not used during the PCA feature transformation.)
Regarding Claim 14:
The claim recites similar limitations as corresponding claim 1 and is rejected for similar reasons as claim 1 using similar teachings and rationale. Additionally, CHOI teaches:
A system, comprising: one or more processors; and a non-transitory, computer-readable storage medium containing instructions which, when executed on the one or more processors, cause the one or more processors to perform a set of actions including: (CHOI [page 6, section 5.1 Experimental Setup] teaches: "All models were implemented with Theano (Bastien et al., 2012) and trained on a machine equipped with two Nvidia Tesla K80 GPUs.")
Regarding Claim 15:
The claim recites similar limitations as corresponding claim 1 and is rejected for similar reasons as claim 1 using similar teachings and rationale. Additionally, CHOI teaches:
A computer-program product tangibly embodied in a non-transitory, machine-readable storage medium, including instructions configured to cause one or more data processors to perform a set of actions including: (CHOI [page 6, section 5.1 Experimental Setup] teaches: "All models were implemented with Theano (Bastien et al., 2012) and trained on a machine equipped with two Nvidia Tesla K80 GPUs.")
Regarding Claim 20:
CHOI in view of SRIVASTAVA teaches the elements of claim 14 as outlined above. Additionally, the claim recites similar limitations as corresponding claim 6 and is rejected for similar reasons as claim 6 using similar teachings and rationale.
Claims 2 and 16 are rejected under 35 U.S.C. 103 as being unpatentable over CHOI in view of SRIVASTAVA as applied respectively above to claims 1 and 14, and further in view of CORRADO (US 20170032242 A1), hereafter CORRADO.
Regarding Claim 2:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. SRIVASTAVA further teaches:
[...] the one or more SMA phenotypes specific to the subject […] an SMA phenotype of the one or more SMA phenotypes. (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype.” SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
CHOI in view of SRIVASTAVA is not relied upon for teaching, but CORRADO teaches: determining that the predicted progression [...] specific to the subject satisfies an early treatment condition, (CORRADO [0010] teaches: "For example, the healthcare professional can be provided with useful information about future health events that may become associated with a current patient, e.g., health events that are likely to be the next health event (i.e., predicted progression) to be associated with the patient or likelihoods that certain conditions will be satisfied by events occurring within a specified time period of the most recent event in the sequence. Additionally, the healthcare professional can be provided with information that identifies the potential effect of a proposed treatment on the likelihoods of the events occurring, e.g., whether a proposed treatment may reduce or increase the likelihood of an undesirable health-related condition being satisfied for the patient in the future." CORRADO [0073] teaches: "In these implementations, the set of future condition scores generated by the recurrent neural network for the last time step in the input temporal sequence is the set of future condition scores for the input temporal sequence." CORRADO [0084] teaches: “For example, a system administrator or other user may designate one or more particular conditions being satisfied as undesirable (i.e., determining that the predicted progression […] specific to the subject satisfied an early treatment condition). The system can then automatically perform the process 600 in response to a new event being added to the temporal sequence and generate an alert to notify the user if the future condition score for one of the undesirable condition crosses the threshold score or increases by more than the threshold increase.”)
wherein satisfying the early treatment condition is indicative of a recommendation to perform a treatment before the subject exhibits […] (CORRADO [0074] teaches: "The likelihood of occurrence of the possible temporal sequence may be, e.g., a product of the next input scores for the health events at each of the additional time steps in the sequence." CORRADO [0075] teaches: "Thus, a doctor may be given an opportunity to consider an additional treatment that could decrease the likelihood of an undesirable condition being satisfied in the future.")
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, and CORRADO before them, to include CORRADO’s future conditions scores outputted by a recurrent neural network in CHOI and SRIVASTAVA’s clinical event prediction method. One would have been motivated to make such a combination in order to consider an additional treatment that could decrease the likelihood of an undesirable condition being satisfied in the future (CORRADO [0075]).
Regarding Claim 16:
CHOI in view of SRIVASTAVA teaches the elements of claim 14 as outlined above. Additionally, the claim recites similar limitations as corresponding claim 2 and is rejected for similar reasons as claim 2 using similar teachings and rationale.
Claims 3 and 17 are rejected under 35 U.S.C. 103 as being unpatentable over CHOI in view of SRIVASTAVA as applied respectively above to claims 1 and 14, and further in view of CORRADO, DALTON (US 20170068789 A1), and THORNTON (US 20190108315 A1), hereafter DALTON and THORNTON.
Regarding Claim 3:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI further teaches:
[…] and the anonymized subject having been diagnosed […]; (CHOI [page 4, section 3. Cohort] teaches: "The dataset consists of de-identified (i.e., an anonymized subject) encounter orders, medication orders, problem list records and procedure orders." CHOI [page 4, section 4. Methods] teaches: "For each patient, the observations are drawn from a multilabel point process in the form of
t
i
,
x
i
for
i
=
1
,
…
,
n
. Each pair represents an event, such as an ambulatory care visit, during which multiple medical codes such as ICD-9 diagnosis codes (i.e., the subject having been diagnosed), procedure codes, or medication codes are documented in the patient record.)
SRIVASTAVA further teaches: […] the one or more SMA phenotypes […]; […] the one or more SMA phenotypes specific to the subject, and the […] subject having been diagnosed with SMA; (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype.” SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
CHOI in view of SRIVASTAVA is not relied upon for teaching, but CORRADO teaches: when the predicted progression […] satisfies the early treatment condition: (CORRADO [0010] teaches: "For example, the healthcare professional can be provided with useful information about future health events that may become associated with a current patient, e.g., health events that are likely to be the next health event (i.e., predicted progression) to be associated with the patient or likelihoods that certain conditions will be satisfied by events occurring within a specified time period of the most recent event in the sequence (i.e., satisfies the early treatment condition). Additionally, the healthcare professional can be provided with information that identifies the potential effect of a proposed treatment on the likelihoods of the events occurring, e.g., whether a proposed treatment may reduce or increase the likelihood of an undesirable health-related condition being satisfied for the patient in the future." CORRADO [0073] teaches: "In these implementations, the set of future condition scores generated by the recurrent neural network for the last time step in the input temporal sequence is the set of future condition scores for the input temporal sequence." CORRADO [0084] teaches: “For example, a system administrator or other user may designate one or more particular conditions being satisfied as undesirable. The system can then automatically perform the process 600 in response to a new event being added to the temporal sequence and generate an alert to notify the user if the future condition score for one of the undesirable condition crosses the threshold score or increases by more than the threshold increase.”)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, and CORRADO before them, to include CORRADO’s future conditions scores outputted by a recurrent neural network in CHOI and SRIVASTAVA’s clinical event prediction method. One would have been motivated to make such a combination in order to consider an additional treatment that could decrease the likelihood of an undesirable condition being satisfied in the future (CORRADO [0075]).
CHOI in view of SRIVASTAVA and CORRADO is not relied upon for teaching:
identifying an existing disease progression associated with an anonymized subject, the existing disease progression matching the predicted progression […] specific to the subject, […]
identifying a user who training the anonymized subject associated with the existing disease progression; and
transmitting a communication to a user device associated with the user, the communication requesting treatment recommendations for the subject.
However, DALTON teaches: identifying an existing disease progression associated with an anonymized subject, the existing disease progression matching the predicted progression […] specific to the subject, […] (DALTON [0016] teaches: "The support tool can be used, for example, to inform clinicians and patients of evidence and knowledge supporting best outcomes for individual patients based on the ability to identify a population of patients who are similar (i.e., identifying an existing disease progression associated with an […] subject, the existing disease progression matching the predicted progression […] specific to the subject), e.g., both by phenotype and genotype.” DALTON [0033] teaches: "information can be anonymized and/or protected (i.e., anonymized subject).”)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, CORRADO, and DALTON before them, to include DALTON’s support tool for identifying patients who are similar in both phenotype and genotype in CHOI, SRIVASTAVA, and CORRADO’s clinical event prediction method. One would have been motivated to make such a combination in order to develop therapeutic and long-term treatment plans based on the past experience of similar patients existing in the data warehouse who are most similar to the candidate patient and present options to the patient and healthcare provider that are predicted to result in “best” diagnostic and therapeutic outcomes for individual patients (DALTON [0007] and [0017]).
CHOI in view of SRIVASTAVA, CORRADO, and DALTON is not relied upon for teaching, but THORNTON teaches: identifying a user who training the […] subject associated with the existing disease progression; (THORNTON [0030] teaches: "In one embodiment, a user, such as a therapy developer, may input via a user-input interface, e.g., a touchscreen, instructions for a processor to search through the database to identify one or more healthcare providers who treated at least one patient meeting the set of predetermined medical criteria and at least one patient meeting the set of predetermined medical criteria." THORNTON [0030] teaches: “Said another way, the present system and method improves therapy recommendations by allowing therapy developers to target and educate the appropriate healthcare providers about therapies that are likely to be beneficial to the type of patient treated by the healthcare providers.” Examiner’s note: As noted in the claim objections above, examiner will construe the limitation as "identifying a user who treated/treats the anonymized subject".)
transmitting a communication to a user device associated with the user, the communication requesting treatment recommendations for the subject. (THORNTON [0037] teaches: "Each of the general healthcare providers 306, 310, 314 can receive the therapy healthcare providers 320, 324, 328 via their respective server 308, 312, 316, therapy information or therapy criteria of patients they have treated using those therapies. The information allows the healthcare providers, or third parties, to match patients with therapy healthcare providers that have provided therapies for patients with similar medical parameters. Thus each therapy healthcare providers' computing systems 322, 326, 330 includes and can transmit a set of criteria that represents the primary and secondary medical parameters that qualify for the therapy, which can represent medical parameters of patients that have responded well to the therapies.")
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, CORRADO, DALTON, and THORNTON before them, to include THORNTON’s identification of healthcare providers for recommending therapies in CHOI, SRIVASTAVA, CORRADO, and DALTON’s clinical event prediction method. One would have been motivated to make such a combination in order to improve therapy recommendations by allowing therapy developers to target and educate the appropriate healthcare providers about therapies that are likely to be beneficial to the type of patient treated by the healthcare providers (THORNTON [0030]).
Regarding Claim 17:
CHOI in view of SRIVASTAVA teaches the elements of claim 14 as outlined above. Additionally, the claim recites similar limitations as corresponding claim 3 and is rejected for similar reasons as claim 3 using similar teachings and rationale.
Claims 4-5 and 18-19 are rejected under 35 U.S.C. 103 as being unpatentable over CHOI in view of SRIVASTAVA as applied respectively above to claims 1 and 14, and further in view of DALTON.
Regarding Claim 4:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI further teaches:
[…] and the anonymized subject having been diagnosed […]; (CHOI [page 4, section 3. Cohort] teaches: "The dataset consists of de-identified (i.e., an anonymized subject) encounter orders, medication orders, problem list records and procedure orders." CHOI [page 4, section 4. Methods] teaches: "For each patient, the observations are drawn from a multilabel point process in the form of
t
i
,
x
i
for
i
=
1
,
…
,
n
. Each pair represents an event, such as an ambulatory care visit, during which multiple medical codes such as ICD-9 diagnosis codes (i.e., the subject having been diagnosed), procedure codes, or medication codes are documented in the patient record.)
SRIVASTAVA further teaches: […] the one or more SMA phenotypes […]; […] the one or more SMA phenotypes specific to the subject, and the […] subject having been diagnosed with SMA; (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype.” SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
CHOI in view of SRIVASTAVA is not relied upon for teaching, but DALTON teaches: when the predicted progression of the one or more […] phenotypes does not satisfy the early treatment condition: (DALTON [0034] teaches: "The information can be used to generate hypotheses about former, current and potential courses of treatment, e.g., that can be used in generating future treatment suggestions." Examiner’s note: Under broadest reasonable interpretation, when […] does not satisfy the early treatment condition can be interpreted as generating the hypothesis without for potential courses of treatment.)
identifying an existing disease progression associated with an anonymized subject, the existing disease progression matching the predicted progression […] specific to the subject, […] (DALTON [0016] teaches: "The support tool can be used, for example, to inform clinicians and patients of evidence and knowledge supporting best outcomes for individual patients based on the ability to identify a population of patients who are similar (i.e., identifying an existing disease progression associated with an […] subject, the existing disease progression matching the predicted progression […] specific to the subject), e.g., both by phenotype and genotype.” DALTON [0033] teaches: "information can be anonymized and/or protected (i.e., anonymized subject).”)
retrieving an anonymized subject record characterizing the anonymized subject; (DALTON [Abstract] teaches: "identifying relevant cohorts that include a population of patients who are similar, both by phenotype and genotype, to a candidate patient;" DALTON [0033] teaches: "information can be anonymized and/or protected". DALTON [Abstract] teaches: "and provide a treatment suggestion to the patient or a care provider based on the selection (i.e., retrieving an anonymized subject record).")
extracting a treatment schedule from the anonymized subject record; and (DALTON [Abstract] teaches: "evaluate the patient cohorts to identify a cohort that is a best match for the candidate patient based on historical treatment information for patients in each candidate patient cohort;" DALTON [0036] In some implementations, the method can further include developing therapeutic and long-term treatment plans based on the past experience of similar patients existing in the data warehouse who are most similar to the candidate patient. As an example, the EBCD engine 102 can generate or modify treatment plans (i.e., extracting a treatment schedule) for population groups that include patients who are most like the candidate patient.” DALTON [0033] teaches: "information can be anonymized and/or protected".)
transmitting the treatment schedule to a user device. (DALTON [Abstract] teaches: "and provide a treatment suggestion to the patient or a care provider based on the selection." DALTON [0031] teaches: “A resulting treatment suggestion can be provided to the candidate patient (e.g., in the patient portal 104) and/or to the candidate patient's physician (e.g., in the healthcare provider portal 106).”)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, and DALTON before them, to include DALTON’s support tool for identifying patients who are similar in both phenotype and genotype in CHOI and SRIVASTAVA’s clinical event prediction method. One would have been motivated to make such a combination in order to develop therapeutic and long-term treatment plans based on the past experience of similar patients existing in the data warehouse who are most similar to the candidate patient and present options to the patient and healthcare provider that are predicted to result in “best” diagnostic and therapeutic outcomes for individual patients (DALTON [0007] and [0017]).
Regarding Claim 5:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. SRIVASTAVA further teaches:
[…] SMA phenotypes […] SMA; (SRIVASTAVA [page 359, Subjects] teaches: "Patients with SMA were identified through the Neuromuscular and SMA Clinics at Children’s Hospital Boston and through the Pediatric Neuromuscular Clinical Research network. Patients with SMA were required to have a positive genetic test for SMA or have a sibling with a known positive SMN gene mutation and the appropriate clinical phenotype.” SRIVASTAVA [page 2, Subjects] teaches: "Patients with SMA were differentiated into those with type 2 and type 3 based on the standard clinical criterion of their maximum level of motor function achieved at any point, including whether the patient was able to walk (type 3) or only sit (type 2).")
CHOI in view of SRIVASTAVA is not relied upon for teaching, but DALTON teaches: matching the completion word or phrase associated with the subject to another one or more […] phenotypes associated with another subject having been previously treated […]; (DALTON [Abstract] teaches: "evaluate the patient cohorts to identify a cohort that is a best match (i.e., matching) for the candidate patient based on historical treatment information for patients in each candidate patient cohort;" DALTON [Abstract] teaches: "The engine is operable to: identify patient cohorts defined by the data in the data warehouse including identifying relevant cohorts that include a population of patients who are similar, both by phenotype and genotype (i.e., completion word or phrase associated with the subject), to a candidate patient (i.e., to another one or more […] phenotypes associated with another subject having been previously treated);" DALTON [0033] teaches: "information can be anonymized and/or protected".)
retrieving an anonymized subject record characterizing the anonymized subject; (DALTON [Abstract] teaches: "identifying relevant cohorts that include a population of patients who are similar, both by phenotype and genotype, to a candidate patient;" DALTON [0033] teaches: "information can be anonymized and/or protected". DALTON [Abstract] teaches: "and provide a treatment suggestion to the patient or a care provider based on the selection (i.e., retrieving an anonymized subject record).")
extracting a treatment schedule from the anonymized subject record; and (DALTON [Abstract] teaches: "evaluate the patient cohorts to identify a cohort that is a best match for the candidate patient based on historical treatment information for patients in each candidate patient cohort;" DALTON [0036] In some implementations, the method can further include developing therapeutic and long-term treatment plans based on the past experience of similar patients existing in the data warehouse who are most similar to the candidate patient. As an example, the EBCD engine 102 can generate or modify treatment plans (i.e., extracting a treatment schedule) for population groups that include patients who are most like the candidate patient.” DALTON [0033] teaches: "information can be anonymized and/or protected".)
transmitting the treatment schedule to a user device. (DALTON [Abstract] teaches: "and provide a treatment suggestion to the patient or a care provider based on the selection." DALTON [0031] teaches: “A resulting treatment suggestion can be provided to the candidate patient (e.g., in the patient portal 104) and/or to the candidate patient's physician (e.g., in the healthcare provider portal 106).”)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, and DALTON before them, to include DALTON’s support tool for identifying patients who are similar in both phenotype and genotype in CHOI and SRIVASTAVA’s clinical event prediction method. One would have been motivated to make such a combination in order to develop therapeutic and long-term treatment plans based on the past experience of similar patients existing in the data warehouse who are most similar to the candidate patient and present options to the patient and healthcare provider that are predicted to result in “best” diagnostic and therapeutic outcomes for individual patients (DALTON [0007] and [0017]).
Regarding Claim 18:
CHOI in view of SRIVASTAVA teaches the elements of claim 14 as outlined above. Additionally, the claim recites similar limitations as corresponding claim 4 and is rejected for similar reasons as claim 4 using similar teachings and rationale.
Regarding Claim 19:
CHOI in view of SRIVASTAVA teaches the elements of claim 14 as outlined above. Additionally, the claim recites similar limitations as corresponding claim 5 and is rejected for similar reasons as claim 5 using similar teachings and rationale.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over CHOI in view of SRIVASTAVA as applied above to claim 1, and further in view of ZHONG (US 20210050084 A1), hereafter ZHONG.
Regarding Claim 7:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI in view of SRIVASTAVA is not relied upon for teaching, but ZHONG teaches:
wherein the disease progression is output at a computing device of the subject using a chatbot. (ZHONG [0101] teaches: "The chatbot feature allows a physician to quickly pull any relevant patient information they may need, both during and between patient visits, simply by speaking into their device's microphone or typing a request on the frontend of the HCP assistant system 102 (e.g., text entry fields that appear on a web page)." ZHONG [0035] teaches: "[0035] The output generated by the HCP assistant system 102 enables HCPs to monitor patient progress in meeting certain behavioral and/or clinical outcome goals."
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, and ZHONG before them, to include ZHONG’s chatbot feature in CHOI and SRIVASTAVA’s clinical event prediction method. One would have been motivated to make such a combination in order to monitor patient progress in meeting certain behavioral and/or clinical outcome goals (ZHONG [0035]).
Claims 11-12 are rejected under 35 U.S.C. 103 as being unpatentable over CHOI in view of SRIVASTAVA as applied above to claims 1, and further in view of CARLINI (“The Secret Sharer: Evaluating and Testing Unintended Memorization in Neural Networks”), hereafter CARLINI.
Regarding Claim 11:
CHOI in view of SRIVASTAVA teaches the elements of claim 1 as outlined above. CHOI in view of SRIVASTAVA is not relied upon for teaching, but CARLINI teaches:
detecting data leakage associated with the NLP model, the data leakage exposing a feature of the set of features included in the subject record characterizing the subject; (CARLINI [page 267, Abstract] teaches: We show that our testing strategy is a practical and easy-to-use first line of defense, e.g., by describing its application to quantitatively limit data exposure […].” CARLINI [page 267, section 1 Introduction] teaches: “The users of such models may discover—either by accident or on purpose—that entering certain text prefixes causes the models to output surprisingly-revealing text completions. For example, users may find that the input “my social-security number is. . . ” gets auto-completed to an obvious secret (such as a valid-looking SSN not their own), or find that other inputs are auto-completed to text with oddly-specific details. […] We demonstrate our algorithm’s effectiveness in experiments, e.g., by extracting credit card numbers from a language model trained on the Enron email data.” CARLINI [page 281, section 10 Related Work and Conclusions] teaches: “In contrast, in our paper, we show that memorization can occur, and training data leaked, even when there is not an attacker present intentionally causing a back-door.” CARLINI [page 271, section 4.2 The Precise Exposure Metric] teaches: “The remainder of this section discusses how we can measure the degree to which an individual canary
s
r
^
is memorized (i.e., detecting data leakage) when inserted in the dataset.”)
in response to detecting data leakage associated with the NLP model, executing a data leakage prevention protocol that prevents or blocks exposure of the feature of the set of features included in the subject record. (CARLINI [page 267, section 1 Introduction] teaches: “Only by using differentially-private training techniques are we able to eliminate the issue completely […].” CARLINI [page 278, section 9 Preventing Unintended Memorization] teaches: “As we have shown, neural networks quickly memorize secret data. This section evaluates (both the efficacy and impact on accuracy) three potential defenses against memorization: regularization, sanitization, and differential privacy.” CARLINI [page 279, section 9.3 Differential Privacy] teaches: “Differential privacy [12, 14, 15] is a property that an algorithm can satisfy which bounds the information it can leak about its inputs. […] Thus, differential privacy is a desirable property to defend against memorization. Consider the case where
D
contains one occurrence of some secret training record
x
(i.e., features included in the […] record), and
D
'
=
D
-
x
. […] We applied the differentially-private stochastic gradient descent algorithm (DP-SGD) from [1] to verify that differential privacy is an effective defense that prevents memorization (i.e., executing a data leakage prevention protocol that prevents or blocks exposure of the feature of the set of features.”)
Accordingly, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the claimed invention, having the teachings of CHOI, SRIVASTAVA, and CARLINI before them, to include CARLINI’s differential privacy training techniques in CHOI and SRIVASTAVA’s clinical event prediction method. One would have been motivated to make such a combination in order to use us differentially-private training techniques to eliminate the data memorization issue completely (CARLINI [page 267, section 1 Introduction]).
Regarding Claim 12:
CHOI in view of SRIVASTAVA and CARLINI teaches the elements of claim 11 as outlined above. CARLINI further teaches:
wherein executing the data leakage prevention protocol includes re-training the NLP model according to a differential privacy model. (CARLINI [page 267, section 1 Introduction] teaches: “Only by using differentially-private training techniques are we able to eliminate the issue completely […].” CARLINI [page 278, section 9 Preventing Unintended Memorization] teaches: “As we have shown, neural networks quickly memorize secret data. This section evaluates (both the efficacy and impact on accuracy) three potential defenses against memorization: regularization, sanitization, and differential privacy.” CARLINI [page 279, section 9.3 Differential Privacy] teaches: “Differential privacy [12, 14, 15] is a property that an algorithm can satisfy which bounds the information it can leak about its inputs. […] Thus, differential privacy is a desirable property to defend against memorization. Consider the case where
D
contains one occurrence of some secret training record
x
(i.e., features included in the […] record), and
D
'
=
D
-
x
. […] We applied the differentially-private stochastic gradient descent algorithm (DP-SGD) from [1] to verify that differential privacy is an effective defense that prevents memorization (i.e., executing a data leakage prevention protocol that prevents or blocks exposure of the feature of the set of features.” Examiner’s note: As noted above in the 112(b) rejection, claim 12 will be construed as depending from claim 11.)
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Alvaro S Laham Bauzo whose telephone number is (571)272-5650. The examiner can normally be reached Mon-Fri 7:30 AM - 11:00 AM | 1:00 PM - 5:30 PM ET.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Usmaan Saeed can be reached on (571) 272-4046. 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.
/A.S.L./Examiner, Art Unit 2146
/USMAAN SAEED/Supervisory Patent Examiner, Art Unit 2146