Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
DETAILED ACTION
Status of Claims
The present Office Action is pursuant to Applicant’s communication on 05-11-2026; application filed 08-03-2022; provisional application No. 62/971,065, filed on Feb. 6, 2020, provisional application No. 63/022, 736, filed on May 11, 2020, provisional application No. 63/089, 304, filed on Oct. 8, 2020.
Examiner’s Note
The rejections below group claims that may not be identical, but whose language and scope are so substantively similar as to lend themselves to grouping, in the interests of clarity and conciseness.
Information Disclosure Statement
The information disclosure statements (IDS) filed on 08-03-2026, have been acknowledged. The submissions are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
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, 3, 5, 7, 9, 10, 14, 15, 16, 18, 31, 32, 34, 35, 38, 39, 40, 41, 42, and 43 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception without significantly more.
I. Independent Claims 1 and 18
A. Step 1 – Is the Claim Directed to an Abstract Idea?
Yes. Independent Claim 1 (system) and Claim 18 (method) are each directed to the abstract idea of comparing two data sets collected at two time points, applying a mathematical/statistical classifier to the comparison, and filtering results based on a calculated score to identify targets. The operative recitations reduce to mental and mathematical concepts:
"receive … (a) a first biological signature … and (b) a second biological signature …" – the recitation of receiving two data sets is an abstract data-gathering step;
"calculate … a set of values representing a relation between said first and second biological signatures" – a mathematical relationship (ratio/difference);
"apply a pretrained Machine Learning (ML)-based classifier to said set of values, to predict a response" – application of a known classification algorithm to a data set;
"apply a statistical test to select a subset of Differentially Expressed Proteins (DEPs)" – use of a well-known statistical test;
"calculate … a Resistance Associated Protein (RAP) score, representing a deviation from a distribution of protein expression" – a mathematical/probabilistic calculation;
"filter the DEPs based on the associated RAP scores" – a threshold-based sorting operation; and
"indicate one or more filtered DEPs as intervention targets" – a mental/observational step.
The specification confirms that these are abstract mathematical and statistical operations. The Specification discloses that the core calculation is a log2 ratio of protein expression between two time points: [0089]: "the present disclosure provides for calculating the relations (e.g., a level of difference in expression values between each biological signature in a pair of signatures associated with a subject, e.g., a difference in, and/or a ratio of, expression values between biological signatures at at-least two time points relative to the therapy, e.g., a T1/T0 ratio." The RAP score itself is disclosed as a pure mathematical formula: [0125]: "RAP score = log2 [P(NR) / P(R)] (Equation 1) … where P(NR) is the probability that the protein can be attributed to the non-responder distribution and P(R) is the probability that the protein can be attributed to the responder distribution."
These are fundamental mathematical and statistical concepts-comparing two measurements, computing a ratio, applying a classifier, performing a statistical test, and filtering by score. They are not tied to any particular technology and are the type of "method of organizing human activity" and "mental process" concepts identified in Alice Corp. v. CLS Bank Int'l, 573 U.S. 208 (2014), and Mayo Collaborative Servs. v. Prometheus Labs., Inc., 566 U.S. 66 (2012). The DEP identification is likewise a well-known statistical operation: [0105]: "DEPs are defined as proteins who's median, mean, variance or other statistical measure differ between responders and non-responders … detectable by numerical measures (e.g., t-test, ANOVA, Kolmogorov-Smirnov test)."
Conclusion of Step 1: Claims 1 and 18 are directed to patent-ineligible abstract ideas (mathematical concepts and mental processes of data comparison, classification, and statistical filtering).
B. Step 2A – Does the Claim Integrate the Abstract Idea into a Practical Application?
No. When the claim is considered as a whole, the additional recitations do not integrate the abstract idea into a significantly more practical application.
The phrase "a target subject having a specified type of disease and receiving a specified therapy for treating said disease" merely provides a field-of-use limitation on the data; it does not alter the abstract nature of the comparison/classification steps. The Specification confirms that the "specified type of disease" is open-ended: [0074]: "the specified disease and/or conditions may be at different stages … the disease is a proliferative disorder … the disease is cancer … the disease is an autoimmune disease … the disease is a genetic disease … the disease is an inflammatory disease" – the field-of-use is essentially any disease, which does not transform the abstract steps.
"Biological sample collected at a first time point … and a second time point" is a well-known, routine data-collection step. The Specification confirms this: [0055]: "biological samples may be obtained from each subject … at specified times before, during, and/or after the conclusion of, the course of therapy … e.g., pre-therapy, on-therapy, and/or post-therapy."
The "pretrained ML-based classifier" is a generic computational operation. The Specification lists conventional algorithms: [0098]: "Support Vector Machine (SVM) … Penalized Logistic Regression (PLR) … Generalized linear model (GLM) … Random forest (RF) … eXtreme Gradient Boosting (XGB)" – all of which are well-known, generic machine-learning techniques.
"Apply a statistical test to select a subset of DEPs" is a well-established statistical practice. The Specification states: [0105]: "detectable by numerical measures (e.g., t-test, ANOVA, Kolmogorov-Smirnov test)."
"Calculate … a RAP score" is the application of a mathematical formula (Equation 1). [0125]: "RAP score = log2 [P(NR)/P(R)]".
"Filter the DEPs based on the associated RAP scores" is a threshold-based sorting operation, i.e., a mental/observational step.
"Indicate one or more filtered DEPs as intervention targets" is an observational/mental step of labeling.
No claim element imposes a meaningful limitation on the technology being employed, nor does the ordered combination effect a transformation of the data into something beyond the abstract result of "a predicted response and a list of intervention targets." The Specification itself acknowledges that the classifier "does not consider the biological function of the proteins": [0147]: "In contrast to the classifier analysis, which does not consider the biological function of the proteins that are part of the classifier, this analysis aims to characterize the host response in the context of differentiating biological processes."
Conclusion of Step 2A: The claim elements, individually and as an ordered combination, do not integrate the abstract idea into a practical application. The claim remains directed to the abstract concept of comparing two data sets, applying a classifier, performing a statistical test, computing a probability-based score, filtering, and labeling.
C. Step 2B – Does the Claim Provide an Inventive Concept?
No. Even assuming the claim recites additional elements beyond the abstract idea, those additional elements do not provide an inventive concept sufficient to make the claim patent-eligible.
Claim 1 (System) – Full Recitation with Additional Elements Emboldened:
A system for predicting a response in a target subject to a specified therapy, comprising: at least one hardware processor; and a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to: receive, for a target subject having a specified type of disease and receiving a specified therapy for treating said disease, (a) a first biological signature associated with a biological sample collected at a first time point relative to said specified therapy, and (b) a second biological signature associated with a biological sample collected at a second time point relative to said specified therapy; calculate, for the target subject, a set of values representing a relation between said first and second biological signatures associated with the target; apply a pretrained Machine Learning (ML)-based classifier to said set of values, to predict a response in the target subject to the specified therapy; when the target subject is predicted as a non-responder subject, apply a statistical test to select a subset of Differentially Expressed Proteins (DEPs), whose expression levels change between the first time point and second time point; calculate, for each DEP of the target subset, a Resistance Associated Protein (RAP) score, representing a deviation from a distribution of protein expression in at least one group selected from (i) subjects responding to the specified therapy, and (ii) subjects not responding to the specified therapy; filter the DEPs based on the associated RAP scores; and indicate one or more filtered DEPs as intervention targets for the target non-responder subject.
The emboldened elements-"at least one hardware processor," "a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor"-are generic computer components. As the Specification confirms, the computer-readable storage medium is a conventional element: [0163]: "The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium may be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing." The medium is further described as including "a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk … and any suitable combination of the foregoing." [0163]
These are generic computing elements that amount to "apply it" instructions. They do not improve the functioning of the computer itself, nor do they impose any meaningful limitation on the abstract idea. The remaining unemboldened steps-receiving data, calculating a relation, applying a classifier, performing a statistical test, computing a RAP score, filtering, and indicating targets-are the abstract idea itself.
The Specification further confirms that the underlying data acquisition is routine: [0076]: "Body fluids may be whole blood, blood plasma, blood serum, peripheral blood mononuclear cells (PBMCs), lymph, urine, saliva, semen, synovial fluid and spinal fluid, fresh or frozen." And the protein profiling is a well-known technique: [0084]: "Protein expression can be assayed by any known method including, an immunoassay, immunoblotting, immunohistochemistry, FACS, ELISA, Western blotting, proteomics arrays, proteome sequencing, proximity-extension assay (PEA)-based assay, aptamer-based assays, multiplex assay and mass spectrometry."
Conclusion of Step 2B for Claim 1: The additional elements (generic processor, storage medium, program instructions) are conventional and do not transform the abstract idea into a patent-eligible application. The claim does not contain an inventive concept. Claim 1 is rejected under 35 U.S.C. § 101.
Claim 18 (Method) – Full Recitation with Additional Elements Emboldened:
A method of predicting a response in a target subject to a specified therapy, comprising: receiving, for a target subject having a specified type of disease and receiving a specified therapy for treating said disease, (a) a first biological signature associated with a biological sample collected at a first time point relative to said specified therapy, and (b) a second biological signature associated with a biological sample collected at a second time point relative to the specified therapy; calculating, for the target subject a set of values representing a relation between said first and second biological signatures associated with the target; applying a pretrained Machine Learning (ML)-based classifier to said set of values, to predict a response in the target subject to the specified therapy; when the target subject is predicted as a non-responder subject, apply a statistical test to select a subset of Differentially Expressed Proteins (DEPs), whose expression levels change between the first time point and second time point; calculating, for each DEP of the target subset, a Resistance Associated Protein (RAP) score, representing a deviation from a distribution of protein expression in at least one group selected from (i) subjects responding to the specified therapy, and (ii) subjects not responding to the specified therapy; filtering the DEPs based on the associated RAP scores; and indicating one or more filtered DEPs as intervention targets for the target non-responder subject.
Claim 18 is a method claim that recites the same abstract steps as Claim 1 but without the "at least one hardware processor" and "non-transitory computer-readable storage medium" preamble. The method steps-receiving, calculating, applying, applying a statistical test, calculating a score, filtering, and indicating-are entirely mental and mathematical in nature. The Specification confirms that the method is a computational/analytical process: [0053]: "the present disclosure provides (i) a computational approach for training a machine learning model to predict a response in patients, as well as (ii) methods for selecting key proteins whose targeting may improve therapy efficacy and/or response to therapy."
No additional element in Claim 18 transforms the abstract steps. The method is a sequence of data-receiving, mathematical computation, classification, statistical testing, scoring, filtering, and labeling-each of which is a well-understood, routine, conventional step. The Specification at [0089] describes the core calculation as "a T1/T0 ratio," and at [0125] the RAP score as a log2 probability ratio-both pure mathematical operations.
Conclusion of Step 2B for Claim 18: The claim does not contain an inventive concept. The method steps, taken as an ordered combination, amount to "apply the well-known mathematical and statistical operations to the data and report the result." Claim 18 is rejected under 35 U.S.C. § 101.
II. Dependent Claims
A. Claim 3 (Depends on Claim 1)
The system of claim 1, wherein said first and second biological signatures are each protein expression profiles, and said sets of values each comprise, with respect to each protein in said protein expression profiles, a ratio or a difference in the levels of expression of said protein in said first and second biological signatures, wherein said protein expression profile comprises expression values for at least two proteins.
Step 1 / 2A / 2B Analysis: Claim 3 further specifies the abstract idea by identifying the "relation" as "a ratio or a difference in the levels of expression." This is a mathematical operation. The Specification confirms: [0089]: "the T1/T0 ratio is a numerical value determined by calculating the ratio of on-treatment and baseline values (pre-treatment)." The limitation "comprises expression values for at least two proteins" is a trivial quantity limitation; the Specification at [0013] states "the protein expression profile comprises expression values for at least two proteins" and at [0107] discloses "The number of proteins measured may vary between 1 and 20,000." Specifying "at least two proteins" does not add any meaningful technological constraint. The claim merely narrows the abstract mathematical concept (ratio/difference of expression values) without adding an inventive concept.
Claim 3 is rejected under 35 U.S.C. § 101 as dependent on ineligible Claim 1 and not itself saving the claim.
B. Claim 5 (Depends on Claim 1)
The system of any one of claim 1, wherein said program instructions are further executable to perform a dimensionality reduction stage with respect to said sets of values, to reduce the number of variables in each of said sets of values, and wherein said dimensionality reduction stage identifies a subset of principal proteins in each of said sets of values.
Analysis: "Dimensionality reduction" is a well-known, conventional data-processing technique. The Specification at [0093] discloses: "any suitable feature selection and/or dimensionality reduction method or technique may be employed, such as, but not limited to: ANOVA with So parameter … SEMMS … L2N … Genetic algorithms … Naïve classifier." The Specification at [0014] states: "the method further comprises performing … a dimensionality reduction stage with respect to the sets of values, to reduce the number of variables." This is a standard statistical/data-reduction step that does not transform the abstract idea. It is a well-understood, routine, conventional application of a known mathematical technique (principal component analysis, variance-based filtering, etc.) to a data set.
Claim 5 is rejected under 35 U.S.C. § 101.
C. Claim 7 (Depends on Claim 5)
The system of claim 5, wherein said training set comprises only said subset of principal proteins in each of said sets of values.
Analysis: Limiting the training set to "only the subset of principal proteins" is a well-known data-preselection step. The Specification at [0017] states: "the training set comprises only the subset of principal proteins in each of the sets of values." This is a conventional data-filtering operation that does not add an inventive concept beyond the abstract idea of selecting a subset of features for a classifier.
Claim 7 is rejected under 35 U.S.C. § 101.
D. Claim 9 (Depends on Claim 1)
The system of claim 1, wherein each of said biological samples is one of: blood plasma, whole blood, blood serum, cerebrospinal fluid (CSF), and peripheral blood mononuclear cells (PBMCs).
Analysis: Specifying the type of biological sample is a field-of-use/source-of-data limitation that does not change the abstract nature of the comparison/classification steps. The Specification at [0076] discloses: "Body fluids may be whole blood, blood plasma, blood serum, peripheral blood mononuclear cells (PBMCs), lymph, urine, saliva, semen, synovial fluid and spinal fluid, fresh or frozen." The choice of sample type is a well-known, routine selection in the field of biomarker analysis and does not transform the abstract mathematical comparison into a practical application.
Claim 9 is rejected under 35 U.S.C. § 101.
E. Claim 10 (Depends on Claim 1)
The system of claim 1, wherein said specified type of disease is a proliferative disease or cancer.
Analysis: Limiting the "specified type of disease" to "a proliferative disease or cancer" is a broad field-of-use limitation. The Specification at [0074] enumerates: "the disease is a proliferative disorder … the disease is cancer … the cancer is a solid cancer … the cancer is a hematopoietic cancer … the cancer is melanoma … the cancer is small cell lung cancer (SCLC) or non-small cell lung cancer (NSCLC) … the cancer is urogenital cancer … the cancer is head and neck cancer." The claim does not recite any specific biological mechanism or technological limitation beyond naming the disease category. The abstract idea of comparing two data sets, applying a classifier, and filtering by score remains unchanged regardless of the disease label.
Claim 10 is rejected under 35 U.S.C. § 101.
F. Claim 14 (Depends on Claim 1)
The system of claim 1, wherein said predicting comprises an indication of secondary effects in said target subject.
Analysis: Adding "an indication of secondary effects" to the prediction output is an observational/mental step. The Specification at [0023] states: "the predicting comprises an indication of secondary effects in the target subject" and at [0064]: "the present disclosure may provide also for predicting adverse events associated with the specified treatment or therapy of a target patient." This is a mental/observational step of noting an outcome category. It does not add a technological limitation or transform the abstract classification into a practical application.
Claim 14 is rejected under 35 U.S.C. § 101.
G. Claim 15 (Depends on Claim 1)
The system of claim 1, wherein said program instructions are further executable to determine, based, at least in part, on said predicting, at least one of: continuing said specified therapy in said target subject, adjusting said specified therapy in said target subject, discontinuing said specified therapy in said target subject, and administering a different therapy to said target subject.
Analysis: The step of "determining … at least one of: continuing … adjusting … discontinuing … and administering a different therapy" is a mental/observational decision step. The Specification at [0025] states: "the method further comprises determining … based, at least in part, on the predicting, at least one of: continuing the specified therapy in the target subject, adjusting the specified therapy in the target subject, discontinuing the specified therapy in the target subject, and administering a different therapy to the target subject." This is a well-known, routine clinical decision-making step. The determination is made "based, at least in part, on" the prediction-i.e., a human-readable output that a clinician would already use. The claim does not recite any technological implementation of the therapy adjustment; it merely identifies the category of clinical decisions that may follow. This is a mental/observational step that does not transform the abstract idea.
Claim 15 is rejected under 35 U.S.C. § 101.
H. Claim 16 (Depends on Claim 1)
The system of claim 1, wherein said specified therapy is an immunotherapy, wherein said immunotherapy is selected from anti-PD-1/PD-L1 therapy, anti-CTLA-4 therapy, and both.
Analysis: Naming a specific therapy modality (immunotherapy, anti-PD-1/PD-L1, anti-CTLA-4) is a field-of-use limitation. The Specification at [0026] states: "the specified therapy is an immunotherapy … the immunotherapy is selected from anti-PD-1/PD-L1 therapy, anti-CTLA-4 therapy, and both." The Specification at [0066] further elaborates: "the immunotherapy comprises immune checkpoint modulation … Immune checkpoint proteins are well known in the art and include, but are not limited to PD-1, PD-L1, PD-L2, CTLA-4 … LAG-3 … TIM-3 … TIGIT and VISTA." Specifying the therapy type does not add a technological limitation to the abstract comparison/classification/filtering steps.
Claim 16 is rejected under 35 U.S.C. § 101.
I. Claim 31 (Depends on Claim 18)
The method of claim 18, further comprising at an inference stage, applying said classifier to a target set of said values associated with a target subject, thereby predicting a response in said target subject to said specified therapy.
Analysis: The "inference stage" of applying a classifier to a new data set is the standard deployment of a trained model. The Specification at [0101] states: "at an inference step 114, a trained machine learning model of the present disclosure may be applied to target data, e.g., a target set of T1/T0 relations … with respect to a target patient … the inference of the trained machine learning model on the target data produces a therapy response prediction or response probability." This is a well-understood, routine, conventional step in machine-learning pipeline deployment and does not add an inventive concept.
Claim 31 is rejected under 35 U.S.C. § 101.
J. Claim 32 (Depends on Claim 31)
(Previously Presented) The method of claim 31, further comprising determining, based, at least in part, on said predicting, at least one of: continuing said specified therapy in said target subject, adjusting said specified therapy in said target subject, discontinuing said specified therapy in said target subject, and administering a different therapy to said target subject.
Analysis: Identical to the analysis of Claim 15. The Specification at [0025] discloses this as a determination step based on the prediction. This is a mental/observational clinical decision step that does not transform the abstract idea.
Claim 32 is rejected under 35 U.S.C. § 101.
K. Claim 34 (Depends on Claim 32)
The method of claim 32, wherein adjusting said specified therapy or administering a different therapy to said target subject is determined by a method comprising: determining DEPs between responders and non-responders; determining, in the sample obtained from said subject, one or more RAPs, selected from the DEPs; and identifying a therapy suitable for balancing the level of the one or more RAPs in said subject.
Analysis: The sub-steps are: (i) "determining DEPs between responders and non-responders" – a well-known statistical identification; (ii) "determining … one or more RAPs, selected from the DEPs" – a filtering step; and (iii) "identifying a therapy suitable for balancing the level of the one or more RAPs" – a mental/observational matching step. The Specification at [0027] discloses: "adjusting the specified therapy or administering a different therapy to said target subject is determined by a method comprising: (i) determining differentially expressed proteins (DEPs) between responders and non-responders; (ii) determining, in the sample obtained from said subject, one or more resistance associated proteins (RAPs) selected from the determined DEPs; and (iii) selecting a therapy suitable for balancing the level of the one or more RAPs in said subject." The Specification at [0135] further discloses the clinical filter: "drugs/INDs that target each RAP are searched in a suitable database and are associated with the RAP." These are well-known, routine steps in the field of biomarker-guided therapy selection.
Claim 34 is rejected under 35 U.S.C. § 101.
L. Claim 35 (Depends on Claim 32)
(Previously Presented) The method of claim 32 wherein determining the one or more RAPs is by providing a probabilistic measurement of a distance of the DEP expression level from a defined group of samples selected from the responder group or the non-responder group, and wherein determining the one or more RAPs is by determining the expression distribution of each DEP in each of the responder and non-responder groups, fitting a probability density function for each group, and calculating for each subject, and based on the DEP expression of said subject, the probability of the DEP to be associated with one of the response groups.
Analysis: This claim recites the mathematical construction of the RAP score in greater detail: fitting a probability density function and calculating a probability. The Specification at [0029] states: "determining the one or more RAPs in a subject is by determining the expression distribution of each DEP in each of the responder and non-responder groups, fitting a probability density function for each group, and calculating for each subject and based on the DEP expression of said subject, the probability of the DEP to be associated with one of the response groups." The Specification at [0125] provides the formula: "RAP score = log2 [P(NR)/P(R)]." Fitting a probability density function and computing a probability ratio is a well-known, routine statistical operation (e.g., Gaussian/Pearson fitting, log-likelihood ratio). This does not add an inventive concept.
Claim 35 is rejected under 35 U.S.C. § 101.
M. Claim 38 (Depends on Claim 1)
The system of claim 1, wherein each one of said first and second biological signatures is selected from a list consisting of: a DNA profile, an RNA profile, a protein profile, a metabolomics profile, microbiome profile, a genomics profile, a transcriptomics profile, a cellular profile, an epigenomics profile, a post-translational modification-based profile, cellular profile, single-cell based analysis and a regulatory RNA profile.
Analysis: This claim enumerates the possible types of biological data that may be compared. The Specification at [0011] states: "the first and second biological signatures are each one of: a DNA profile, an RNA profile, a protein profile, a metabolomics profile, microbiome profile, a transcriptomics profile, a genomics profile, an epigenomics profile, a cellular profile, a post-translational modification-based profile, single-cell based analysis, and a regulatory RNA profile." These are all well-known types of biological data. Enumerating them does not add a technological limitation; the abstract comparison/classification/filtering steps remain the same regardless of the data type.
Claim 38 is rejected under 35 U.S.C. § 101.
N. Claim 39 (Depends on Claim 1)
The system of claim 1 wherein the at least one processor is configured to calculate the RAP score per DEP of the target subject based on a ratio between a probability that the DEP is attributed to distribution of protein expression in non-responding subjects, and a probability that the DEP is attributed to distribution of protein expression in responding subjects.
Analysis: This claim recites the RAP score calculation as a "ratio between a probability … and a probability." The Specification at [0125] discloses: "RAP score = log2 [P(NR)/P(R)]" and at [0126]: "P(NR) is the probability that the protein can be attributed to the non-responder distribution and P(R) is the probability that the protein can be attributed to the responder distribution." This is a pure mathematical ratio of two probability values. It is a well-known statistical operation and does not add an inventive concept.
Claim 39 is rejected under 35 U.S.C. § 101.
O. Claim 40 (Depends on Claim 1)
The system of claim 1 wherein the at least one processor is configured to filter the DEPs by selecting DEPs as proteins whose expression levels change between responding subjects and non-responding subjects, and between the first time point and second time point.
Analysis: This recites the DEP identification as "proteins whose expression levels change between responding subjects and non-responding subjects, and between the first time point and second time point." The Specification at [0105] discloses: "DEPs are defined as proteins who's median, mean, variance or other statistical measure differ between responders and non-responders … detectable by numerical measures (e.g., t-test, ANOVA, Kolmogorov-Smirnov test)." And at [0121]: "DEPs are proteins whose levels change between responders and non-responders and may also change between T0 and T1 in responders and non-responders." This is a well-known, routine statistical filtering step.
Claim 40 is rejected under 35 U.S.C. § 101.
P. Claim 41 (Depends on Claim 1)
The system of claim 1 wherein the at least one processor is configured to filter the DEPs by selecting DEPs that are targeted by available drugs or Investigational New Drugs (INDs).
Analysis: This recites a database-lookup step: identifying DEPs that correspond to known drug targets. The Specification at [0135] discloses: "First, drugs/INDs that target each RAP are searched in a suitable database and are associated with the RAP." This is a well-known, conventional database-search operation. It does not add a technological limitation to the abstract idea of comparing, classifying, and filtering.
Claim 41 is rejected under 35 U.S.C. § 101.
Q. Claim 42 (Depends on Claim 1)
The system of claim 1, wherein the at least one processor is further configured to: receive, for each of a plurality of subjects having the specified type of disease and receiving the specified therapy (a) a respective first biological signature and (b) a respective second biological signature; calculate, for each of the plurality of subjects, a respective set of values representing a relation between the respective first and second biological signatures; train the classifier on a training set comprising: (i) the calculated sets of values, and (ii) labels associated with an outcome of said specified therapy in each of said subjects, to predict a response in the target subject to said specified therapy.
Analysis: This claim recites the training phase of the ML classifier: receiving data from a cohort, calculating relations, and training a classifier on labeled data. The Specification at [0048] states: "the present disclosure provides for training a machine learning model using a training dataset comprising a biological profile (e.g., protein expression profile) of biological samples or biological signatures obtained from a plurality of subjects … having a specified type of disease and receiving a specified type of treatment." And at [0094]: "a training dataset for training a machine learning model … may be constructed, comprising sets of values representing relations (e.g., ratios or difference in expression values) of the biological signatures at multiple time points relative to the therapy, with respect to at least some of the subjects in the cohort." The Specification at [0098] lists the training algorithms: "Support Vector Machine (SVM) … Penalized Logistic Regression (PLR) … Generalized linear model (GLM) … Random forest (RF) … eXtreme Gradient Boosting (XGB)." Training a classifier on labeled data is a well-understood, routine, conventional step in machine learning. The Specification at [0143] confirms: "Using this approach, 10 differentially expressed proteins were identified. … A logistic regression model (GLM) was generated using the 10 selected proteins as predictors and the true response status as the dependent variable."
Claim 42 is rejected under 35 U.S.C. § 101.
R. Claim 43 (Depends on Claim 42)
The method of claim 42, wherein: said training set further comprises, with respect to at least some of said subjects, labels associated with clinical data, or wherein the statistical test is selected from a student's t-test, an Analysis of Variance (ANOVA) test and a Kolmogorov-Smirnov test.
Analysis: This claim adds two alternative limitations: (1) including clinical data labels in the training set, and (2) selecting a specific statistical test. The Specification at [0021] states: "the training set further comprises, with respect to at least some of the subjects, labels associated with clinical data." The Specification at [0105] discloses: "detectable by numerical measures (e.g., t-test, ANOVA, Kolmogorov-Smirnov test)." Both are well-known, conventional practices. Including clinical metadata in a training set is a standard data-enrichment step. Selecting among well-known statistical tests (t-test, ANOVA, Kolmogorov-Smirnov) is a routine choice. Neither limitation adds an inventive concept.
Claim 43 is rejected under 35 U.S.C. § 101.
III. Summary of Rejections
Claim
Status
Basis
1
Rejected § 101
Directed to abstract idea (mathematical comparison, ML classification, statistical filtering); generic computer components do not provide an inventive concept
3
Rejected § 101
Depends on 1; adds "ratio or difference" (mathematical) and "at least two proteins" (trivial quantity)
5
Rejected § 101
Depends on 1; "dimensionality reduction" is a well-known, routine statistical technique
7
Rejected § 101
Depends on 5; limiting training set to principal proteins is conventional data filtering
9
Rejected § 101
Depends on 1; specifying sample type (blood plasma, CSF, etc.) is a field-of-use/source-of-data limitation
10
Rejected § 101
Depends on 1; "proliferative disease or cancer" is a broad field-of-use limitation
14
Rejected § 101
Depends on 1; "indication of secondary effects" is a mental/observational step
15
Rejected § 101
Depends on 1; "determining continuing/adjusting/discontinuing/administering" is a mental clinical decision step
16
Rejected § 101
Depends on 1; specifying "immunotherapy, anti-PD-1/PD-L1, anti-CTLA-4" is a field-of-use limitation
18
Rejected § 101
Method claim directed to same abstract idea as Claim 1; no additional inventive concept
31
Rejected § 101
Depends on 18; "applying classifier at inference stage" is a routine ML deployment step
32
Rejected § 101
Depends on 31; "determining therapy continuation/adjustment/discontinuation" is a mental/observational step
34
Rejected § 101
Depends on 32; sub-steps of determining DEPs, RAPs, and matching to therapy are well-known, routine steps
35
Rejected § 101
Depends on 32; "fitting a probability density function" and "calculating probability" are well-known statistical operations
38
Rejected § 101
Depends on 1; enumerating biological signature types is a field-of-use/source-of-data limitation
39
Rejected § 101
Depends on 1; RAP score as a "ratio between a probability … and a probability" is a pure mathematical operation
40
Rejected § 101
Depends on 1; filtering by "expression levels change" is a well-known statistical filtering step
41
Rejected § 101
Depends on 1; "selecting DEPs targeted by available drugs or INDs" is a conventional database lookup
42
Rejected § 101
Depends on 1; "receiving data, calculating relations, training a classifier" is a well-known, routine ML training process
43
Rejected § 101
Depends on 42; adding clinical data labels and selecting from well-known statistical tests (t-test, ANOVA, KS) are routine practices
IV. Overall Conclusion
All claims are directed to the abstract idea of comparing two biological data sets collected at different time points, applying a machine-learning classifier to predict a treatment response, performing a statistical test to identify differentially expressed proteins, computing a probability-based score (RAP score), filtering proteins by score, and identifying intervention targets. The additional elements-generic computer hardware (processor, storage medium, program instructions), specific biological sample types, specific disease categories, specific therapy types, and well-known statistical/ML techniques-are either conventional implementations of the abstract idea or mere field-of-use limitations. None of these elements, individually or as an ordered combination, transform the abstract idea into a patent-eligible practical application or provide an inventive concept sufficient to overcome the Alice/Mayo framework.
Accordingly, claims 1, 3, 5, 7, 9, 10, 14, 15, 16, 18, 31, 32, 34, 35, 38, 39, 40, 41, 42, and 43 are rejected under 35 U.S.C. § 101 as being directed to non-statutory subject matter.
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 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 date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1, 3, 5, 7, 9, 10, 14, 15, 16, 18, 31, 32, 34, 35, 38, 39, 40, 41, 42, and 43 are rejected under 35 U.S.C. § 103 as being unpatentable over Regev (WO 2019/070755 A1) in view of Curiel1 (US 2021/0332136 A1) and further in view of Vermeire2 (US 2022/0364171 A1).
I. Rejection of Claims 1, 18, 31, 32, 34, 35, 38, 39, 40, 41, 42, 43 (Independent and Dependent Claims)
A. Claim 1 / Claim 18 – Limitation-by-Limitation Mapping
1[a] / 18[a]: "A system for predicting a response in a target subject to a specified therapy" / "A method of predicting a response in a target subject to a specified therapy"
Regev discloses a system/method for predicting response to immunotherapy (immune checkpoint inhibitors, ICI). Regev [0171]: "the immunotherapy resistance signature is referred to as 'ITR,' 'immunotherapy resistance signature,' 'ICR,' 'immune checkpoint inhibitor resistance,' 'mICR,' 'malignant immune checkpoint inhibitor resistance,' 'PIT,' 'post-immunotherapy,' 'oncogenic-ICR,' 'unified-ICR,' 'uICR,' 'uICR-up,' 'uICR-down,' 'refined uICR,' 'immune resistant,' 'refined immune resistant,' 'functional immune resistance,' 'post treatment,' 'exclusion-up,' or 'exclusion-down.' All of these terms may be used in reference to a gene signature in malignant cells from a subject that is resistant to immune checkpoint inhibitors (ICI)."
Regev [0172]: "the term 'cancer-specific survival' refers to the percentage of patients with a specific type and stage of cancer who have not died from their cancer during a certain period of time after diagnosis."
1[b]: "at least one hardware processor; and a non-transitory computer-readable storage medium having stored thereon program instructions, the program instructions executable by the at least one hardware processor to:"
This is a standard computer-implemented system limitation. Regev discloses a computational system for analyzing gene expression data and predicting immunotherapy response. Regev [0171]: "a gene signature in malignant cells from a subject that is resistant to immune checkpoint inhibitors (ICI)" implies a computational analysis system. The use of a hardware processor and non-transitory computer3-readable storage medium is well-known in the art for implementing any computational method, and no specific hardware limitation is claimed beyond the generic "at least one hardware processor."
1[c] / 18[c]: "receive, for a target subject having a specified type of disease and receiving a specified therapy for treating said disease, (a) a first biological signature associated with a biological sample collected at a first time point relative to said specified therapy, and (b) a second biological signature associated with a biological sample collected at a second time point relative to said specified therapy"
Regev discloses receiving biological signatures at two distinct time points relative to immunotherapy. Regev [0171]: "In regards to 'oncogenic ICR,' 'mICR,' 'malignant immune checkpoint inhibitor resistance,' 'Post-treatment,' 'PIT,' or 'post-immunotherapy,' these terms all refer to genes differentially expressed in malignant cells after immunotherapy." This is the second time point (post-treatment). Regev [0171]: "In regards to the exclusion signatures, these signatures refer to signatures in malignant cells that correlate to immune cell exclusion. In other words, exclusion-up refers to genes that are upregulated in malignant cells and that are correlated with exclusion, while exclusion-down refer to genes downregulated in malignant cells that are correlated with exclusion." This is the first time point (baseline/pre-treatment).
Curiel further discloses the two-time-point framework in the context of cancer and immunotherapy. Curiel [0132]: "At baseline, tumor cytoplasmic PD-L1 was lower in responders versus non-responders. Strikingly, cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders (FIG. 17B) more easily seen as the significant drop in cytoplasmic/surface ratio in responders versus a significant increase in non-responders (FIG. 17C) whereas total tumor PD-L1 was essentially unchanged and predicted neither treatment response nor immune cell content."
Curiel [0132]: "In a clinical trial of ovarian cancer patients receiving nivolumab (anti-PD-1), bevacizumab and cyclophosphamide (ClinicalTrials.gov Identifier: NCT02853318), cytoplasmic PD-L1 was measured in tumor biopsies of selected patients… Responders were defined as surviving ≥24 months after treatment initiation before disease progression. Non-responders were defined as surviving <4 months after treatment initiation before disease progression."
1[d] / 18[d]: "calculate, for the target subject, a set of values representing a relation between said first and second biological signatures associated with the target"
Regev discloses calculating a unified signature representing the relation between the baseline (exclusion) and post-treatment (PIT) signatures. Regev [0171]: "All of 'unified-ICR,' 'uICR,' 'uICR-up,' 'uICR-down,' 'refined uICR,' 'refined immune resistant,' 'functional immune resistance' refer to an immunotherapy resistant signature that includes genes from the post immunotherapy and exclusion signatures."
Curiel discloses calculating a ratio representing the relation between two biological measurements. Curiel [0132]: "cytoplasmic versus surface PD-L1 fell significantly… more easily seen as the significant drop in cytoplasmic/surface ratio in responders versus a significant increase in non-responders (FIG. 17C)."
Curiel [0132]: "Fold-change in cell prevalence at baseline versus on-treatment" (FIG. 19A description).
1[e] / 18[e]: "apply a pretrained Machine Learning (ML)-based classifier to said set of values, to predict a response in the target subject to the specified therapy"
Regev does not explicitly disclose as disclosed by Vermeire, Vermeire disclosing applying a pretrained ML-based classifier to predict response to therapy. Vermeire [0460]: "The gene expression data layers identified as the contributing -omic layers were subjected to feature selection and machine learning based predictive modelling using the DaMirSeq R package (M. Chiesa, et al. Bioinformatics 34, 1416-1418 (2018)), to identify genes whose expression signatures can discern patient groups (patients with/without endoscopic response, patients with/without reduction of faecal calprotectin)."
Vermeire [0477]: "The top 5 features selected based on the score generated by the multi-variate filter RReliefF (ELOVL4, FGL2, CTSW, DDX11, LYZ), were used for performing the Machine-Learning based classification using multiple classifiers including a stacked ensemble classifier. Mean accuracy rates for the 5 selected features from the CD4+ transcriptomic datasets using a 6-stack ensemble classifier were reported at 92%."
Vermeire [0460]: "In order to avoid over-fitting, the samples were first split into a training (TR1) and test set (TS 1) by bootstrap sampling. Another pair of training (TR2) and test set (TS2) were obtained from TR1. While TR2 is used to train the six individual classifiers, TS2 was used to test their accuracies."
Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
1[f] / 18[f]: "when the target subject is predicted as a non-responder subject, apply a statistical test to select a subset of Differentially Expressed Proteins (DEPs), whose expression levels change between the first time point and second time point"
Regev discloses selecting differentially expressed genes (analogous to DEPs) between responders and non-responders. Regev [0171]: "exclusion-up refers to genes that are upregulated in malignant cells and that are correlated with exclusion, while exclusion-down refer to genes downregulated in malignant cells that are correlated with exclusion."
Regev [0171]: "In regards to 'oncogenic ICR,' 'mICR,' 'malignant immune checkpoint inhibitor resistance,' 'Post-treatment,' 'PIT,' or 'post-immunotherapy,' these terms all refer to genes differentially expressed in malignant cells after immunotherapy."
Regev4 does not explicitly disclose as disclosed by Curiel the use of statistical tests to distinguish responders from non-responders. Curiel [0132]: "The mean of 3 regions of interest/patient was assessed by 2-way ANOVA with Tukey's multiple comparisons test."
Curiel [0132]: "cytoplasmic PD-L1 negatively predicted CD3+ T cell increase on treatment (R=-0.88, p=0.004)" – a statistical correlation test.
Curiel [0120]: "In vivo tumor growth was analyzed by 2-way ANOVA plus Bonferroni post-tests. All other single measurement assays were assessed with unpaired student's t test. P values <0.05 were considered significant."
1[g] / 18[g]: "calculate, for each DEP of the target subset, a Resistance Associated Protein (RAP) score, representing a deviation from a distribution of protein expression in at least one group selected from (i) subjects responding to the specified therapy, and (ii) subjects not responding to the specified therapy"
Regev discloses calculating a score representing deviation from a distribution in responder vs. non-responder groups. Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
Regev [0171]: "All of 'unified-ICR,' 'uICR,' 'uICR-up,' 'uICR-down,' 'refined uICR,' 'refined immune resistant,' 'functional immune resistance' refer to an immunotherapy resistant signature that includes genes from the post immunotherapy and exclusion signatures."
Curiel discloses a score representing deviation from a distribution. Curiel [0132]: "cytoplasmic PD-L1 negatively predicted CD3+ T cell increase on treatment (R=-0.88, p=0.004)" – a correlation coefficient representing deviation from the expected distribution.
Curiel [0008]: "increased expression of any of the 5 LAMTOR proteins (referred to herein as 'LAMTOR') was observed to be associated with poor prognosis and reduced survival in cancers" – a score associated with deviation in a non-responder group.
1[h] / 18[h]: "filter the DEPs based on the associated RAP scores"
Regev discloses filtering genes based on their performance scores. Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
Regev does not explicitly disclose as disclosed by Vermeire, Vermeire explicitly disclosing filtering features based on scores. Vermeire [0477]: "After dimensionality reduction using feature prioritization, we identified 30 CD4+ expressed genes, which could explain the endoscopic remission-based segregation of patients. The top 5 features selected based on the score generated by the multi-variate filter RReliefF (ELOVL4, FGL2, CTSW, DDX11, LYZ), were used for performing the Machine-Learning based classification."
1[i] / 18[i]: "indicate one or more filtered DEPs as intervention targets for the target non-responder subject"
Regev does not explicitly disclose as disclosed by Curiel; Curiel discloses indicating specific proteins as intervention targets for non-responders. Curiel [0008]: "methods of predicting cancer treatment responses and patient prognosis based on expression of tumor cytoplasmic PD-L1, or the ratio of cytoplasmic PD-L1 to surface-expressed PD-L1, are also provided."
Curiel [0009]: "An analysis of tumors obtained from a clinical trial of pembrolizumab, bevacizumab plus cyclophosphamide to treat patients with epithelial ovarian cancer revealed that tumor cytoplasmic PD-L1 predicted the clinical response of patients to the therapy, whereas surface-expressed PD-L1 and total tumor PD-L1 were uninformative for predicting clinical response. Considering that epithelial ovarian cancers typically do not respond to anti-PD-1 or anti-PD-L1 immunotherapy alone, such methods may be used, e.g., to identify patients with epithelial ovarian cancers can benefit from an anti-PD-1 or anti-PD-L1 immunotherapy."
Curiel [0108]: "tumor PD-L1 may be used as a biomarker for response to DDRi agents alone or combined with chemotherapy or immunotherapy."
B. Claim 3 – Dependent on Claim 1
3[a]: "said first and second biological signatures are each protein expression profiles"
Curiel discloses protein expression profiles as biological signatures. Curiel [0132]: "Tumor biopsies were formalin-fixed, paraffin-embedded tumors, cut into 4 micron slices and stained with specific antibodies coupled to barcoded UV-photocleavable oligonucleotides and fluorescent tags from NanoString. Percent surface and cytoplasmic PD-L1 was determined by assessing oligonucleotide location in relationship to tumors in selected regions of interest using FIJI software at 0.325 µm/pixel resolution to map x-y coordinates through the focal plane."
Curiel [0078]: "cytoplasmic PD-L1 and/or surface PD-L1 can be detected or measured in a tissue sample or cell using an immunological or antibody-based detection method."
3[b]: "said sets of values each comprise, with respect to each protein in said protein expression profiles, a ratio or a difference in the levels of expression of said protein in said first and second biological signatures"
Curiel discloses calculating a ratio between two time points. Curiel [0132]: "cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders (FIG. 17B) more easily seen as the significant drop in cytoplasmic/surface ratio in responders versus a significant increase in non-responders (FIG. 17C)."
Curiel [0132]: "Fold-change in cell prevalence at baseline versus on-treatment" (FIG. 19A).
3[c]: "wherein said protein expression profile comprises expression values for at least two proteins"
Curiel [0008]: "increased expression of any of the 5 LAMTOR proteins (referred to herein as 'LAMTOR') was observed to be associated with poor prognosis and reduced survival in cancers." Curiel [0083]: "Lamtor proteins include LAMTOR1, LAMTOR2, LAMTOR3, LAMTOR4, and LAMTOR5."
C. Claim 5 – Dependent on Claim 1
5[a]: "said program instructions are further executable to perform a dimensionality reduction stage with respect to said sets of values, to reduce the number of variables in each of said sets of values, and wherein said dimensionality reduction stage identifies a subset of principal proteins in each of said sets of values"
Vermeire discloses a dimensionality reduction / feature selection stage. Vermeire [0477]: "To test and verify this, we used feature reduction techniques to infer genes within this particular gene expression layer which could properly differentiate patients who did and did not endoscopically respond to anti-TNF. After dimensionality reduction using feature prioritization, we identified 30 CD4+ expressed genes, which could explain the endoscopic remission-based segregation of patients."
Vermeire [0460]: "The gene expression data layers identified as the contributing -omic layers were subjected to feature selection and machine learning based predictive modelling using the DaMirSeq R package… to identify genes whose expression signatures can discern patient groups."
D. Claim 7 – Dependent on Claim 5
7[a]: "said training set comprises only said subset of principal proteins in each of said sets of values"
Vermeire discloses training a classifier using only the selected subset of features. Vermeire [0477]: "The top 5 features selected based on the score generated by the multi-variate filter RReliefF (ELOVL4, FGL2, CTSW, DDX11, LYZ), were used for performing the Machine-Learning based classification using multiple classifiers including a stacked ensemble classifier."
Vermeire [0460]: "While TR2 is used to train the six individual classifiers, TS2 was used to test their accuracies."
E. Claim 9 – Dependent on Claim 1
9[a]: "each of said biological samples is one of: blood plasma, whole blood, blood serum, cerebrospinal fluid (CSF), and peripheral blood mononuclear cells (PBMCs)"
Curiel discloses biological samples for measuring PD-L1. Curiel [0079]: "cytoplasmic PD-L1 and/or surface PD-L1 can be detected or measured in a tissue sample or cell using an immunological or antibody-based detection method." Curiel [0079]: "immunohistochemistry or confocal microscopy can be used to detect or measure intracellular PD-L1. Cell-fractionation may be combined with an immunological detection method such as a Western blot."
Curiel [0152]: "Mice were sacrificed by cervical dislocation after induction of deep isoflurane anesthesia. ID8agg tumors were collected and cells were stained as described… using LSR II hardware and analyzed by FACS Diva (BD Bioscience) or Flow Jo software."
The specific sample types (blood plasma, whole blood, etc.) are well-known biological sample types in the art for proteomic analysis, and Curiel's disclosure of using tumor biopsies and cell-based assays is analogous5.
F. Claim 10 – Dependent on Claim 1
10[a]: "said specified type of disease is a proliferative disease or cancer"
Regev [0172]: "the term 'cancer-specific survival' refers to the percentage of patients with a specific type and stage of cancer who have not died from their cancer during a certain period of time after diagnosis."
Curiel [0004]: "Programmed death ligand-1 (PD-L1, CD274, B7-H1), a target of several FDA-approved cancer immunotherapies, is highly expressed on some cancers and enables immune evasion by inhibiting tumor-specific CD8+ T cells."
Curiel [0007]: "Cancers continue to present a serious, often complex clinical problem that results in significant mortality. Clearly, there is a need for new and improved methods for treating cancers, personalizing cancer therapies, and identifying patients more likely to respond to specific anti-cancer therapies."
G. Claim 14 – Dependent on Claim 1
14[a]: "said predicting comprises an indication of secondary effects in said target subject"
Curiel discloses secondary effects of therapy. Curiel [0106]: "Tumor-intrinsic PD-L1 exerts diverse effects on pivotal cellular survival, metabolic, and signaling pathways including promoting mTOR, tumor initiating (stem) cells, Ras/Mek/Erk, PI3K/AKT, JAK/STAT signaling, and inhibiting autophagy and mRNA degradation."
Curiel [0162]: "In ID8agg tumors, CAMB plus aPD-L1, but not either single treatment, significantly increased tumor-infiltrating activated CD8+ T cells (FIG. 37A, left panel) with enhanced IFN-γ production (FIG. 37A, right panel)."
H. Claim 15 – Dependent on Claim 1
15[a]: "said program instructions are further executable to determine, based, at least in part, on said predicting, at least one of: continuing said specified therapy in said target subject, adjusting said specified therapy in said target subject, discontinuing said specified therapy in said target subject, and administering a different therapy to said target subject"
Curiel discloses determining therapy based on prediction. Curiel [0008]: "methods of predicting cancer treatment responses and patient prognosis based on expression of tumor cytoplasmic PD-L1, or the ratio of cytoplasmic PD-L1 to surface-expressed PD-L1, are also provided."
Curiel [0009]: "such methods may be used, e.g., to identify patients with epithelial ovarian cancers can benefit from an anti-PD-1 or anti-PD-L1 immunotherapy."
Curiel [0108]: "the DDRi could be administered alone to patients with tumors that express little or no PD-L1, and the DDRi can optionally be combined with an additional immunotherapy or chemotherapy."
Curiel [0145]: "aPD-L1 plus rabusertib was essentially 100% lethal in control RT4 cells with no effect on PD-L1KO RT4 cells" – indicating a different therapy (combination) for non-responders.
I. Claim 16 – Dependent on Claim 1
16[a]: "said specified therapy is an immunotherapy, wherein said immunotherapy is selected from anti-PD-1/PD-L1 therapy, anti-CTLA-4 therapy, and both"
Regev [0171]: "a gene signature in malignant cells from a subject that is resistant to immune checkpoint inhibitors (ICI)."
Curiel [0004]: "Programmed death ligand-1 (PD-L1, CD274, B7-H1), a target of several FDA-approved cancer immunotherapies, is highly expressed on some cancers and enables immune evasion by inhibiting tumor-specific CD8+ T cells."
Curiel [0011]: "The immune blockade therapy may be an antibody that selectively binds PD-L1 or PD-1. In some embodiments, the antibody selectively binds PD-1. In some embodiments, the antibody is cemiplimab, nivolumab, pembrolizumab, spartalizumab (PDR001), camrelizumab (SHR1210), sintilimab (IBI308), tislelizumab (BGB-A317), toripalimab (JS 001), AMP-224, or AMP-514. In some embodiments, the antibody selectively binds PD-L1. In some embodiments, the antibody is atezolizumab, avelumab, durvalumab, KN035, CK-301, AUNP12, CA-170, or BMS-986189."
J. Claim 31 – Dependent on Claim 18
31[a]: "further comprising at an inference stage, applying said classifier to a target set of said values associated with a target subject, thereby predicting a response in said target subject to said specified therapy"
Vermeire [0460]: "While TR2 is used to train the six individual classifiers, TS2 was used to test their accuracies."
Vermeire [0477]: "The top 5 features selected based on the score generated by the multi-variate filter RReliefF (ELOVL4, FGL2, CTSW, DDX11, LYZ), were used for performing the Machine-Learning based classification using multiple classifiers including a stacked ensemble classifier. Mean accuracy rates for the 5 selected features from the CD4+ transcriptomic datasets using a 6-stack ensemble classifier were reported at 92%."
Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
K. Claim 32 – Dependent on Claim 31
32[a]: "further comprising determining, based, at least in part, on said predicting, at least one of: continuing said specified therapy in said target subject, adjusting said specified therapy in said target subject, discontinuing said specified therapy in said target subject, and administering a different therapy to said target subject"
Curiel [0009]: "such methods may be used, e.g., to identify patients with epithelial ovarian cancers can benefit from an anti-PD-1 or anti-PD-L1 immunotherapy."
Curiel [0108]: "the DDRi could be administered alone to patients with tumors that express little or no PD-L1, and the DDRi can optionally be combined with an additional immunotherapy or chemotherapy."
Curiel [0145]: "aPD-L1 plus rabusertib was essentially 100% lethal in control RT4 cells with no effect on PD-L1KO RT4 cells (FIGS. 27A-B)" – indicating a different therapy (combination) for non-responders.
L. Claim 34 – Dependent on Claim 32
34[a]: "determining DEPs between responders and non-responders"
Regev [0171]: "exclusion-up refers to genes that are upregulated in malignant cells and that are correlated with exclusion, while exclusion-down refer to genes downregulated in malignant cells that are correlated with exclusion."
Curiel [0132]: "At baseline, tumor cytoplasmic PD-L1 was lower in responders versus non-responders. Strikingly, cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders."
34[b]: "determining, in the sample obtained from said subject, one or more RAPs, selected from the DEPs"
Curiel [0008]: "increased expression of any of the 5 LAMTOR proteins (referred to herein as 'LAMTOR') was observed to be associated with poor prognosis and reduced survival in cancers."
Curiel [0132]: "cytoplasmic PD-L1 negatively predicted CD3+ T cell increase on treatment (R=-0.88, p=0.004)" – identifying specific proteins (RAPs) from the DEPs.
34[c]: "identifying a therapy suitable for balancing the level of the one or more RAPs in said subject"
Curiel [0009]: "such methods may be used, e.g., to identify patients with epithelial ovarian cancers can benefit from an anti-PD-1 or anti-PD-L1 immunotherapy."
Curiel [0108]: "tumor PD-L1 may be used as a biomarker for response to DDRi agents alone or combined with chemotherapy or immunotherapy."
Curiel [0145]: "aPD-L1 plus rabusertib was essentially 100% lethal in control RT4 cells" – identifying a therapy (combination of aPD-L1 and Chk1 inhibitor) suitable for the subject.
M. Claim 35 – Dependent on Claim 32
35[a]: "determining the one or more RAPs is by providing a probabilistic measurement of a distance of the DEP expression level from a defined group of samples selected from the responder group or the non-responder group"
Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
Curiel [0132]: "cytoplasmic PD-L1 negatively predicted CD3+ T cell increase on treatment (R=-0.88, p=0.004)" – a probabilistic/correlation-based measurement of distance from a group.
35[b]: "determining the one or more RAPs is by determining the expression distribution of each DEP in each of the responder and non-responder groups, fitting a probability density function for each group, and calculating for each subject, and based on the DEP expression of said subject, the probability of the DEP to be associated with one of the response groups"
Curiel [0132]: "The mean of 3 regions of interest/patient was assessed by 2-way ANOVA with Tukey's multiple comparisons test."
Curiel [0120]: "In vivo tumor growth was analyzed by 2-way ANOVA plus Bonferroni post-tests. All other single measurement assays were assessed with unpaired student's t test. P values <0.05 were considered significant."
Vermeire [0477]: "The top 5 features selected based on the score generated by the multi-variate filter RReliefF (ELOVL4, FGL2, CTSW, DDX11, LYZ), were used for performing the Machine-Learning based classification" – RReliefF is a probabilistic feature selection method that calculates the probability of a feature being associated with a class.
N. Claim 38 – Dependent on Claim 1
38[a]: "each one of said first and second biological signatures is selected from a list consisting of: a DNA profile, an RNA profile, a protein profile, a metabolomics profile, microbiome profile, a genomics profile, a transcriptomics profile, a cellular profile, an epigenomics profile, a post-translational modification-based profile, cellular profile, single-cell based analysis and a regulatory RNA profile"
Regev [0171]: "a gene signature in malignant cells from a subject that is resistant to immune checkpoint inhibitors (ICI)" – gene signatures are RNA/transcriptomic profiles.
Curiel [0004]: "Cancers typically have genomic instability, and mutations in cancers are common."
Curiel [0075]: "Programmed death-ligand 1 (PD-L1) is expressed by many human cancers. Binding of PD-L1 to its receptor PD-1 on activated T cells can inhibit anti-tumor immunity by counteracting T cell-activating signals."
Vermeire [0433]: "Multiple factors are believed to contribute to a patient's response to biological therapies including host genomics, transcriptomics and possibly also metagenomics, metabolomics and proteomics."
Vermeire [0433]: "As artificial intelligence, machine learning and systems biology have now opened the avenue for the efficient integration and interpretation of big datasets, technological advancements including next-generation sequencing, high-throughput omic data generation and molecular networks should be applied to dissect the underlying complexity and predict response to therapy."
O. Claim 39 – Dependent on Claim 1
39[a]: "the at least one processor is configured to calculate the RAP score per DEP of the target subject based on a ratio between a probability that the DEP is attributed to distribution of protein expression in non-responding subjects, and a probability that the DEP is attributed to distribution of protein expression in responding subjects"
Curiel [0132]: "cytoplasmic PD-L1 negatively predicted CD3+ T cell increase on treatment (R=-0.88, p=0.004)" – a ratio/correlation between responder and non-responder distributions.
Curiel [0132]: "At baseline, tumor cytoplasmic PD-L1 was lower in responders versus non-responders. Strikingly, cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders (FIG. 17B) more easily seen as the significant drop in cytoplasmic/surface ratio in responders versus a significant increase in non-responders (FIG. 17C)."
Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
P. Claim 40 – Dependent on Claim 1
40[a]: "the at least one processor is configured to filter the DEPs by selecting DEPs as proteins whose expression levels change between responding subjects and non-responding subjects, and between the first time point and second time point"
Regev [0171]: "exclusion-up refers to genes that are upregulated in malignant cells and that are correlated with exclusion, while exclusion-down refer to genes downregulated in malignant cells that are correlated with exclusion."
Regev [0171]: "In regards to 'oncogenic ICR,' 'mICR,' 'malignant immune checkpoint inhibitor resistance,' 'Post-treatment,' 'PIT,' or 'post-immunotherapy,' these terms all refer to genes differentially expressed in malignant cells after immunotherapy."
Curiel [0132]: "At baseline, tumor cytoplasmic PD-L1 was lower in responders versus non-responders. Strikingly, cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders."
Vermeire [0477]: "After dimensionality reduction using feature prioritization, we identified 30 CD4+ expressed genes, which could explain the endoscopic remission-based segregation of patients."
Q. Claim 41 – Dependent on Claim 1
41[a]: "the at least one processor is configured to filter the DEPs by selecting DEPs that are targeted by available drugs or Investigational New Drugs (INDs)"
Curiel [0008]: "methods of predicting cancer treatment responses and patient prognosis based on expression of tumor cytoplasmic PD-L1, or the ratio of cytoplasmic PD-L1 to surface-expressed PD-L1, are also provided."
Curiel [0011]: "The DDR inhibitor may be a Chk1 inhibitor (Chk1i), a PARP inhibitor (PARPi), ATM inhibitor (ATMi), or an ATR inhibitor (ATRi). In some embodiments, the PARP inhibitor is rucaprib, olaparib, or niraparib. In some embodiments, the ATM inhibitor is AZD0156 or KU-55933. In some embodiments, the ATR inhibitor is VE-821, AZD6738, or VX970. In some embodiments, the Chk1 inhibitor is MK8776 (SCH900776), LY2603618, CCT245737, or GDC-0575."
Curiel [0108]: "tumor PD-L1 may be used as a biomarker for response to DDRi agents alone or combined with chemotherapy or immunotherapy."
R. Claim 42 – Dependent on Claim 1
42[a]: "receive, for each of a plurality of subjects having the specified type of disease and receiving the specified therapy (a) a respective first biological signature and (b) a respective second biological signature"
Curiel [0132]: "We tested 3 responder and 3 non-responder ovarian cancer patients. Tumor biopsies were formalin-fixed, paraffin-embedded tumors, cut into 4 micron slices and stained with specific antibodies coupled to barcoded UV-photocleavable oligonucleotides and fluorescent tags from NanoString."
Regev [0481]: "the program is detected in TCGA tumors, which were all untreated. Second, while the program is more predominant in the malignant cells of the post-treatment resistant patients, it is also overexpressed in a subset of the malignant cells from untreated patients."
42[b]: "calculate, for each of the plurality of subjects, a respective set of values representing a relation between the respective first and second biological signatures"
Curiel [0132]: "Fold-change in cell prevalence at baseline versus on-treatment" (FIG. 19A).
Curiel [0132]: "cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders (FIG. 17B) more easily seen as the significant drop in cytoplasmic/surface ratio in responders versus a significant increase in non-responders (FIG. 17C)."
42[c]: "train the classifier on a training set comprising: (i) the calculated sets of values, and (ii) labels associated with an outcome of said specified therapy in each of said subjects, to predict a response in the target subject to said specified therapy"
Vermeire [0460]: "In order to avoid over-fitting, the samples were first split into a training (TR1) and test set (TS 1) by bootstrap sampling. Another pair of training (TR2) and test set (TS2) were obtained from TR1. While TR2 is used to train the six individual classifiers, TS2 was used to test their accuracies."
Vermeire [0477]: "The top 5 features selected based on the score generated by the multi-variate filter RReliefF (ELOVL4, FGL2, CTSW, DDX11, LYZ), were used for performing the Machine-Learning based classification using multiple classifiers including a stacked ensemble classifier. Mean accuracy rates for the 5 selected features from the CD4+ transcriptomic datasets using a 6-stack ensemble classifier were reported at 92% (Table 12", FIG. 4A")."
Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
S. Claim 43 – Dependent on Claim 42
43[a]: "said training set further comprises, with respect to at least some of said subjects, labels associated with clinical data"
Curiel [0132]: "Responders were defined as surviving ≥24 months after treatment initiation before disease progression. Non-responders were defined as surviving <4 months after treatment initiation before disease progression."
Curiel [0123]: "Tumor biopsies from 3 responder (R) and 3 non-responder (NR) ovarian cancer patients in a trial of pembrolizumab plus cyclophosphamide plus bevacizumab were studied. Responders were defined as surviving >24 months after treatment initiation before disease progression. Non-responders were defined as surviving <4 months after treatment initiation before disease progression."
43[b]: "wherein the statistical test is selected from a student's t-test, an Analysis of Variance (ANOVA) test and a Kolmogorov-Smirnov test"
Curiel [0120]: "In vivo tumor growth was analyzed by 2-way ANOVA plus Bonferroni post-tests. All other single measurement assays were assessed with unpaired student's t test. P values <0.05 were considered significant."
Curiel [0132]: "The mean of 3 regions of interest/patient was assessed by 2-way ANOVA with Tukey's multiple comparisons test."
Curiel [0155]: "We used two-way ANOVA corrected with Bonferroni post-tests to compare repeated means in tumor growth analyses. The log-rank test was used to analyze mouse survival from survival plots generated with the method of Kaplan-Meier. We used an unpaired t-test for other analyses and considered P<0.05 as statistically significant."
II. Rationale for Obviousness (35 U.S.C. § 103)
It would have been obvious to one of ordinary skill in the art before the effective filing date to combine the teachings of Regev, Curiel, and Vermeire for the following reasons:
Regev provides the foundational framework for predicting immunotherapy (ICI) response using gene signatures at two time points (baseline/exclusion and post-treatment/PIT), and the concept of a "refined" or "functional" immune resistance signature that filters the best-performing genes for predicting sensitivity. Regev [0171]: "The 'functional immune resistance,' 'refined uICR' and 'refined immune resistant' signatures are shortened lists from the immune resistance signature that include the best performing genes from the exclusion and post treatment signatures for predicting immunotherapy sensitivity."
Curiel provides the specific application to cancer (a proliferative disease) with anti-PD-1/PD-L1 immunotherapy, the two-time-point measurement (baseline vs. on-treatment), the use of statistical tests (ANOVA, t-test) to distinguish responders from non-responders, and the identification of specific proteins (PD-L1, LAMTORs, Chk2) as biomarkers and intervention targets. Curiel [0132]: "At baseline, tumor cytoplasmic PD-L1 was lower in responders versus non-responders. Strikingly, cytoplasmic versus surface PD-L1 fell significantly, whereas it rose significantly in non-responders."
Vermeire provides the specific ML-based classifier implementation (DaMirSeq R package, 6-stack ensemble classifier), the feature selection/dimensionality reduction (RReliefF), and the training/test set methodology. Vermeire [0460]: "The gene expression data layers identified as the contributing -omic layers were subjected to feature selection and machine learning based predictive modelling using the DaMirSeq R package… to identify genes whose expression signatures can discern patient groups."
The combination is motivated by the common goal of predicting therapy response in cancer patients. Regev [0481]: "Applicants hypothesized that the immune resistance program, while more pronounced in the malignant cell of patients after ICI, in fact reflects an intrinsic resistance mechanism, present even before immunotherapy." Curiel [0007]: "Cancers continue to present a serious, often complex clinical problem that results in significant mortality. Clearly, there is a need for new and improved methods for treating cancers, personalizing cancer therapies, and identifying patients more likely to respond to specific anti-cancer therapies." Vermeire [0433]: "As artificial intelligence, machine learning and systems biology have now opened the avenue for the efficient integration and interpretation of big datasets, technological advancements including next-generation sequencing, high-throughput omic data generation and molecular networks should be applied to dissect the underlying complexity and predict response to therapy."
The "RAP score" as a deviation from a distribution in responder vs. non-responder groups is a straightforward application of the statistical methods (ANOVA, t-test, correlation) already disclosed in Curiel [0120] and [0132], combined with the "refined" gene selection concept of Regev [0171]. The "intervention targets" are the specific proteins (PD-L1, LAMTORs, Chk2) identified in Curiel [0008] and [0108] as biomarkers for therapy selection.
III. Summary of Rejections
Claim(s)
Rejection Basis
Primary Ref.
Secondary Ref(s)
1, 18
§ 103
Regev
Curiel, Vermeire
3
§ 103
Regev
Curiel, Vermeire
5, 7
§ 103
Regev
Curiel, Vermeire
9
§ 103
Regev
Curiel, Vermeire
10
§ 103
Regev
Curiel, Vermeire
14
§ 103
Regev
Curiel, Vermeire
15, 32
§ 103
Regev
Curiel, Vermeire
16
§ 103
Regev
Curiel, Vermeire
31
§ 103
Regev
Curiel, Vermeire
34
§ 103
Regev
Curiel, Vermeire
35
§ 103
Regev
Curiel, Vermeire
38
§ 103
Regev
Curiel, Vermeire
39
§ 103
Regev
Curiel, Vermeire
40
§ 103
Regev
Curiel, Vermeire
41
§ 103
Regev
Curiel, Vermeire
42, 43
§ 103
Regev
Curiel, Vermeire
Response To Arguments
Applicant contentions are fully considered and responded as follows:
101 Arguments
Regarding 101 rejections Applicant contentions are moot in view of updated 101 rejections pursuant to Applicant Amendments.
103 Arguments
Applicant contentions are moot in view of Applicant claim amendments and pursuant to change in grounds from said amendments.
Conclusion
This action is a final rejection and closes the prosecution of this application. Applicant’s reply under 37 CFR 1.113 to this action is limited to an appeal to the Patent Trial and Appeal Board, an amendment complying with the requirements set forth below, or a request for continued examination (RCE) to reopen prosecution where permitted.
General information on the Patent Trial and Appeal Board is available at: www.uspto.gov/patents/patent-trial- and-appeal-board/about-ptab/new-ptab. The information at this page includes guidance on time limited options that may assist the applicant contemplating appealing an examiner’s rejection. It also includes information on pro bono (free) legal services and advice available for those who are under-resourced and considering an appeal at: https://www.uspto.gov/patents/patent-trial-and-appeal-board/patent-trial-and-appeal-board-pro-bono-program-independent. The page is best reviewed promptly after applicant has received a final rejection or the claims have been twice rejected because some of the noted assistance must be requested within one month from the date of the latest rejection. See MPEP § 1204 for more information on filing a notice of appeal.
If applicant should desire to appeal any rejection made by the examiner, a Notice of Appeal must be filed within the period for reply. The Notice of Appeal must be accompanied by the fee required by 37 CFR 41.20(b)(1). The current fee amount is available at: www.uspto.gov/Fees.
If applicant should desire to file an after-final amendment, entry of the proposed amendment cannot be made as a matter of right unless it merely cancels claims or complies with a formal requirement made in a previous Office action. Amendments touching the merits of the application which otherwise might not be proper may be admitted upon a showing of good and sufficient reasons why they are necessary and why they were not presented earlier.
A reply under 37 CFR 1.113 to a final rejection must include cancellation of or appeal from the rejection of, each rejected claim. The filing of an amendment after final rejection, whether or not it is entered, does not stop the running of the statutory period for reply to the final rejection unless the examiner holds all of the claims to be in condition for allowance.
If applicant should desire to continue prosecution in a utility or plant application filed on or after May 29, 2000 and have the finality of this Office action withdrawn, an RCE under 37 CFR 1.114 may be filed within the period for reply. See MPEP § 706.07(h) for more information on the requirements for filing an RCE.
The application will become abandoned unless a Notice of Appeal, an after final replay that places the application in condition for allowance, or an RCE has been filed properly within the period for reply, or any extension of this period obtained under either 37 CFR 1.136(a) or (b).
The prior art made of record6 and NOT relied upon is considered pertinent to applicant's disclosure:
Davicioni (US 2021/0130902)
The present invention relates to methods, systems and kits for the diagnosis, prognosis and the determination of progression of cancer in a subject. The invention also provides biomarkers that define subgroups of prostate cancer, clinically useful classifiers for distinguishing prostate cancer subtypes, bioinformatic methods for determining clinically useful classifiers, and methods of use of each of the foregoing. The methods, systems and kits can provide expression-based analysis of biomarkers for purposes of subtyping prostate cancer in a subject. Further disclosed herein, in certain instances, are probe sets for use in subtyping prostate cancer in a subject. Classifiers for subtyping a prostate cancer are provided. Methods of treating cancer based on molecular subtyping are also provided.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHAEL EZEWOKO whose telephone number is 571 272 7850. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Fonya Long can be reached on 571 270 5096. The fax phone number for the organization where this application or proceeding is assigned is 571-273-7850.
Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/MICHAEL I EZEWOKO/Primary Examiner, Art Unit 3682
1 See Section “II. Rationale for Obviousness (35 U.S.C. § 103)” for obviousness rationale
2 See Section “II. Rationale for Obviousness (35 U.S.C. § 103)” for obviousness rationale
3 Regev discloses a computer capable of running computer instructions [0363]
4 However Regev does disclose the use of ANOVA between treated and untreated populations [0527]
5 Regev discloses collections of biopsies [0389], [0283]: biopsies and associated blood source
6Please see Form 892 for complete listing
