DETAILED ACTION
Status of Claims
This action is in reply to the application filed on 30 May, 2025.
Claims 1 – 20 are currently pending and have been examined.
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 .
Summary of the Invention
The specification describes conventional techniques that use machine learning models, trained to predict a characteristic of a subject. The specification describes the problem associated with these conventional techniques that the invention purports to solve. In particular, that there is insufficient training data available; and conventional techniques for acquiring such data involve physically sequencing blood samples, which is inefficient, time consuming, and may be inconsistently performed. The invention purports to solve the problem by receiving, from a data store, cfRNA expression data previously obtained from a biological fluid sample, and using the obtained data as training data. The specification teaches that the use of previously obtained data improves the inefficiency of collecting data without requiring the obtaining, preparing and sequencing of blood samples to obtain the same data. (@ 0122, 0123 in the specification as published (PGPUB 2025/0372255 A1)) The claims are directed to generating training data – i.e. artificial cfRNA expression data – that combines healthy expression profiles and tumor expression profiles that were “previously obtained from biological samples”.
Notably, receiving cfRNA expression data from a data store inherently requires first: obtaining, preparing and sequencing the biological samples. As such, it is unclear to the Examiner how the timing of sample collection and sequencing has any effect on the efficiency of collection the information. Sample collection and sequencing is required for both the data store RNA expression data and the contemporaneously collected RNA expression data.
Claim Interpretation
The claims recite terms that are subject to interpretation in order to obtain a sufficient understanding of the invention.
For example, the claims recite the term: “characteristic of the subject”. The specification discloses examples of characteristics that may be determined including: that the subject has, or does not have a disease or condition such as cancer; that the subject has, or does not have, liver metastasis; a predicted fraction of malignant B cells relative to the total number of B cells; a predicted PD-1 status, a predicted response to a particular therapy, a predicted survival prognosis, whether the subject has minimal residual disease (MRD), whether the subject is likely to experience an adverse event, and “other characteristics”. (@ 0091 – 0094, 0121, 0139) The claims allow for a broad interpretation of the term “characteristic” to include ANY characteristic of the subject.
The claims recite the term: “artificial” cfRNA expression data. The term “artificial data” is commonly referred to as “synthetic data” that is artificially generated and mimics real world data without containing any actual information. (Wikipedia) The claims use the term “artificial” even though the data is obtained from real-world data sources and contains actual information – “a plurality of RNA expression profiles previously obtained from biological samples from healthy subjects”. Nonetheless, Applicants can be their own lexicographer. Here, the term “artificial cfRNA expression data” denotes data obtained from a plurality of subjects that was used to train the model; and is distinguished in the claim from “cfRNA expression data”, which denotes data from the subject of interest that is processed by the trained model to determine a characteristic of the subject. As such, the term “artificial cfRNA expression data” is construed by the Examiner to be what is commonly known as “training data” for a machine learning model.
The claims recite the term “combining” relative to generating the healthy expression profile and the artificial cfRNA expression profile. Combining a plurality of RNA expression profiles is described in the specification as “determining a weighted sum of the plurality of RNA expression profiles”. Similarly, combining a healthy expression profile and a tumor expression profile is described as “determining a weighted sum of the healthy expression profile component and the tumor expression profile component”, such as by using a plurality of gene counts. (@ 0048 - 0050) Weighted sums may be determined using equations expressly disclosed in the specification (Equation 1 – 3), although the claims are not limited with respect to how profiles are combined.
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 a judicial exception (i.e. a law of nature, a natural phenomenon, or an abstract idea), and does not include additional elements that either: 1) integrate the abstract idea into a practical application, or 2) that provide an inventive concept – i.e. element that amount to significantly more than the abstract idea. The Claims are directed to an abstract idea because, when considered as a whole, the plain focus of the claims is on an abstract idea.
Claim 1 is representative. Claim 1 recites:
A method of predicting a characteristic of a subject based on cell-free RNA (cfRNA) expression data previously-obtained from a biological fluid sample from the subject, the method comprising:
using at least one computer hardware processor to perform:
obtaining the cfRNA expression data; and
processing the cfRNA expression data using a machine learning model trained to process cfRNA expression data from a subject and produce an output indicative of the characteristic of the subject,
wherein the machine learning model was trained using artificial cfRNA expression data,
the artificial cfRNA expression data comprising a plurality of artificial cfRNA expression profiles,
an artificial cfRNA expression profile of the plurality of artificial cfRNA expression profiles having been generated by:
generating a healthy expression profile component by:
receiving a plurality of RNA expression profiles previously-obtained from biological samples from healthy subjects,
the plurality of RNA expression profiles including a respective RNA expression profile for each of one or more cell types and/or each of one or more types of cell-containing samples; and
generating the healthy expression profile component by combining the plurality of RNA expression profiles;
generating a tumor expression profile component; and
generating the artificial cfRNA expression profile by combining the healthy expression profile component and the tumor expression profile component.
Claim 20 recites non-transitory medium with instructions executed by a processor, and Claim 19 recites a system that executes the steps of the method recited in Claim 1.
STEP 1
The claims are directed to a system, a method, and non-transitory computer readable medium which are included in the statutory categories of invention.
STEP 2A PRONG ONE
The claims, as illustrated by Claim 1, recite limitations that encompass an abstract idea including:
A method of predicting a characteristic of a subject based on cell-free RNA (cfRNA) expression data previously-obtained from a biological fluid sample from the subject, the method comprising:
obtaining the cfRNA expression data; and processing the cfRNA expression data and produce an output indicative of the characteristic of the subject.
The claims, as illustrated by Claim 1, recite limitations that encompass an abstract idea within the “mental processes” grouping – concepts performed in the human mind including observation, evaluation, judgment and opinion. The claims recite obtaining (i.e. receiving) and processing (i.e. analyzing) cfRNA data to produce a characteristic of the subject of interest. The cfRNA data was “previously obtained” using techniques that are well-known to those of ordinary skill in the art. In particular, the specification teaches that the cfRNA data may be obtained from a data store, or by user input. (@ 0157, 0161, 0168, 0178 – 0184) Receiving and analyzing cfRNA expression data, and analyzing the data to produce a characteristic of the subject, is a process that, except for generic computer implementation steps, can be performed in the human mind. One of ordinary skill in the art, if given cfRNA expression data, can recognize the gene expression levels for particular genes that indicate a patient characteristic, including judging that the expression level itself is outside of a normal range.
Collecting information, including when limited to particular content, is within the realm of abstract ideas, and analyzing information by steps people go through in their minds, or by mathematical algorithms, without more, are mental processes within the abstract idea category (Electric Power Group v. Alstom S.A. (Fed Cir, 2015-1778, 8/1/2016).
As such, the claims recite an abstract idea within the mental process grouping.
The claims, as illustrated by Claim 1, recite limitations that encompass an abstract idea within the “certain methods of organizing human activity” grouping –
managing personal behavior or relationships or interactions between people including social activities, teaching, and following rules or instructions.
The claims recite predicting a characteristic of a subject based on cfRNA expression data. This process is typical in medicine, where a doctor orders genetic sequencing tests which are “read” to determine if any RNA expression levels indicate a patient characteristic, and is process that merely organizes this human activity. (See MPEP 2016.04 (a)(2) II C finding that “a mental process that a neurologist should follow when testing a patient for nervous system malfunctions” is a method of organizing human activity, In re Meyer, 688 F.2d 789, 791-93, 215 USPQ 193, 194-96 (CCPA 1982).
As such, the claims recite an abstract idea within the certain methods of organizing human activity grouping.
The claims, as illustrated by Claim 1, also recite limitations that encompass an abstract idea within the law of nature or natural phenomenon grouping.
The claims start with a biological fluid sample comprising a cell-free RNA – a naturally occurring material that circulates freely in the patient’s blood stream. The cfRNA expression levels are determined and analyzed to determine a characteristic of the subject. The claims rely on the natural relationship between the cfRNA expression data and respective subject characteristics. As such, the claims recite a law of nature or natural phenomenon.
The claims, as illustrated by Claim 1, recite limitations that encompass an additional abstract idea including:
the artificial cfRNA expression data comprising a plurality of artificial cfRNA expression profiles, an artificial cfRNA expression profile of the plurality of artificial cfRNA expression profiles having been generated by:
generating a healthy expression profile component by: receiving a plurality of RNA expression profiles previously-obtained from biological samples from healthy subjects, the plurality of RNA expression profiles including a respective RNA expression profile for each of one or more cell types and/or each of one or more types of cell-containing samples; and
generating the healthy expression profile component by combining the plurality of RNA expression profiles;
generating a tumor expression profile component; and
generating the artificial cfRNA expression profile by combining the healthy expression profile component and the tumor expression profile component.
The claims, as illustrated by Claim 1, also recite limitations that encompass an abstract idea within the mathematical formula or relationship grouping. The claims generate artificial cfRNA expression data – i.e. training data - by receiving and combining a plurality of healthy profiles, generating a tumor expression profile and combining the health profile and the tumor profile. The cfRNA data was “previously obtained” using techniques that are well-known to those of ordinary skill in the art. In particular, the specification teaches that the cfRNA data may be obtained from a data store, or by user input. (@ 0157, 0161, 0168, 0178 – 0184) The specification discloses that combining includes Combining a plurality of RNA expression profiles is described as “determining a weighted sum of the plurality of RNA expression profiles”. Similarly, combining a healthy expression profile and a tumor expression profile is described as “determining a weighted sum of the healthy expression profile component and the tumor expression profile component”, such as by using a plurality of gene counts. (@ 0048 - 0050) Weighted sums may be determined to expressly disclosed equations (Equation 1 – 3). As such, the claims recite a mathematical formula or relationship.
STEP 2A PRONG TWO
The claims recite limitations that include additional elements beyond those that encompass the abstract idea above including:
using at least one computer hardware processor;
using a machine learning model trained to process cfRNA expression data from a subject wherein the machine learning model was trained using artificial cfRNA expression data.
However, these additional elements do not integrate the abstract idea into a practical application of that idea in accordance with the MPEP. (see MPEP 2106.05)
The computer hardware processor is recited at a high level of generality such that it amounts to no more than instructions to apply the abstract idea using a generic computer component. These elements merely add instructions to implement the abstract idea on a computer, and generally link the abstract idea to a particular technological environment.
In particular, the claims replace the knowledge and experience of a physician by applying established and generic methods of machine learning to an abstract diagnostic process in a new data environment with a specific context – i.e. applying a trained model to the cfRNA expression data of a subject. The specification teaches that the machine learning model may be trained in a training phase to output patient characteristics using cfRNA expression data; using any suitable model and correctly labelled biological information (@ 0033, 0043, 0138, 0158, ). Machine learning limitations reciting broad, functionally described, well-known techniques executed by generic and conventional computing devices do not provide a practical application of the abstract diagnostic process. The Courts have found that the way machine learning works is the inputs are defined, the model is trained; and then the algorithm is actually updated and improved over time based on the input; “Today we hold only that patents that do no more than claim the application of generic machine learning to new data environments, without disclosing improvements to the machine learning models to be applied, are patent ineligible under §101.” Nothing in the claims recites an improvement to the machine learning model itself; rather, the model is merely limited to a particular field of use. (Recentive Analytics, Inc. v. Fox Corp. (Fed. Cir. 2025)).
Nothing in the claim recites specific limitations directed to an improved technology or technological process. Similarly, the specification is silent with respect to these kinds of improvements. Rather, the improvement described in the specification is related to collecting training data; and in particular, to using data that was previously collected instead of performing a sequencing process on a biological sample. This data may be obtained from one or more known databases such as the “Gen Expression Omnibus (GEO) database; the Cancer Genome Atlas (TCGA) database; the BioStudies database; the European Nucleotide Archive (ENA) database.” (@ 0158).
Determining a characteristic of a subject by applying conventional machine learning techniques does not provide an inventive concept. A general purpose computer that applies a judicial exception by use of conventional computer functions, as is the case here, does not qualify as a particular machine, nor does the recitation of a generic computer impose meaningful limits in the claimed process. (see Ultramercial, Inc. v. Hulu, LLC, 772 F.3d 709, 716-17 (Fed. Cir. 2014)). As such, the additional elements recited in the claim do not integrate the abstract diagnostic process into a practical application of that process.
STEP 2B
The additional elements identified above do not amount to significantly more than the abstract diagnostic process. The claims recite “obtaining and receiving” information, (i.e. the cfRNA expression data; a plurality of RNA expression profiles previously-obtained from biological samples from healthy subjects), for example over a network. While receiving and obtaining information is part of the abstract mental process described above, these features may be further characterized as insignificant data gathering – i.e. a well-understood, routine and conventional computer function – i.e. receiving or transmitting data over a network as in Symantec, TLI, OIP and buySAFE.
The additional structural elements or combination of elements in the claims, other than the abstract idea per se, amount to no more than a recitation of generic computer structure (i.e. a computing hardware processor, a trained machine learning model, computer-readable medium). Each of the above components are disclosed in the specification as being purely conventional and/or known in the industry. In particular, the specification describes the machine learning technology a being purely conventional – “any suitable type of machine learning model” may be used as “aspects of the technology are not limited in this respect.” (@ 0138) Because the specification describes these additional elements in general terms, without describing particulars, Examiner concludes that the claim limitations may be broadly, but reasonably construed, as reciting well-understood, routine and conventional computer components and techniques. The specification describes the elements in a manner that indicates that they are sufficiently well-known that the specification does not need to describe the particulars in order to satisfy U.S.C. 112. Considered as an ordered combination the limitations recited in the claims add nothing that is not already present when the steps are considered individually. As such, the additional elements recited in the claim do not provide significantly more than the abstract diagnostic process, or an inventive concept.
The dependent claims add additional features including:
those that merely serve to further narrow the abstract idea above such as:
further limiting the characteristic to a particular type (Claim 2, 12);
further limiting the model to a particular type (Claim 13);
further limiting the number of profiles (Claim 17);
those that recite additional abstract ideas such as:
generating recommendations for diagnostic tests and performing the diagnostic test; (Claim 3 – 6, 8);
determining whether the subject has CLL (Claim 7);
combining profiles (Claim 15, 18);
those that recite well-understood, routine and conventional activity or computer functions such as:
using training data from healthy and tumor patients (Claim 9 - 11);
training the model using training data (Claim 16);
those that recite insignificant extra-solution activities such as:
obtaining cfRNA expression data by sequencing (Claim 14);
or those that are an ancillary part of the abstract idea.
The limitations recited in the dependent claims, in combination with those recited in the independent claims add nothing that integrates the abstract idea into a practical application, or that amounts to significantly more. As such, the additional element do not integrate the abstract idea into a practical application, or provide an inventive concept that transforms the claims into a patent eligible invention.
The apparatus claims are no different from the method claims in substance. “The equivalence of the method, system and media claims is readily apparent.” “The only difference between the claims is the form in which they were drafted.” (Bancorp). The method claims recite the abstract idea implemented on a generic computer, while the apparatus claims recite generic computer components configured to implement the same idea. Specifically, Claims 19 and 20 merely add the generic hardware noted above that nearly every computer will include. The apparatus claim’s requirement that the same method be performed with a programmed computer does not alter the method’s patentability under U.S.C. 101 (In re Grams). Therefore, the claims are rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter.
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.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1, 9, 13 – 16 and 18 – 20 are rejected under 35 U.S.C. 103 as being unpatentable over Applicant’s admission as to the state of the prior art (AOA) in view of Cahan et al.: (US PGPUB 2021/0193267 A1).
CLAIMS 1, 19 and 20
Applicant admits that following features recited in the claims are known in the art:
A method of predicting a characteristic of a subject based on cell-free RNA (cfRNA) expression data previously-obtained from a biological fluid sample from the subject, the method comprising:
using at least one computer hardware processor to perform:
obtaining the cfRNA expression data; and processing the cfRNA expression data using a machine learning model trained to process cfRNA expression data from a subject and produce an output indicative of the characteristic of the subject, wherein the machine learning model was trained using artificial cfRNA expression data, the artificial cfRNA expression data comprising a plurality of artificial cfRNA expression profiles.
In the specification, Applicant discloses: “Conventional techniques for predicting a characteristic of a subject using cfRNA expression data involve processing measured cfRNA expression data using a machine learning model trained to predict the particular characteristic of the subject.” (see US PGPUB 2025/0372255 A1 @ 0122) The disclosure of such conventional techniques inherently requires using at least one computer hardware processor. Further “the process of training the machine learning model is required for any machine learning model . . . and using a machine learning technique necessarily includes iterative training” (see Recentive).
With respect to the following limitations:
an artificial cfRNA expression profile of the plurality of artificial cfRNA expression profiles having been generated by:
generating a healthy expression profile component by: receiving a plurality of RNA expression profiles previously-obtained from biological samples from healthy subjects, the plurality of RNA expression profiles including a respective RNA expression profile for each of one or more cell types and/or each of one or more types of cell-containing samples; and generating the healthy expression profile component by combining the plurality of RNA expression profiles;
generating a tumor expression profile component; and
generating the artificial cfRNA expression profile by combining the healthy expression profile component and the tumor expression profile component. (Cahan 0005, 0006, 0055, 0074, 0077, 0081, 0083, 0092).
The claim requires generating artificial cfRNA expression data for training the machine learning model, (i.e. training data), by combining, (i.e. determining a weighted sum for), a tumor expression profile component and a healthy expression profile component. The healthy expression profile component is generated by combining a plurality of RNA expression profiles of health subjects. Applicant’s disclosure of prior art practices does not expressly include generating training data using RNA expression profiles from healthy and tumor samples.
Cahan teaches a system and method for evaluating cancer models that includes a machine learning classifier, trained to classify biological samples according to their cancer tumor type. Training data used to train the model is generated by receiving RNA expression data obtained from The Cancer Genome Atlas (TCGA), GEO, and other well-known sources for both normal tissue samples and cancer tissue samples. The specification in the present application discloses receiving the RNA expression data from the same databases. (@ 0158) Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model so as to have included generating training data using RNA expression profiles from healthy and tumor samples, including from well-known databases, in accordance with the teaching of Cahan, in order to allow for robust model training.
CLAIM 9
The combination of AOA/Cahan discloses the limitations above relative to Claim 1. Additionally Cahan discloses the following limitations:
wherein the trained machine learning model is a machine learning model that has been trained to predict whether the subject has cancer using training data comprising at least some of the artificial cfRNA expression data; (Cohen 0005, 0006);
the artificial cfRNA expression data including: a first plurality of artificial cfRNA expression profiles generated using a first plurality of healthy expression profile components, and a second plurality of artificial cfRNA expression profiles generated using a second plurality of healthy expression profile components and a plurality of tumor expression profile components, the plurality of tumor expression profile components having been generated using tumor expression profiles from tumor samples obtained from subjects having cancer; (Cahan 0005, 0006, 0055, 0074, 0077, 0081, 0083, 0092).
Cahan teaches a system and method for evaluating cancer models that includes a machine learning classifier, trained to classify biological samples according to their cancer tumor type. Training data used to train the model is generated by receiving RNA expression data obtained from The Cancer Genome Atlas (TCGA), GEO, and other well-known sources for both normal tissue samples and cancer tissue samples. The specification in the present application discloses receiving the RNA expression data from the same databases. (@ 0158) Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model so as to have included generating training data using RNA expression profiles from healthy and tumor samples, including from well-known databases, in accordance with the teaching of Cahan, in order to allow for robust model training.
CLAIMS 13 – 16 and 18
The combination of AOA/Cahan discloses the limitations above relative to Claim 1. Additionally Cahan discloses the following limitations:
wherein the trained machine learning model is a decision tree model, a gradient boosted decision tree model, a linear regression model, a non-linear regression model, a support vector machine, a Gaussian mixture model, a random forest model, or a neural network model; (Cahan 0053);
obtaining the cfRNA expression data from the biological fluid sample from the subject by sequencing the biological fluid sample; (Cahan 0077);
wherein generating the healthy expression profile component by combining the plurality of RNA expression profiles comprises combining the plurality of RNA expression profiles and a cfRNA expression profile previously-obtained from a biological fluid sample from a healthy subject; (Cahan 0005, 0006, 0055, 0074, 0077, 0081, 0083, 0092).
training the trained machine learning model to predict the characteristic of the subject using the artificial cfRNA expression data including the plurality of artificial cfRNA expression profiles; (Cahan 0006 – 0009, 0053, 0058);
generating the artificial cfRNA expression data by generating each particular artificial cfRNA expression profile of the plurality of artificial cfRNA expression profiles; (Cahan 0007, 0008, 0012).
Cahan discloses machine learning models including a neural network model, a decision tree model, a random forest model, a support vector machine, a linear regression model, a non-linear regression model, (i.e. partial least squares regression), and others. Cahan teaches generating training data using RNA expression data obtained from biological samples from healthy subjects and subject with tumors; and training the model with the training data. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model so as to have included one of various known machine learning algorithms and generating training data using RNA expression profiles from healthy and tumor samples, including from well-known databases, in accordance with the teaching of Cahan, in order to allow for robust model training.
CLAIM 17
The combination of AOA/Cahan discloses the limitations above relative to Claim 16. Additionally Cahan discloses the following limitations:
wherein the plurality of artificial cfRNA expression profiles comprise at least 100 artificial cfRNA expression profiles, at least 250 artificial cfRNA expression profiles, at least 500 artificial cfRNA expression profiles, at least 1,000 artificial cfRNA expression profiles, at least 1,500 artificial cfRNA expression profiles, at least 2,000 artificial cfRNA expression profiles, at least 2,500 artificial cfRNA expression profiles, at least 3,000 artificial cfRNA expression profiles, at least 4,000 artificial cfRNA expression profiles, at least 5,000 artificial cfRNA expression profiles, or at least 10,000 artificial cfRNA expression profiles; (Cahan 0074).
Cahan discloses obtaining training data from the TCGA database and the GEO database including 9288 patient profiles. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model disclosed by the Applicant, so as to have included various amounts of training data, in accordance with the teaching of Cahan, in order to allow for robust model training.
CLAIM 2
The combination of AOA/Cahan discloses the limitations above relative to Claim 16. Additionally Cahan discloses the following limitations:
wherein the trained machine learning model is: a machine learning model that has been trained to predict whether the subject has cancer; (Cahan 0051).
AOA discloses training a machine learning model to predict a characteristic of a subject, but not expressly to predict cancer. Cahan discloses training a machine learning model to predict cancer in a subject. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model disclosed by the Applicant, so as to have included training the model to predict cancer, in accordance with the teaching of Cahan, in order to allow for the prediction of various medical conditions.
With respect to the following limitations:
a machine learning model that has been trained to predict whether the subject has liver metastasis;
a machine learning model that has been trained to predict a fraction of malignant B cells relative to total number of B cells in the biological fluid sample from the subject, or
a machine learning model that has been trained to predict a PD-1 status for the subject, wherein the PD-1 status is indicative of whether PDCD 1 is expressed in tumor cells of the subject.
Cahan teaches a machine learning model that can predict cancer using any type of criterion to define the cancer type. (@ 0051) Cahan does not expressly disclose predicting liver metastasis, a fraction of malignant B cells, or a PD-1 status; however, these predictions are optional (i.e. A, B, C or D). A trained model that predict cancer meets the scope of the claim.
Claims 3 – 6, 8 and 10 - 12 are rejected under 35 U.S.C. 103 as being unpatentable over Applicant’s admission as to the state of the prior art (AOA) in view of Cahan et al.: (US PGPUB 2021/0193267 A1) and in view of Official Notice.
CLAIMS 3 and 4
The combination of AOA/Cahan discloses the limitations above relative to Claim 2. Additionally Cahan discloses the following limitations:
when the trained machine learning model is the machine learning model that has been trained to predict whether the subject has cancer and when the output of the trained machine learning model indicates that the subject has the cancer, wherein the cancer is breast cancer or basal breast cancer, (Cahan 0051).
The combination of AOA/Cahan discloses a machine learning model trained to predict a characteristic of a subject, including breast cancer, but does not expressly disclose the following limitations:
generating a recommendation to perform a diagnostic test and/or performing the diagnostic test; wherein the diagnostic test comprises a mammography and/or a biopsy.
Nonetheless, performing diagnostic testing to confirm a cancer diagnosis is notoriously old and well-known, a fact for which Examiner takes Official Notice.
CLAIMS 5, 6 and 8
The combination of AOA/Cahan discloses the limitations above relative to Claim 2. Claims 5, 6 and 8 recite predicting different conditions and generating recommendations specific to the condition.
when the subject has liver metastasis, recommending an ultrasound and/or a biopsy, and/or performing the ultrasound and/or biopsy;
when the fraction of malignant B cells, recommending an anti-cancer treatment; and/or administering the anti-cancer treatment.
when the PD-1 status, recommending anti-cancer treatment and/or administering the anti-cancer treatment.
AOA/Cahan does not expressly disclose predicting the recited conditions or making the recited recommendations; nonetheless, one of ordinary skill in the art would know how to diagnose, confirm and treat these conditions, a fact for which Examiner takes Official Notice. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model disclosed by the Applicant, so as to have included training the model to predict various recited conditions, and to make relevant recommendations, in accordance with the Official Notice taken, in order to allow for providing advice for a range of conditions.
CLAIMS 10 - 12
The combination of AOA/Cahan discloses the limitations above relative to Claim 1. Claims 10 - 12 recite predicting different conditions using training data that includes healthy and tumor expression profiles.
a machine learning model that has been trained to predict whether the subject has liver metastasis;
a machine learning model that has been trained to predict a PD-1 status of the subject;
a machine learning model that has been trained to predict a fraction of malignant B cells relative to a total number of B cells in the biological fluid sample from the subject;
using training data comprising the plurality of artificial cfRNA expression profiles.
AOA/Cahan does not expressly disclose predicting the recited conditions; nonetheless, one of ordinary skill in the art would know how to predict these conditions, a fact for which Examiner takes Official Notice. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model disclosed by the Applicant, so as to have included training the model to predict various recited conditions, and to make relevant recommendations, in accordance with the Official Notice taken, in order to allow for providing advice for a range of conditions.
With respect to the following limitations:
a first plurality of artificial cfRNA expression profiles generated using a first plurality of healthy expression profile components and a first plurality of tumor expression profile components, and a second plurality of artificial cfRNA expression profiles generated using a second plurality of healthy expression profile components and a plurality of tumor expression profile components; (Cahan 0005, 0006, 0055, 0074, 0077, 0081, 0083, 0092).
Cahan teaches a machine learning model, trained to classify biological samples according to their cancer tumor type. Training data used to train the model is generated by receiving RNA expression data obtained from The Cancer Genome Atlas (TCGA), GEO, and other well-known sources for both normal tissue samples and cancer tissue samples. The RNA expression data is from subject who have a tumor, as well as healthy subjects. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model so as to have included generating training data using RNA expression profiles from healthy and tumor samples, including from well-known databases, in accordance with the teaching of Cahan, in order to allow for robust model training.
With respect to the following:
the first plurality of healthy expression profile components having been generated using at least one RNA expression profile previously-obtained from liver tissue/ from tumor samples that express PDCD1 (PDCD1+), and
the second plurality of healthy expression profile components having been generated without using at least one RNA expression profile previously-obtained from liver tissue/ from tumor samples that do not express PDCD1 (PDCD1-).
The limitations above require obtaining samples from liver tissue, when predicting liver metastasis; or samples that express or do not express PDCD1, when predicting a PD-1 status. AOA/Cahan does not disclose obtaining a biological sample from the tissue of interest; nonetheless, Examiner takes Official Notice that one of ordinary skill in the art would know that RNA expression data from liver tissue would be used to predict liver metastasis, and that RNA expression data from tissue samples that do, and do not, express PDCD1 would be used to predict PD-1 status. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model so as to have included generating training data using RNA expression profiles from samples related to the characteristic of interest (i.e. liver metastasis, PD-1 status), in accordance with the Official Notice taken, in order to allow for robust model training.
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over Applicant’s admission as to the state of the prior art (AOA) in view of Cahan et al.: (US PGPUB 2021/0193267 A1) and in view of Official Notice, and in view of Hayes et al.: “Isolation of malignant B cells from patients with chronic lymphocytic leukemia (CLL) for analysis of cell proliferation: Validation of a simplified method suitable for multi-center clinical studies”; June 2010.
CLAIM 7
The combination of AOA/Cahan discloses the limitations above relative to Claim 6. With respect to the following limitation:
determining, based on the fraction of malignant B cells, whether the subject has chronic lymphocytic leukemia (CLL); (Hayes et al. at least Introduction)
AOA/Cahan does not disclose determining CLL based on malignant B cells; however, Hayes does. Diagnosis of CLL using malignant B cell is well-known in the regard. Therefore, it would have been obvious to one of ordinary skill in the art, before the effective filing data of the claimed invention, to have modified the well-known trained machine learning model so as to have included determining CLL for malignant B cell fractions, in accordance with the teachings of Hayes, in order to allow for robust model training.
CONCLUSION
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
US PGPUB 2018/0046754 A1 to Batagov et al discloses a system and method for measuring gene expression used to for survival prediction of a particular subgroup.
Any inquiry of a general nature or relating to the status of this application or concerning this communication or earlier communications from the Examiner should be directed to John A. Pauls whose telephone number is (571) 270-5557. The Examiner can normally be reached on Mon. - Fri. 8:00 - 5:00 Eastern. If attempts to reach the examiner by telephone are unsuccessful, the Examiner’s supervisor, Robert Morgan can be reached at (571) 272-6773.
Official replies to this Office action may now be submitted electronically by registered users of the EFS-Web system. Information on EFS-Web tools is available on the Internet at: http://www.uspto.gov/patents/process/file/efs/guidance/index.jsp. An EFS-Web Quick-Start Guide is available at: http://www.uspto.gov/ebc/portal/efs/quick-start.pdf.
Alternatively, official replies to this Office action may still be submitted by any one of fax, mail, or hand delivery. Faxed replies should be directed to the central fax at (571) 273-8300. Mailed replies should be addressed to “Commissioner for Patents, PO Box 1450, Alexandria, VA 22313-1450.” Hand delivered replies should be delivered to the “Customer Service Window, Randolph Building, 401 Dulany Street, Alexandria, VA 22314.”
/JOHN A PAULS/Primary Examiner, Art Unit 3683
Date: 6 July, 2026
Prosecution Insights