DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Receipt of Applicant’s Amendment filed July 21, 2025, is acknowledged.
Response to Amendment
Claim 18 has been amended. Claims 4, 9, and 20-22 have been canceled. Claims 23-25 are new. Claims 1-3, 5-8, 10-19, and 23-25 are pending and are provided to be examined upon their merits.
Response to Arguments
Applicant’s arguments with respect to claims 1-3, 5-8, 10-19, and 23-25 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument. A response is provided below in bold where appropriate.
Applicant argues 35 USC §101 Rejection, starting pg. 9 of Remarks (footnotes omitted):
The Rejections Under 35 U.S.C. § 101 Should be Withdrawn
In the Office Action, claims 1-20 were rejected under 35 U.S.C. § 101 as being directed to an abstract idea without significantly more. In particular, the Office Action asserted that the independent claims included limitations which, when given their broadest reasonable interpretation, covered performance of the claimed method as a mental process.' It then argued that the limitations which allegedly covered a mental process were not integrated into a practical application because the claims’ hardware were generic components recited at a high level of generality such that they did not meaningfully limit the claims’ and that the claims lacked an inventive concept which ensured that they amounted to significantly more than an abstract idea for the same reason."
In response to these rejections, the applicant notes that, even a claim which recites an abstract idea and is implemented using generic hardware (points which the applicant does not concede are the case for the pending independent claims) can be patent eligible if it provides an improvement to the functioning of a computer or to another technology or technical field. The applicant also notes that under the patent office’s guidance for artificial intelligence inventions, a claim should be treated as providing such an improvement if:
1) The accompanying specification provides a technical explanation of an asserted improvement; and
2) The claim reflects the asserted improvement.”
As set forth below, it is incontrovertible that those requirements are satisfied for the instant application.
Starting with the first requirement, it is undeniable that the specification provides a technical explanation of an asserted improvement. Specifically, the specification explains that conventional approaches to training machine learning models for analyzing pathology images are, slow, inefficient and can lead to increased risks of errors by the model being trained. It also explains that the disclosed technology improves on those conventional approaches by allowing machine learning models to be trained without outlining objects of interest by focusing on a number of objects instead,’ and that combining this number with a weight allows for additional functionality such as aggregating predictions across images. Thus, the specification asserts and provides a technical explanation of an improvement in the process of training machine learning models, and the first requirement set forth in the patent office’s guidance for artificial intelligence inventions is therefore satisfied.
Machine learning model is a generic learning model. Using machine learning is not improving machine learning. For example, if there was an improvement to machine learning technology, the specific type of machine learning (e.g., neural network) and the actual improvement (steps) to the machine learning model would be claimed.
From MPEP 2106.05(a)…
“If it is asserted that the invention improves upon conventional functioning of a computer, or upon conventional technology or technological processes, a technical explanation as to how to implement the invention should be present in the specification. That is, the disclosure must provide sufficient details such that one of ordinary skill in the art would recognize the claimed invention as providing an improvement. The specification need not explicitly set forth the improvement, but it must describe the invention such that the improvement would be apparent to one of ordinary skill in the art. Conversely, if the specification explicitly sets forth an improvement but in a conclusory manner (i.e., a bare assertion of an improvement without the detail necessary to be apparent to a person of ordinary skill in the art), the examiner should not determine the claim improves technology. An indication that the claimed invention provides an improvement can include a discussion in the specification that identifies a technical problem and explains the details of an unconventional technical solution expressed in the claim, or identifies technical improvements realized by the claim over the prior art. For example, in McRO, the court relied on the specification’s explanation of how the particular rules recited in the claim enabled the automation of specific animation tasks that previously could only be performed subjectively by humans, when determining that the claims were directed to improvements in computer animation instead of an abstract idea. McRO, 837F.3d at 1313-14, 120 USPQ2d at 1100-01. In contrast, the court in Affinity Labs of Tex. v. DirecTV, LLC relied on the specification’s failure to provide details regarding the manner in which the invention accomplished the alleged improvement when holding the claimed methods of delivering broadcast content to cellphones ineligible. 838 F.3d 1253, 1263-64, 120 USPQ2d 1201, 1207-08(Fed. Cir. 2016).”
“After the examiner has consulted the specification and determined that the disclosed invention improves technology, the claim must be evaluated to ensure the claim itself reflects the disclosed improvement in technology. Intellectual Ventures I LLC v. Symantec Corp., 838 F.3d 1307, 1316, 120 USPQ2d 1353, 1359 (Fed. Cir. 2016) (patent owner argued that the claimed email filtering system improved technology by shrinking the protection gap and mooting the volume problem, but the court disagreed because the claims themselves did not have any limitations that addressed these issues). That is, the claim must include the components or steps of the invention that provide the improvement described in the specification. However, the claim itself does not need to explicitly recite the improvement described in the specification (e.g., “thereby increasing the bandwidth of the channel”). The full scope of the claim under the BRI should be considered to determine if the claim reflects an improvement in technology (e.g., the improvement described in the specification). In making this determination, it is critical that examiners look at the claim “as a whole,” in other words, the claim should be evaluated “as an ordered combination, without ignoring the
requirements of the individual steps.” When performing this evaluation, examiners should be “careful to avoid oversimplifying the claims” by looking at them generally and failing to account for the specific requirements of the claims. McRO, 837 F.3d at 1313,120 USPQ2d at 1100.”
Applicant should cite the technical problem found in their disclosure (with current machine learning technology) and ensure the improvement is in the claims.
Turning to the second requirement, the applicant submits that it is also undeniable that the claims reflect the improvement described in the specification. Specifically, each of the independent claims includes features for training a machine learning model with a training dataset identifying numbers of objects and weights for slide images, and then implementing that model to predict numbers and weights of objects in other slide images.” As noted above, these are the exact features the application as filed describes as improving on conventional outline-based approaches to training machine learning models for slide image analysis, and so the second requirement from the patent office’s guidance for artificial intelligence inventions is also satisfied. Accordingly, the applicant submits that, even if the claims do recite an abstract idea (which, as noted above, the applicant does not concede to be the case), they should still be treated as patent eligible because they provide an improvement in machine learning model training technology.
Training is claimed at a high level of generality. See for example the July 2024, Subject Matter Eligibility”, Example 47, Claim 3 where training itself was either abstract as a mathematical concept or a mental process as just detecting anomalies, where no details were provided as to how the trained ANN operates or how detection is made, therefore mental process (pg. 10, paragraphs 4 and 5). In Claim 3, there were additional steps (d)-(f) that provided security solutions, improving network security (pg. 12 bottom paragraph).
It's also pointed out that “generating training set data comprising:…” is just generating a data set for training. This can be done mentally and/or with pen and paper. A person can come up with a set of data that will be used for training. The actual training is just “training a machine learning model based on the training set data;…” which is at a very high level of generality using a generic machine.
While the applicant believes that the arguments set forth above should be sufficient to establish the eligibility of the present claims, out of a desire to advance prosecution, the applicant would also like to address example 47 of the patent office’s eligibility examples. In the Interview, the Examiner indicated that that example may weight against the eligibility of the present claims. However, the applicant submits that, to the extent example 47 is relevant, the treatment of that example undermines, rather than supports, the present rejections under section 101. In example 47, a claim to an anomaly detection method which included training an artificial neural network was found to be ineligible. While that claim, like the present claims, included training a machine learning model, that is not the reason it was found to be ineligible. Rather, the significance of the training step in example 47 was that it was claimed in terms of specific mathematical calculations —i.e., back propagation and gradient descent.'? However, there are no analogous mathematical calculations in any of the pending independent claims, meaning that the analysis of training in example 47 cannot be applied to establish that the present independent claims are ineligible. Indeed, the applicant submits that, based on comparing example 47 with the treatment of training in example 39 it is clear that — contrary to the position advanced in the Office Action — training a machine learning model (in the absence of particular mathematical calculations) should not be treated as an ineligible abstract idea. Specifically, the claim of example 39, which was titled “Method for Training a Neural Network for Facial Detection” was found to be patent eligible because, despite including two explicit steps of “training the neural network,” it does not recite a judicial exception.'' Notably, the patent office’s analysis of that example specifically addressed whether the claim recited a mental process (i.e., the type of abstract idea the Office Action asserted was recited in the present claims) and stated that it did not “because the steps are not practically performed in the human mind.”” Accordingly, the applicant submits that patent office’s examples, to the extent relevant, support the eligibility of the present claims, and indicate that the rejections under 35 U.S.C. § 101 should be withdrawn.
The generate training steps were pointed out to be mental processes, also indicated in the SME where detecting was mental observation. A person mentally can identify the number of objects in a first slide and weigh data identifying the first weight. Applicant’s own disclosure teaches annotated by a user for ground truth purposes (para. [0162]).
Further, unlike Claim 3, Applicant is not improving network technology or other technology.
Based on the above response, the rejection is respectfully maintained.
Applicant argues 35 USC §103 Rejection, starting pg. 9 of Remarks (footnotes omitted):
The Rejections Under 35 U.S.C. § 103 Should be Withdrawn
In the Office Action, claims 1-20 were rejected under 35 U.S.C. § 103 as being obvious over U.S. published patent application 202 1/0209753 (“Dogdas”) in view of U.S. published patent application 2020/0250817 (“Leng”). In response, the applicant submits that each of the pending claims includes features which are not taught, suggested, or otherwise rendered obvious over the cited art. Accordingly, the applicant submits that the rejections under 35 U.S.C. § 103 should be reconsidered and withdrawn.
Remarks addressing the obviousness rejections on a claim by claim basis are set forth below.
Claim 1
Turning first to claim 1, the applicant submits that the rejection of that claim should be reconsidered and withdrawn because the cited art does not cover “implement[ing] the machine learning model, wherein the machine learning model predicts a number of a second plurality of objects in a second slide image and a second weight.” In its rejection, the Office Action asserted that that feature, which is recited in the final clause of claim 1, is covered by the teaching of paragraph 32 of Dogdas that “multiple (second) training images may be used to adjust weight, therefore, second slide image and second weight” when considered in combination with the disclosure from Dogdas’ paragraph 55 of “output quantification based on weights.”!? However, the applicant submits that those disclosures cannot be treated as covering the final clause of claim 1 because that clause recites that “the machine learning model predicts ... a second weight,” (emphasis added) rather than reciting that weights are adjusted or used to quantify outputs (e.g., a number of cancer cells, as described in paragraph 55 of Dogdas). Moreover, no reasoned explanation was provided for why it would have been obvious to modify Dogdas’ technology to provide a weight as a prediction as recited in claim 1, either based on Dogdas, Leng, the common sense of one skilled in the art, or a combination of the foregoing. Accordingly, the applicant submits that the rejection of claim 1 under 35 U.S.C. § 103 cannot properly be maintained and so respectfully requests that it be reconsidered and withdrawn.
From Applicant’s argument above…
>>”However, the applicant submits that those disclosures cannot be treated as covering the final clause of claim 1 because that clause recites that “the machine learning model predicts ... a second weight,” (emphasis added) rather than reciting that weights are adjusted or used to quantify outputs (e.g., a number of cancer cells, as described in paragraph 55 of Dogdas).”<<
The idea behind using machine learning and training would be to predict. Adjust a weight is a second weight as it cannot be the first weight.
Also from Dogdas…
“The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
“FIG. 5 shows an example implementation of the detection analysis as disclosed herein in reference to FIG. 2. FIG. 5 shows a flowchart for a process of predicting pCR and/or MRD in prostate cancer with treatment effects. In the example of prostate cancer, pCR and MRD may be used as endpoints for clinical trials that assess neoadjuvant treatments. However, a major challenge is the presence of treatment effects in assessing cancer.” [0072]
“The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
Claims 2-12
Turning next to claims 2-12, the applicant notes that each of those claims depends from claim 1, and so incorporates each of the innovative features identified as being absent from the cited art in the context of that claim. Additionally, as addressed in more detail below for claims 5, 10 and 11, claims 2-12 include their own features which can provide separate basis for distinguishing the cited art. Accordingly, the applicant submits that the rejections of claims 2-12 should be reconsidered and withdrawn based both on those claims’ relationship to claim 1, and on the additional features which those claims recite.
Claim 5
With respect to claim 5, that claim recites that a first slide image included in training data used to train the machine learning model corresponds to a portion of an image and is identified in user input from a user computing device. In its rejections, the Office Action asserted that claim 5 was covered by the disclosure of selected areas of interest from paragraph 29 of Dogdas.'* However, that paragraph simply describes the types of images in which cancer cells may be detected. It does not describe machine learning training, and so cannot be treated as teaching, suggesting, or otherwise rendering obvious the inclusion of user identified portions of images in training data for a machine learning model. Accordingly, the applicant submits that the rejection of claim 5 should be reconsidered and withdrawn even if the rejection of claim 1 is maintained.
Claim 5 is equal to claim 1 and claim 5. Claim 1 provides teaching on training and machine learning. Claim 5 simply adds obtain user input identifying the portion of the image.
Claim 10
With respect to claim 10, that claim recites that the training data set for the machine learning model of claim 1 comprises both a first slide image which corresponds to a first portion of an image and a third slide image which corresponds to a second portion of the image. In rejecting claim 10, the Office action asserted that the use of the third image in addition to the first image was taught by the disclosure of multiple images in paragraph 6 of Dogdas, and that the first and third images being slide images was taught by the disclosure of a slide intake tool and slide scanner from paragraph 42 of that same reference.'° However, claim 10 does not simply recite multiple images which are slide images, and instead requires that slide images which each corresponding to different portions of an image be included in a machine learning model’s training data, something which is not included in, and which no reason has been given to add to, the teachings of Dogdas. Accordingly, the applicant submits that the rejection of claim 10 should be reconsidered and withdrawn, even if the rejection of claim 1 is maintained.
From Applicant’s argument above….
>>”However, claim 10 does not simply recite multiple images which are slide images, and instead requires that slide images which each corresponding to different portions of an image be included in a machine learning model’s training data, something which is not included in, and which no reason has been given to add to, the teachings of Dogdas. Accordingly, the applicant submits that the rejection of claim 10 should be reconsidered and withdrawn, even if the rejection of claim 1 is maintained.”<<
Dogdas in para. [0055] continues to teach identify regions (plural) and model based on training with images that provide the same…
“The detection machine learning model may be trained to output cancer qualifications and quantifications, as disclosed herein. Cancer qualifications may be output for one or more of a plurality of different cancer types. The detection machine learning model may be trained using images from the one or more of the plurality of different cancer types. For example, the training images may include images related to breast cancer, prostate cancer, and lung cancer. Accordingly, the generated detection machine learning model may receive a digital image at 202 of FIG. 2 and may qualify the image as representing tissue that includes cancer cells, the number of cancer cells, and/or the type of cancer cells. The detection machine learning model may output the cancer qualification and/or quantification based on weights and/or layers trained during its training process. Based on the weights and/or layers, the detection machine learning model may identify regions of a digital image that may more strongly be used as evidence for the presence or absence of cancer and, further, the extent of the cancer. The model may then evaluate some or all of those regions to determine the presence, absence, and/or extent of cancer cells based on training with images that provide the same. Feedback (e.g., pathologist confirmation, correction, adjustment, etc.) may further train the detection machine learning model during operation of the model.” [0055]
Also from paragraph [0032] and training to identify regions…
“…The detection machine learning model may be trained to detect cancer cells based on, for example, training images having tagged cancer cells. The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.
Claim 11
With respect to claim 11, that claim recites implementing a second machine learning model which aggregates a plurality of predictions from the machine learning model of claim 1. In rejecting claim 11, the Office Action asserted that this was covered by paragraph 80 of Dogdas because that paragraph disclosed the number 37, which was interpreted as a number in an image 4 Office Action at 16. '5 Office Action at 41-42. However, the applicant asserts that this application of paragraph 80 to claim 11 is improper. First, the number 37 in paragraph 80 was not a number in an image. Rather, it was a number of malignant WSIs (i.e., whole slide images) which were correctly identified by the machine learning model described in Dogdas, which was presented in support of the statement the potential benefits of that model.'” Second, there is nothing in Dogdas to indicate that the number 37 in paragraph 80 of Dogdas was aggregated from a plurality of machine learning model predictions by a second machine learning model. Instead, given that the number 37 was derived by adding the number of WSIs which contained tumors less than 5 mm (three) and the number of WSIs which contained tumors greater than 5 mm (34), it seems much more likely that the number 37 was derived by a human simply adding two numbers together, without any involvement of a second machine learning model at all. Accordingly, the applicant submits that paragraph 80 of Dogdas cannot properly be treated as covering claim 11, and so the rejection of that claim should be reconsidered and withdrawn, even if the arguments provided for claim 1 are not found to be persuasive.
From Applicant’s argument above…
>>”With respect to claim 11, that claim recites implementing a second machine learning model which aggregates a plurality of predictions from the machine learning model of claim 1.”<<
The clam recites “aggregates a plurality of predictions…” which is not limited by the ML model in claim 1, and there is no antecedence for the predictions.
From Applicant’s argument above…
>>”In rejecting claim 11, the Office Action asserted that this was covered by paragraph 80 of Dogdas because that paragraph disclosed the number 37, which was interpreted as a number in an image 4 Office Action at 16. '5 Office Action at 41-42. However, the applicant asserts that this application of paragraph 80 to claim 11 is improper. First, the number 37 in paragraph 80 was not a number in an image. Rather, it was a number of malignant WSIs (i.e., whole slide images) which were correctly identified by the machine learning model described in Dogdas, which was presented in support of the statement the potential benefits of that model.'””<<
It's the number in the image, for the whole slide.
>>”Second, there is nothing in Dogdas to indicate that the number 37 in paragraph 80 of Dogdas was aggregated from a plurality of machine learning model predictions by a second machine learning model. Instead, given that the number 37 was derived by adding the number of WSIs which contained tumors less than 5 mm (three) and the number of WSIs which contained tumors greater than 5 mm (34), it seems much more likely that the number 37 was derived by a human simply adding two numbers together, without any involvement of a second machine learning model at all.”<<
Use of “WSIs” is multiple slides, where presumably the number 37 was totaled from the different images provided by the slides. Further, in the sentence before it teaches “Ground truth was established by pathologist annotations” therefore this was provided by ML and proved by a pathologist.
Claims 13 and 14
Turning next to claims 13 and 14, the applicant notes that those claims include features similar to those discussed above in the context of claim 1, and that claims 13 and 14 were treated as interchangeable with claim 1 in the rejections under 35 U.S.C. § 103.!8 Accordingly, the applicant submits that the arguments set forth above for claim 1 can also be applied to claims 13 and 14, and so requests that the rejections of claims 13 and 14 be reconsidered and withdrawn based on the similarity between those claims and claim 1.
The rejection is respectfully maintained based on reasons provided above for Claim 1.
Claims 15-19
Turning next to claims 15-19, the applicant notes that each of those claims depends from claim 14, and so incorporates each of the innovative features identified as being absent from the cited art in the context of that claim. Additionally, claims 15-19 include their own features which can provide separate basis for distinguishing the cited art (e.g., amended claim 18 recites features similar to those discussed above as distinguishing claim 10). Accordingly, the applicant submits that the rejections of claims 15-19 should be reconsidered and withdrawn based both on those claims’ relationship to claim 14, and on the additional features which those claims recite.
The rejection is respectfully maintained based on reasons provided above for Claim 1.
Applicant argues new claims, pg. 16 of Remarks:
The Office Action’s Rejections Cannot be Applied to the New Claims
Turning next to new claims 23-25, the applicant submits that the Office Action’s rejections cannot be applied to those claims at least because those claims depend from claim 13 and therefore incorporate the features from claim 13 discussed above as overcoming the Office Action’s rejections. Additionally, claims 23-25 include their own features which can establish the patentability of those claims even if the rejection of claim 13 is maintained. For example, claim 23 (and therefor claim 24, which depends from claim 23) recites that the first slide image is a portion of an image identified by a user, similar to claim 5 discussed above. Likewise, claim 25 recites implementing a second machine learning model which aggregates a plurality of predictions, similar to claim 11, discussed above. Accordingly, the applicant submits that the Office Action’s rejections cannot be applied to the new claims, based both on those claims depending from claim 13, and on the additional features they recite.
With all due respect, new claims are subject to 35 USC 102/103 requirements. There is nothing in the MPEP that would allow for not examining claims and applying relevant prior art.
The rejection is respectfully maintained but modified for the claim amendments.
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, 5-8, 10-19, and 23-25 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Claims 1-3, 5-8, 10-19, and 23-25 are directed to a system, method, or product, which are statutory categories of invention. (Step 1: YES).
The Examiner has identified method Claim 13 as the claim that represents the claimed invention for analysis and is similar to system Claim 1 and product Claim 14. Claim 13 recites the limitations of:
A computer-implemented method comprising:
obtaining a first slide image comprising a first plurality of objects;
determining a number of the first plurality of objects in the first slide image and a first weight;
generating training set data comprising:
the first slide image,
object data identifying the number of the first plurality of objects in the first slide image, and
weight data identifying the first weight;
training a machine learning model based on the training set data; and
implementing the machine learning model, wherein the machine learning model predicts a number of a second plurality of objects in a second slide image and a second weight.
These above limitations, under their broadest reasonable interpretation, cover performance of the limitation as mental processes. The claim recites elements, in non-bold above, covers performance of the limitation that can be concepts performed in the mind of a person or with pen and paper. For example, a person can obtain a slide image, determine objects on the slide, generate a training set of data comprising a slide image, data identifying a number of objects in the slide, weigh data identifying a first weight, and implement a model. See also MPEP 2106.04(a)(2) III C, where using a generic computer for an abstract idea was shown to be non-statutory. If a claim limitation, under its broadest reasonable interpretation, covers performance of the limitation as a mental process, then it falls within the “Mental Processes” grouping of abstract ideas. Accordingly, the claim recites an abstract idea. Claims 1 and 14 are also abstract for similar reasons. (Step 2A-Prong 1: YES. The claims are abstract)
This judicial exception is not integrated into a practical application. In particular, the claims only recite: memory circuit, hardware processing unit, machine (claim 1); computer, machine (Claim 13); computer-readable medium, computing devices, machine (Claim 14). The computer hardware is recited at a high-level of generality (i.e., as a generic processor performing a generic computer function) such that it amounts no more than mere instructions to apply the exception using a generic computer component. The machine is a generic machine. The training and machine learning model are being applied at a high level of generality. Accordingly, these additional elements, when considered separately and as an ordered combination, do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Therefore claims 1, 13, and 14 are directed to an abstract idea without a practical application. (Step 2A-Prong 2: NO. The additional claimed elements are not integrated into a practical application)
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, when considered separately and as an ordered combination, they do not add significantly more (also known as an “inventive concept”) to the exception. As discussed above with respect to integration of the abstract idea into a practical application, the additional element of using a computer hardware amounts to no more than mere instructions to apply the exception using a generic computer component. Mere instructions to apply an exception using a generic computer component cannot provide an inventive concept. Accordingly, these additional elements, when considered separately and as an ordered combination, do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Steps such as obtaining (receiving) are steps that are considered insignificant extra solution activity and mere instructions to apply the exception using general computer components (see MPEP 2106.05(d), II). Thus claims 1, 13, and 14 are not patent eligible. (Step 2B: NO. The claims do not provide significantly more)
Dependent claims 2, 3, 5-8, 10-12, 15-20, and 23-25 further define the abstract idea that is present in their respective independent claims 1, 13, and 14 and thus correspond to Mental Processes and hence are abstract for the reasons presented above. The dependent claims do not include any additional elements that integrate the abstract idea into a practical application or are sufficient to amount to significantly more than the judicial exception when considered both individually and as an ordered combination. Claim 4 recites machine learning model and convolutional neural network at a high level of generality. Claims 11 and 20 recite machine learning and training applied at a high level of generality. Therefore, the claims 2, 3, 5-8, 10-12, 15-20, and 23-25 are directed to an abstract idea. Thus, the claims 1-3, 5-8, 10-19, and 23-25 are not patent-eligible.
Examiner Request
The Applicant is requested to indicate where in the specification there is support for amendments to claims should Applicant amend. The purpose of this is to reduce potential 35 U.S.C. §112(a) or §112 1st paragraph issues that can arise when claims are amended without support in the specification. The Examiner thanks the Applicant in advance.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
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 1-3, 5-8, 10-19, and 25 are rejected under 35 U.S.C. 103 as being unpatentable over Pub. No. US 2021/0209753 to Dogdas et al.in view of Pub. No. US 2020/0250817 to Leng et al.
Regarding claims 1, 13, and 14
(claim 1) An apparatus comprising:
a memory circuit storing computer-executable instructions; and
Dogdas teaches:
Memory storing instructions…
“A system for outputting cancer detection results includes a memory storing instructions and a processor executing the instructions to perform a process including receiving a digital image corresponding to a target specimen associated with a pathology category, wherein the digital image is an image of tissue specimen, determining a detection machine learning model, the detection machine learning model being generated by processing a plurality of training images to output a cancer qualification and further a cancer quantification if the cancer qualification is a confirmed cancer qualification, providing the digital image as an input to the detection machine learning model, receiving one of a pathological complete response (pCR) cancer qualification or a confirmed cancer quantification as an output from the detection machine learning model, and outputting the pCR cancer qualification or the confirmed cancer quantification.” [0008]
a hardware processing unit configured to execute the computer-executable instructions, wherein execution of the computer-executable instructions causes the hardware processing unit to:
Processor executing instructions…
“A system for outputting cancer detection results includes a memory storing instructions and a processor executing the instructions to perform a process including receiving a digital image corresponding to a target specimen associated with a pathology category, wherein the digital image is an image of tissue specimen, determining a detection machine learning model, the detection machine learning model being generated by processing a plurality of training images to output a cancer qualification and further a cancer quantification if the cancer qualification is a confirmed cancer qualification, providing the digital image as an input to the detection machine learning model, receiving one of a pathological complete response (pCR) cancer qualification or a confirmed cancer quantification as an output from the detection machine learning model, and outputting the pCR cancer qualification or the confirmed cancer quantification.” [0008]
obtain a first slide image comprising a first plurality of objects;
Slide intake (obtain) tool and slide scanner (slide image)…
“Specifically, FIG. 1B depicts components of the machine learning module 100, according to one embodiment. For example, the machine learning module 100 may include a detection tool 101, a data ingestion tool 102, a slide intake tool 103, a slide scanner 104, a slide manager 105, a storage 106, and a viewing application tool 108. For clarification, the machine learning module 100 shown in FIGS. 1A and 1B is a previously trained and generated machine learning model (e.g., a detection machine learning model that may include a treatment effects machine learning model). Additional disclosure is provided herein for training and generating different types of machine learning models that may be used as machine learning module 100.” [0042]
Cancer cells (plurality of objects)…
“The detection tool 101 refers to a process and system for determining a cancer qualification, and if a confirmed cancer qualification is present, determining a cancer quantification. The cancer qualification may be a confirmed cancer qualification, a pCR cancer qualification (e.g., no cancer cells detected), or the like. A confirmed cancer qualification may indicate that one or more cancer cells were detected in the digital image of a tissue specimen. A cancer quantification may indicate the number of cancer cells detected, a ratio of cancer cells to non-cancer cells, or a degree of cancer. A subset of the cancer quantification is a MRD cancer qualification which may indicate whether the number of cancer cells are below a MRD threshold. The MRD threshold may be protocol specific, cancer type specific, institution specific, pathologist specific, or the like. The detection tool 101 may include a plurality of machine learning models or may load one machine learning model at a time. For example, the detection tool 101 may include a treatment effects machine learning model that may be trained based on a different or additional training data set then the detection machine learning model disclosed herein.” [0043]
determine a number of the first plurality of objects in the first slide image and a first weight;
Example of make (determine) cancer quantifications (number of objects)…
“Pathologists may evaluate cancer and other disease pathology slides in for cancer detection. The present disclosure presents an automated way to identify cancer cells and to make cancer qualifications and, if applicable, cancer quantifications. In particular, the present disclosure describes various exemplary AI tools that may be integrated into the workflow to expedite and improve a pathologist's work.” [0024]
Example of weights based on identify regions to have cancer cells…
“The detection machine learning model may be trained to detect cancer cells based on, for example, training images having tagged cancer cells. The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
See Weight below.
generate training set data comprising:
Example of train using cancer cells and weights (training set data)…
“The detection machine learning model may be trained based on supervised, semi-supervised, weakly-supervised or un-supervised training including but not limited to multiple instance learning. Training images may be from the same pathology category as the respective digital images input to the detection machine learning model. According to an implementation, multiple different training images from a plurality of pathology categories may be used to train the detection machine learning model across pathology categories. According to this implementation, an input to the detection machine learning model may include the pathology category of the digital image. Pathology categories may include, but are not limited to, histology, cytology, frozen section, immunohistochemistry (IHC), immunofluorescence (IF), hematoxylin and eosin (H&E), hematoxylin alone, molecular pathology, 3D imaging, or the like. The detection machine learning model may be trained to detect cancer cells based on, for example, training images having tagged cancer cells. The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
the first slide image,
Image from a slide…
“Histopathology refers to the study of a specimen that has been placed onto a slide. For example, a digital pathology image may be comprised of a digitized image of a microscope slide containing the specimen (e.g., a smear). One method a pathologist may use to analyze an image on a slide is to identify nuclei and classify whether a nucleus is normal (e.g., benign) or abnormal (e.g., malignant)…” [0028]
object data identifying the number of the first plurality of objects in the first slide image, and
Example of quantification (identifying) number of cancer cells in the digital image…
“Implementations of the disclosed subject matter include systems and methods for using a detection machine learning model to determine the presence or absence of cancer cells in a WSI. The detection machine learning model may be generated to determine a cancer qualification. The cancer qualification may include an indication of whether cells represented in a digital image of a tissue sample are cancer cells or if no cancer cells are identified in the digital image. According to an implementation, the cancer qualification may also include a type of cancer (e.g., breast, prostate, bladder, colorectal, etc.). If a cancer qualification is a confirmed cancer qualification, then a cancer quantification may also be output by the detection machine learning model. The cancer quantification may indicate the number, ratio, or degree of cancer cells identified from the digital image and may be a minimal residual disease (MRD) designation based on an established MRD criteria (e.g., 1 cell per million or less). If the cancer qualification output by the detection machine learning model indicates no cancer cells, a pathological complete response (pCR) cancer qualification may be output.” [0031]
weight data identifying the first weight;
Example of weights based on identify regions to have cancer cells…
“The detection machine learning model may be trained to detect cancer cells based on, for example, training images having tagged cancer cells. The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
See Weight below.
train a machine learning model based on the training set data; and
Machine learning model trained…
“The detection machine learning model may be trained based on supervised, semi-supervised, weakly-supervised or un-supervised training including but not limited to multiple instance learning. Training images may be from the same pathology category as the respective digital images input to the detection machine learning model. According to an implementation, multiple different training images from a plurality of pathology categories may be used to train the detection machine learning model across pathology categories. According to this implementation, an input to the detection machine learning model may include the pathology category of the digital image. Pathology categories may include, but are not limited to, histology, cytology, frozen section, immunohistochemistry (IHC), immunofluorescence (IF), hematoxylin and eosin (H&E), hematoxylin alone, molecular pathology, 3D imaging, or the like. The detection machine learning model may be trained to detect cancer cells based on, for example, training images having tagged cancer cells. The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
implement the machine learning model, wherein the machine learning model predicts a number of a second plurality of objects in a second slide image and a second weight.
Multiple (second) training images may be used and adjust weight, therefore, second slide and second weight…
“The detection machine learning model may be trained based on supervised, semi-supervised, weakly-supervised or un-supervised training including but not limited to multiple instance learning. Training images may be from the same pathology category as the respective digital images input to the detection machine learning model. According to an implementation, multiple different training images from a plurality of pathology categories may be used to train the detection machine learning model across pathology categories. According to this implementation, an input to the detection machine learning model may include the pathology category of the digital image. Pathology categories may include, but are not limited to, histology, cytology, frozen section, immunohistochemistry (IHC), immunofluorescence (IF), hematoxylin and eosin (H&E), hematoxylin alone, molecular pathology, 3D imaging, or the like. The detection machine learning model may be trained to detect cancer cells based on, for example, training images having tagged cancer cells. The detection machine learning model may adjust weights in one or more layers to identify regions likely to have cancer cells based on a known or determined cancer type and may further adjust weights in one or more layers based on identifying cancer cells or not finding cancer cells within those regions.” [0032]
Output quantification based on weights…
“The detection machine learning model may be trained to output cancer qualifications and quantifications, as disclosed herein. Cancer qualifications may be output for one or more of a plurality of different cancer types. The detection machine learning model may be trained using images from the one or more of the plurality of different cancer types. For example, the training images may include images related to breast cancer, prostate cancer, and lung cancer. Accordingly, the generated detection machine learning model may receive a digital image at 202 of FIG. 2 and may qualify the image as representing tissue that includes cancer cells, the number of cancer cells, and/or the type of cancer cells. The detection machine learning model may output the cancer qualification and/or quantification based on weights and/or layers trained during its training process. Based on the weights and/or layers, the detection machine learning model may identify regions of a digital image that may more strongly be used as evidence for the presence or absence of cancer and, further, the extent of the cancer. The model may then evaluate some or all of those regions to determine the presence, absence, and/or extent of cancer cells based on training with images that provide the same. Feedback (e.g., pathologist confirmation, correction,