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 .
Reopening of Prosecution After Appeal Brief
In view of the appeal brief filed on 3/19/2026, PROSECUTION IS HEREBY REOPENED. New grounds of rejection are set forth below.
The claim amendments submitted 2/17/2026 are entered.
Terminal Disclaimer
The terminal disclaimer filed on 2/17/2026 disclaiming the terminal portion of any patent granted on this application which would extend beyond the expiration date of US11983868B2 has been reviewed and is accepted. The terminal disclaimer has been recorded.
Priority
Upon further consideration, this priority is given herein and entered into record:
Applicant’s claim for the benefit of a prior-filed application under 35 U.S.C. 119(e) or under 35 U.S.C. 120, 121, or 365(c) is acknowledged. Priority of US application 62/633,311 filed 02/21/2018 is acknowledged.
Withdrawn Rejections/Objections
The double patenting rejections to claims 1, 5, 10, 15, 17 and 20 in the Office action mailed 01 October 2025 are withdrawn in view of the terminal disclaimer filed on 2/17/2026.
The rejection to claims 1-20 under 35 USC § 103 in the Office action mailed 01 October 2025 are withdrawn in view of the terminal disclaimer filed on 2/17/2026. However, new 103 rejections are applied.
Claim Status
Claims 1-20 are pending and are examined on the merits
Claim Rejections - 35 USC § 101
The instant rejection is maintained from the Office Action of 10/1/2025 and modified to address amendments filed 02/17/2026.
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 USC 101 because the claimed invention is directed to an non-subject matter.
Step 1: Process, Machine, Manufacture, or Composition of Matter
Claims 1-9 are drawn to a process, here a “method”, with functional steps like “providing”, “extracting”, “controlling”, and “generating”.
Claim 10-16 are drawn to a machine or manufacturer, here a "system," with structural components like "one or more computing devices”.
Claim 17-20 are drawn to another machine or manufacturer, here a "non-transitory computer-readable storage”.
Step 2A Prong One: Identification of Judicial Exceptions
The claims recite:
Extracting a set of radiomic features from the pre-treatment image.
This element recites “extracting” a set of radiomic features, but provide nothing more than mere instructions to implement an abstract idea on the pre-treatment image.
The “extracting” is used to generally apply the abstract idea without limiting how the “extracting” functions. These limitations only recite the outcomes of “extracting” without any details about how the outcomes (here “a set of radiomic features”) are accomplished.
Under a BRI, “extracting a set of radiomic features” reads on data collection (or data outputting of “a set of radiomic features”). However, since it recited at such a high-level of generality (no feature is specified; no details on how the pre-treatment radiology image is acquired; no specification on how “extracting” is performed), it is directed to an insignificant extra-solution activity (MPEP §2106.05(g)) because extracting features (aka “feature engineering”) is a routine and necessary step before machine learning model training.
Providing the set of radiomic features to a machine learning model, the machine learning model having been trained to generate a second prediction as to whether the region of tissue will respond to the pharmaceutic treatment based on the set of radiomic features.
Under a broadest reasonable interpretation (BRI), “machine learning model’ is interpreted as a mathematical model, e.g. linear regression model, and generating a second prediction from radiomic features equates to a mathematical operation. Therefore, this step is classified into an abstract idea of mathematical concepts.Controlling the deep learning model to generate a first prediction.
This limitation does not specify the structure for the deep learning model, nor does it clear how the deep learning model is controlled. The deep learning model is recited at a high level of generality and amounts to no more than mere instructions to apply the exception using the deep learning model.
Controlling the machine learning model to generate the second prediction
This limitation does not specify the structure for the machine learning model, nor does it clear how the machine learning model is controlled. The machine learning model is recited at a high level of generality and amounts to no more than mere instructions to apply the exception using the machine learning model.
Generating a classification of the region of tissue as a responder or non-responder to the pharmaceutical treatment prior to administering the pharmaceutic treatment based, at least in part, on the first prediction and the second prediction.
Under a BRI, “generating a classification of the region of tissue as a responder or non-responder”, “based, at least in part, on the first prediction and the second prediction” is interpreted as one probability times another probability, which can be achieved in mind with the help of a pen/paper. Therefore this step equates to an abstract idea of mental processes.
Dependent claims 2-9, 11-16 and 18-20 further recite limitations characterizing the data input, output and analysis processes, that therefore read on a judicial exception.
Step 2A, Prong Two: Consideration of a Practical Application
The claims result in a step outputting the classification of the region of tissue as a responder or non-responder (claims 1, 10 and 17), and a personalized treatment plan (claim 8). The claims do not include any additional elements that integrate the recited judicial exceptions (abstract ideas and natural correlation) into a practical application.
This judicial exception is not integrated into a practical application because the claims do not meet any of the following criteria:
An additional element reflects an improvement in the functioning of a computer, or an improvement to other technology or technical field;
an additional element that applies or uses a judicial exception to effect a particular treatment or prophylaxis for a disease or medical condition;
an additional element implements a judicial exception with, or uses a judicial exception in conjunction with, a particular machine or manufacture that is integral to the claim;
an additional element effects a transformation or reduction of a particular article to a different state or thing; and
an additional element applies or uses the judicial exception in some other meaningful way beyond generally linking the use of the judicial exception to a particular technological environment, such that the claim as a whole is more than
a drafting effort designed to monopolize the exception.
Step 2B: Consideration of Additional Elements and Significantly More
The preamble of claim 17 recites one or more computer devices and non-transitory computer-readable storage devices. These are generic computing components and they merely provide technical environment to execute the abstract idea (MPEP § 2106.05(h)).
The claimed method also recites "additional elements" that are not limitations drawn to an abstract idea:
Providing a pre-treatment radiology image of a region of tissue of a patient to a deep learning model, the pre-treatment radiology image including at least one tumor, the deep learning model having been trained to generate a first prediction as to whether the region of tissue will respond to a pharmaceutical treatment based on the pre-treatment radiology image or portions thereof;
This step equates to an additional element of inputting radiology image to a deep learning model and get a data output.
However, using radiology images as data input, and using deep learning as tools, to classify/stratify tumors/cancers are conventional in laboratory and/or clinical environment, as exemplified by:
Joseph B. DeGrandchamp, et al.: "Predicting response before initiation of neoadjuvant chemotherapy in breast cancer using new methods for the analysis of dynamic contrast enhanced MRI (DCE MRI) data," Proc. SPIE 9788, Medical Imaging 2016: Biomedical Applications in Molecular, Structural, and Functional Imaging, 978811, 29 March 2016. Cited on 12/19/2023 IDS.
Natasha Antropova, et al.: "Performance comparison of deep learning and segmentation-based radiomic methods in the task of distinguishing benign and malignant breast lesions on DCE-MRI," Proc. SPIE 10134, Medical Imaging 2017: Computer-Aided Diagnosis, 101341G, 3 March 2017. Cited on the 12/19/2023 IDS.
Stefano Marrone, et al.: “An Investigation of Deep Learning for Lesions Malignancy Classification in Breast DCE-MRI”, Image Analysis and Processing – ICIAP 2017, 19th International Conference, Catania, Italy, September 11–15, 2017. Cited on the 12/19/2023 IDS.
Prasanna, P., Tiwari, P. & Madabhushi, A. “Co-occurrence of Local Anisotropic Gradient Orientations (CoLlAGe): A new radiomics descriptor”. Sci Rep 6, 37241 (2016) . Cited on the 12/19/2023 IDS.
Mitrea, S. Nedevschi, M. Abrudean and R. Badea, "Colorectal cancer recognition from ultrasound images, using complex textural microstructure cooccurrence matrices, based on Laws' features," 2015 38th International Conference on Telecommunications and Signal Processing (TSP), 2015, pp. 458-462. Cited on the 12/19/2023 IDS.
Fan, Ming, et al. "Radiomic analysis of DCE-MRI for prediction of response to neoadjuvant chemotherapy in breast cancer patients." European journal of radiology 94 (2017): 140-147. Cited on the 12/19/2023 IDS.
Reza Rasti, Mohammad Teshnehlab, Son Lam Phung: “Breast cancer diagnosis in DCE-MRI using mixture ensemble of convolutional neural networks”, Pattern Recognition 72 (2017) 381–390. Cited on the 12/19/2023 IDS.
Braman “Intratumoral and peritumoral radiomics for the pretreatment prediction of pathological complete response to neoadjuvant chemotherapy based on breast DCE-MRI”, Breast Cancer Research (2017) 19:57. Cited on the 12/19/2023 IDS.
A. DeGrandchamp et al. evidences predicting response to NAC on breast cancer using the DCE-MRI image data and machine learning.
B. Natasha Antropova, et al. evidences classifying breast cancers using DCE-MRI image based on machine learning of CNNs. In this reference radiomic features are extracted from radiology images and are used as input to the deep learning model.
C. Stefano Marrone, et al. evidences classifying breast cancers using DCE-MRI image based SVM (para 3, pg 484) using the features extracted by pre-trained CNN, (Figure 1, pg. 483).
D and E. Prasanna, P., Tiwari, P. & Madabhushi, A. and Mitrea, S. Nedevschi, M. Abrudean and R. Badea, evidence image processing. Combined D and E evidence the eight radiomic parameters that have been shown to contribute most to the classification power in the instant invention.
F. Fan, Ming, et al. evidence extracting radiomic features from DCE-MRI images of breast cancer patients to predict response to neoadjuvant chemotherapy using a classifier.
G. Reza Rasti, Mohammad Teshnehlab, Son Lam Phung reviewed the computer-aided applications in breast cancer diagnosis based on DCE-MRI images (Table 1, pg 382). G also reviewed the CNN methods in medical applications (Table 2, pg 383).
H. Braman evidences DCE-MRI image for breast cancer prognosis through DLDA, LDA and other machine learning algorithms. H also teaches extracting radiomic features out or the radiology images
These references demonstrated that recited additional element in current claims, such as radiology images, and Deep learning, have been applied in breast cancer diagnosis and prognosis for years and by multiple groups. Therefore, the additional element recited in instant claims are well-known and conventional at the time the current invention was filed.
The claim(s) does/do not include additional elements that are sufficient to amount to significantly more than the judicial exception. Viewed as a whole, these additional claim element(s) do not provide meaningful limitation(s) to transform the abstract idea recited in the instantly presented claims into a patent eligible application of the abstract idea such that the claim(s) amounts to significantly more than the abstract idea itself. Therefore, the claim(s) are rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter.
Response to Applicant’s Arguments
The Remarks filed 2/17/2026 is responded on the Advisory Action (PTOL-303) filed 3/4/2026.
In the Appeal Brief filed 23 Mar 2026 (page 9 through page 16), Applicant argued (page 9, last para through page 10, 1st para) that “these claims, considered as a whole and in the context of the Specification, are more correctly viewed as reflecting improvements to machine learning that allow a computer, equipped with appropriately-trained machine learning models, to predict which patients will respond to a pharmaceutical treatment, such as neo-adjuvant chemotherapy (NAC), prior to treatment”. Applicant’s argument relates to Step 2A/Prong two, relating to whether claims recite a technical improvement to machine learning.
In response, Applicant’s argument is not persuasive. First Examiners don’t read specification into claims during examination; Second we don’t know what is improved. In the instant independent claims both the deep learning and the machine learning models are recited at a high level of generality. Neither the model structures, nor the model inputs are specified; Third, claims don’t recite any limitations to apply and to reflect the improvement.
In the Appeal Brief, Applicant argued (page 10, 2nd para) that claims here provide something that cannot be done by conventional means. Applicant’s argument refers to Step 2B in the 101 analysis, relating to whether claims recite non-conventional combinations.
In response, Applicant’s argument is not persuasive. As discussed in the 101 analysis, elements like “Generating a classification of the region of tissue as a responder or non-responder”, or estimating prognostic outcomes, are common and known in the art. Applicant simply combined two machine learning models with an expectation for success.
Furthermore, Applicant claims that there is an improvement, a correlation between types of tumors captures by radiological images and response to NAC treatment outcome. Applicant is citing a natural correlation. It is not an improvement to machine learning technology.
Nothing in the instant claims is specific about oncology. The claims recite generically any tumor radiology images, generically a deep learning model, and generically a machine learning model. The deep learning model and the machine learning model are used to generally apply the abstract idea without limiting how the trained deep learning model functions and how the machine learning model functions. The limitations only recite the outcomes of “to generate the first prediction” and “to generate the second prediction” and without any details about how the outcomes are accomplished.
In the Appeal Brief, Applicant argued (page 10, 3rd para through page 11, last para) that Examiner failed to consider recent clarifications in Office policy and the case law underlying them:
Ex Parte Desjardins et al., Appeal No. 2024-000567;
Enfish, LLC V. Microsoft Corp., 822 F.3d 1327 (Fed. Cir. 2016),
Memorandum from Charles Kim, Deputy Commissioner for Patents, to technology Centers 2100, 2600, and 3600 (August 14, 2025), related to 35 USC §101.
In response, Applicant’s argument is not persuasive. Instant claim 1 is not similar to Desjardins, or Enfish. The instant invention is about a medical technology, not a computing technology. As outlined in the 101 rejection, instant claim 1 does recite mathematical concept when extracting radiomic features and when combining the first prediction and the second prediction to acquire a outputting classification.
Applicants review (page 10-11) the guidance in Desjardins and argue that the Examiner is analyzing the claims at a high level of generality, dissecting the claims and ignoring clear evidence of an improvement to computer technology. Applicants argue that the improvements are analogous to those in Enfish
Applicant argues evidence in Spec. at [0003, 0005] to support the improvement.
Applicant argues (page 10, last two paras) “the Examiner fails to consider recent clarifications in Office policy and the case law underlying them”,
More specifically, Applicant argues that Examiner “analyzing the claims at too high a level of generality, dissecting the claims, and ignoring clear evidence of improvement to computer technology that is set forth in the Specification” (page 11, 2nd para).
In response, Applicant’s argument is not persuasive. Examiner has to weight what is recited in claims, Improvement has to be recited in the claims.
As discussed above, claims do not recite any limitations that reflect an improvement the machine learning technology or to the deep learning technology. Examiner is analyzing claims using BRI. Because claims don’t recite any specific radiomic features used as input for the machine learning models, nothing is specified about the structures of the deep learning model and the machine learning model, the claims don’t recite specifically what is predicted. A case of technological improvement to the machine learning or the deep learning is not built here. As the to the alleged “improvement to computer technology”, computer is not recited in instant claims 1 and 10.
In the Appeal Brief, Applicant argues (page 12, last two paras through page 13, penultimate para) that Examiner misused the BRI standard, regarding whether BRI should be applied “consistent with” or “in light of” the specification.
In response, it is improper to import claim limitations from the specification (MPEP §2111.01.II). Examination is not done using the method in Phillips (Phillips V. AWH Corp., 415 F.3d 1303, 1316 (Fed. Cir. 2005). Examination does not read the specification into the claims (MPEP §2111.01).
Applicants argue (Remarks, page 13, par. 1) that the tasks recite in the claims cannot be performed by the human mind
Applicant argues (page 13, par. 2) that Examiner admits that the claims reflect an improvement to technology and argue that consideration to whether the claims improves the functioning of a computer should be given. In response, as quoted by Applicant, it was noted in the Advisory Action that there might be an improvement in image classification. However, mere image classification under BRI is an abstract idea. The “controlling of a deep learning model,” “controlling of the machine learning model” and “generating a classification,” are recited at such a high degree of generality that any improvement is not reflected by the claims. The step of “controlling the deep learning model to generate the first prediction,” is analogous to Example 47, claim 2. Instantly, generating the first prediction is an abstract idea and the deep learning model is used to generally apply the abstract idea without limiting how the deep learning model functions. The deep learning model is described at a high level such that it amounts to using a computer with a generic deep learning model to apply the abstract idea. Therefore, one cannot conclude that the claims represent any improvement to a computer or computer product but instead merely implement a generically recited deep learning model as a software-type calculator for processing efficiency.
In the Appeal Brief, Applicant further argues (page 13, penultimate para through page 14, 2nd para) that Examiner “admits that the claims reflect an improvement to technology a traditional indicator that the claims are patent-eligible. ‘In determining patent eligibility, examiners should consider whether the claim purport(s) to improve the functioning of the computer itself or any other technology or technical field.’ MPEP §2106.05”).
In response, the limitation recited are broader than the specific embodiment in the specification; the specification is not read into the claims (MPEP §2111.01). “The words of a claim must be given their "plain meaning" unless such meaning is inconsistent with the specification” (MPEP §2111.01.I).
Furthermore, “it is improper to import claim limitations from the specification” (MPEP §2111.01.II)
In the Appeal Brief, Applicant again argues (page 14, 3rd para through page 16, 2nd para) the same arguments discussed above over page 10-11, relating to alleged technological improvement. Applicant again argues the Desjardins argument (Appeals, page 14-15); Applicant cited the claims in Desjardins and Applicant summarize the claims was patent eligible at Step 2A/Prong two, because the claims are integrated into a practical application due to a technological improvement in machine learning techniques.
In response, the claims in Desjardins was drawn to a particular machine learning technique, and it was found that the claims resulted in an improvement to the technology wherein the improvement was specifically recited in the claims. However, Applicant’s claims are not specified to any specifical machine learning technique. Applicant’s machine learning and deep learning models are recited as “just apply it” as described in MPEP §2106.05(g)).
Additionally, with regard to Applicant’s comparison of the instant claims to Enfish (page 16, 2nd para), the claims in Enfish were founded to be directed to an improved computerized technique, i.e. a self-referential table. In contrast, the only computer technologies recited in the instant claims is a generically recited deep learning model and a generically recited machine learning model. There is no specific radiomic features recited; There is no specific structure on the deep learning model and the machine learning model recited; There is no specific improvement recited.
Hence, the 101 rejections are maintained.
Claim Rejections - 35 USC § 103
This is a newly applied rejection. Necessitated by priority adjustment and claim amendments.
The instant rejection is maintained from the Office Action of 10/1/2025 and modified to address amendments filed 02/17/2026 and the adjustment of priority date.
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective 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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Fan ("Radiomic analysis of DCE-MRI for prediction of response to neoadjuvant chemotherapy in breast cancer patients." European journal of radiology 94 (2017): 140-147. Cited on the 12/19/2023 IDS), in view of Litjens et al.:(“A survey on deep learning in medical image analysis,” Medical Image Analysis 42:60-88 (2017). Newly cited) and Ren et. al.: (Ensemble classification and regression-recent developments, applications and future directions." IEEE Computational intelligence magazine 11.1 (2016): 41-53. Newly cited).
Claim 1 is interpreted as an in silico method to prognosis pharmaceutic treatment outcome. Regarding claim 1, Fan provides (page 140, section “Abstract/Materials and methods”) “A dataset of 57 cancer patients with breast DCE-MR images acquired before NAC was used”, which teaches pre-treatment breast DCE-MRI images before chemotherapy. The images necessarily depict breast cancer/tumor regions because Fan segments breast regions/tumors and predicts response in breast cancer patients. Fan does not teach “a deep learning model”.
Litjens provides (page 60, section “Abstract”) “Deep learning algorithms, in particular convolutional networks, have rapidly become a methodology of choice for analyzing medical images” and (page 60, col 2, 2nd para) “The most successful type of models for image analysis to date are convolutional neural networks (CNNs). CNNs contain many layers that transform their input with convolution filters of a small extent”, which teaches deep learning for analyzing medical images.
Fan provides (page 140, section “Abstract/Materials and methods”) “a total of 158 radiomic features were computed to represent the morphologic, dynamic, and the texture of the tumors as well as the background parenchymal features”, which teaches extracting a set of radiomic features.
Fan provides (page 140, section “Abstract/Materials and methods”) “The optimal subset of features was selected … The classifier was trained and tested … to classify Responder and non-Responder cases”, which teaches providing the set of radiomic features to a machine learning model that are trained to generate a second prediction based on the set of radiomic features.
Fan teaches classifier trained/tested on selected radiomic features to classify responder/non-responder. Litjens teaches deep learning for medical image classification.
Fan provides (page 140, section “Abstract/Materials and methods”) “classifier was trained and tested … to classify Responder and non-Responder cases”, which is inherent to control the machine learning model to generate the second prediction.
Neither Fan nor Litjens teaches generating a classification based on two existing predictions. Ren provides (page 41, col 2, last para) “The main idea behind the ensemble methodology is to aggregate multiple weighted models to obtain a combined model that outperforms every single model in it”, which teaches combining/fusing multiple classifier outputs.
Fan teaches responder/non-responder classification before NAC; Litjens supplies the deep-learning first prediction. Ren supplies combining/fusing multiple classifier outputs; It would be obvious to combine the two predictions from previous steps to generate a final output using Ren’s method.
Fan provides (page 140, section “Abstract/Materials and methods”) “The classifier was trained and tested using a leave-one-out cross-validation (LOOCV) method to classify Responder and non-Responder cases. The area under a receiver operating characteristic curve (AUC) was computed to assess the classifier performance”, which teaches generating output report. Afterall, the claim limitation “outputting the classification” is an ordinary computer implementation.
Regarding claim 2, Fan provides (page 140, section “Abstract/Objectives”) “neoadjuvant chemotherapy (NAC)”, which teaches the pharmaceutical treatment comprises chemotherapy.
Regarding claim 3, Fan provides (page 140, section “Abstract/Objectives”) “neoadjuvant chemotherapy (NAC)”, which teaches neo-adjuvant chemotherapy.
Regarding claim 4, Fan provides (page 140, section “Abstract/Materials and methods”) “breast DCE-MR images” and (page 140, section “Abstract/Objectives”) “breast cancer patients”, which teaches the region of tissue comprises a breast or portion of a breast.
Regarding claim 5, Fan provides (page 142, col 2, last para) “Table 1 summarizes the basic clinical and pathology characteristics of the response and non-response case groups”, which teaches involvement of clinical variables in prediction of the response and non-response.
Regarding claim 6, Fan provides (page 142) Table 1 (table 1 summarizes the basic clinical and pathology characteristics of the response and non-response case groups), which teaches ages and HER2 (Human Epidermal Growth Factor Receptor 2) status of the patients.
Regarding claim 7, Fan provides (page 143, col 2, 3rd para) “Gray-Level Co-Occurrence Matrix (GLCM)-based features were computed in the breast tumor. The calculated GLCM features included sum average, sum entropy, sum variance, variance, contrast, correlation, energy, entropy, correlation and homogeneity”, which suggests a Haralick entropy feature, because the Haralick entropy is computed from GLCM.
Regarding claim 8, Fan provides (page 142, col 2, last para) “Table 1 summarizes the basic clinical and pathology characteristics of the response and non-response case groups”, which teaches involvement of clinical variables (such as ages and HER2 status) in prediction of the response and non-response. Fan also demonstrates pre-treatment prediction of response to NAC (page 140, section “Abstract/Materials and methods”). It is obvious that Fan’s responder/non-responder prediction would naturally inform whether to proceed with NAC or consider alternative treatment.
Regarding claim 9, Fan provides (page 140, section “Abstract/Materials and methods”) “breast DCE-MR images acquired before NAC”, which teaches pre-treatment radiology image is a DCE-MRI image.
Claim 10 is the “system” version of the claim 1 “method”. Regarding claim 10, Fan provides (page 143, col 1, last para) “a computerized scheme was also applied to automatically segment the left and right breast areas”, which suggests a computer system. Hence the art applied to claim 1 also teaches claim 10.
Regarding claim 11, Fan provides (page 140, section “Abstract/Materials and methods”) “breast DCE-MR images” and (page 140, section “Abstract/Objectives”) “breast cancer patients”, which teaches the region of tissue comprises a breast or portion of a breast.
Regarding claim 12, Fan provides (page 140, section “Abstract/Objectives”) “neoadjuvant chemotherapy (NAC)”, which teaches the pharmaceutical treatment comprises chemotherapy.
Regarding claim 13, Fan provides (page 140, section “Abstract/Objectives”) “neoadjuvant chemotherapy (NAC)”, which teaches neo-adjuvant chemotherapy.
Regarding claim 14, Fan provides (page 140, section “Abstract/Materials and methods”) “breast DCE-MR images acquired before NAC”, which teaches pre-treatment radiology image is a DCE-MRI image.
Regarding claim 15, Fan provides (page 142, col 2, last para) “Table 1 summarizes the basic clinical and pathology characteristics of the response and non-response case groups”, which teaches involvement of clinical variables in prediction of the response and non-response.
Regarding claim 16, Fan provides (page 142) Table 1 (table 1 summarizes the basic clinical and pathology characteristics of the response and non-response case groups), which teaches ages and HER2 (Human Epidermal Growth Factor Receptor 2) status of the patients.
Claim 17 is the CRM version of the claim 1 method. Regarding claim 10, Fan provides (page 143, col 1, last para) “a computerized scheme was also applied to automatically segment the left and right breast areas”, which suggests a non-transitory computer-readable storage device. Therefore, the art applied to claim 1 also teaches claim 17
Regarding claim 18, Fan provides (page 143, col 2, last para) “we trained a leave-one-out single feature logistic regression classifier using each of the features individually”, which teaches the .logistic regression classifier.
Regarding claim 19, Fan provides (page 140, section “Abstract/Materials and methods”) “breast DCE-MR images” and (page 140, section “Abstract/Objectives”) “breast cancer patients”, which teaches the region of tissue comprises a breast cancer pathology.
Regarding claim 20, Fan does not teach a classification prediction is a probability. Litjens provides (page 70, col 2, 2nd para) “Estimating the probability that an individual has lung cancer from a CT scan is an important topic”; and (page 70, col 2, 6th para) “In their approach, the task of segmenting membranes of neurons was performed by mild smoothing and thresholding of the output of a CNN, which computes pixel probabilities”, which teaches two classification predictions are probabilities. Hence the first prediction is a first probability and the second prediction is a second probability.
It would have been prima facie obvious to a person of ordinary skill in the art, before the priority date, to modify Fan’s radiomics feature-based NAC-response prediction method by adding or substituting a deep-learning image classifier as taught by Litjens, because deep learning convolutional networks were known to be useful for medical-image classification and could automatically learn image features from radiology images.
A person of ordinary skill would have had a reasonable expectation of success because Fan already demonstrated that pre-treatment DCE-MRI contained predictive information for responder/non-responder classification, Litjens taught the applicability of deep learning to medical-image classification including breast imaging.
It would have been prima facie obvious to a person of ordinary skill in the art, before the priority date, to combine the prediction output of the combined Fan and Litjens pipeline that classify medical images using deep learning method, and the prediction output of Fan’s radiomics feature-based NAC-response prediction using machine learning method, with Ren’s teaching of a simple classifier-fusion rules for combining classifier outputs. Because deep learning convolutional networks were known to be useful for medical-image classification but the radiomic features need be handed separately (Litjens: page 78, Table 11, “CNN outperforms classical radiomics features in patients with esophageal cancer”).
A person of ordinary skill would have had a reasonable expectation of success because combining the prediction from Fan’s radiomic-feature classifier with the prediction from Litjens deep-learning classifier using the classifier-combination teachings of Ren, because combining classifier outputs was a known technique for improving robustness and exploiting complementary information from different classifiers/features.
Response to Applicant’s Arguments
The Remarks filed 2/17/2026 is responded on the Advisory Action (PTOL-303) filed 3/4/2026.
In the Appeal Brief filed 23 Mar 2026 (page 16 through page 23), Applicant argued (page 16, last para through page 20, 1st para) against the reference of Ha. Since Ha is removed due to adjust of the priority date, Fan and Litjens are used instead.
In the Appeal Brief (page 20 paras 2-3), Applicant argued that the reference of Fan and Tax do not remedy the defect of Ha. Since the new combination of Fan, Litjens and Ren are used instead. The argument is no longer effective.
In the Appeal Brief (page 20, 4th para through page 23, 1st para), Applicant argued that the rationale for reference combination falls short. Since the new combination of Fan, Litjens and Ren are different from the previous version. The argument is no longer effective.
Hence, the 103 rejection is maintained.
Conclusion
No claims are allowed.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to GUOZHEN LIU whose telephone number is (571)272-0224. The examiner can normally be reached Monday-Friday 8-5.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Larry D Riggs can be reached at (571) 270-3062. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/GL/
Patent Examiner
Art Unit 1686
/Anna Skibinsky/
Primary Examiner, AU 1635
Prosecution Insights