DETAILED ACTION
Response to Arguments
Applicant’s arguments and amendments in the Amendment filed January 22, 2026 (herein “Amendment”), with respect to the objections to the Specification have been fully considered and are persuasive in part. However, some of the issues, including use of the term “Public Service Announcement” value as it pertains to prostate cancer analysis, have not been corrected, and accordingly, some objections have been maintained.
Applicants arguments and amendments in the Amendment regarding the rejection and objection to previously claims 1–19 are moot because these claims have all been canceled, and a new set of claims, claims 20–39 are presented for examination. The examiner has not found the originally presented claimed invention to have changed and therefore, is examining these claims on the merits, which is proper in this Final action as the claimed subject matter in the previous claims was not examinable due to indefiniteness issues which are resolved in the new set of claims.
Specification
The specification is objected to as failing to provide proper antecedent basis for the claimed subject matter. See 37 CFR 1.75(d)(1) and MPEP § 608.01(o). Correction of the following is required: page 12 recites “a Public Service Announcement (PSA) value,” however, this is new matter as it is added by way of the Amendment filed 1/22/2026, and has no plain meaning or significance in the technological art field of image processing. “PSA” is known however in the field of image processing of prostate cancer images to mean “Prostate Specific Antigen.” Because claim 32 recites “PSA value,” the claimed PSA therefore lacks antecedent support in the Specification. Appropriate correction is required.
Claim Objections
Claim 20 and therefore claims 21–37, and claim 38 and therefore claim 39 which depends therefrom, are objected to because of the following informalities: line 14 in claim 20, and line 9 in claim 38 recites “wherein pre-processing” but should recite “wherein the pre-processing.” Appropriate correction is required.
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 20–21, 25–31, and 36–38 are rejected under 35 U.S.C. 103 as being unpatentable over Fischer et al., “A Radiogenomic Approach for Decoding Molecular Mechanisms Underlying Tumor Progression in Prostate Cancer,” Cancers (Basel). 2019 Sep 2;11(9):1293. doi: 10.3390/cancers11091293. PMID: 31480766; PMCID: PMC6770738 (herein “Fischer”) further in view of Cosma et al., “Prediction of Pathological Stage in Patients with Prostate Cancer: A Neuro-Fuzzy Model,” PLOS One, June 3, 2016, DOI: 10.1371/journal.pone.0155856 (herein “Cosma”) in view of Madabhushi et al., US Patent Application Publication No. US 2020/0000396 A1 (herein “Madabhushi”) and in view of Mopur et al., US Patent Application Publication No. US 2023/0316710 A1 (herein “Mopur”).
Regarding claims 20 and 38, with substantive differences between the claims noted in curly brackets {}, with deficiencies in Fischer noted in square brackets [], and with claim 20 as exemplary, Fischer teaches {a system – claim 20 / a computer-implemented method – claim 38} for predicting diseases in an early phase using artificial intelligence, the {system – claim 20 / method – claim 38} comprising {at least one hardware processor; and at least one non-transitory computer-readable memory storing program instructions that, when executed by the at least one hardware processor, cause the at least one hardware processor to: - claim 20 only} (Fischer page 3, fig. 1, various data is input into a model learning and prediction model for predicting a pathological stage of cancer (disease) including an earlier T2c stage (early phase), where pages 12–13, sections 3.2, 3.3, and 3.5–3.7, teach that the R software package was used to create and execute the disclosed AI/ML operations, and where it would be understood by a person having ordinary skill in the art that in order to use a software package, both memory and processors are required):
collect {a plurality of – claim 38} medical images in digital format from a plurality of medical prediction centers and a plurality of medical record databases (Fischer page 12, sections 3.1 and 3.2.3, MRI imaging data from the TCGA-PRAD collection of the Cancer Imaging Archive, a collection of data from multiple institutions, medical centers and hospitals, and the respective record databases thereof), wherein the {collected – claim 20 / plurality – claim 38} medical images are captured using {at least one of – claim 38} a general-purpose camera or real-time medical image capturing tools including computed tomography (CT), radiology, magnetic resonance imaging (MRI), ultrasound, or nuclear medicine imaging (Fischer page 12, sections 3.1 and 3.2.3, MRI data is specifically used by the disclosed system, and additionally, the TCGA-PRAD Cancer Imaging Archive collection has all of MRI, CT and PT formats of images);
pre-process the {collected – claim 20 / plurality – claim 38} medical images to {enhance visual quality – claim 20 / generate pre-processed images – claim 38} by reducing noise and identifying image texture, color, and shape {to produce clean images – claim 20} (Fischer page 12, section 3.2.3, pre-processing including determining imaging feature categories C2 and C3, which include texture, Gray-level co-occurrence matrix features (color) and homogeneity (shape)), wherein pre-processing comprises [resizing the {collected/plurality of} medical images to a lower pixel resolution to reduce processing time], [cropping] the {collected/plurality of} medical images to remove unnecessary areas while retaining a region of interest, and [transforming a red-green-blue (RGB) color space of the {collected/plurality of} medical images to grayscale intensity to remove undesired color variations] (Fischer page 12, section 3.2, MRI imaging data is subjected to median filters for noise reduction and segmented using the NIH ImageJ software to determine a region of interest);
segment the {clean – claim 1 / pre-processed – claim 38} images to extract the region of interest from an image background [by identifying pixel characteristics and dividing each {pre-processed – claim 38} image into segments consisting of similar characteristic {pixels – claim 20}] (Fischer page 12, section 3.2, MRI imaging data is segmented using the NIH ImageJ software to determine the region of interest);
extract a set of {image – claim 38} features from the region of interest, the set of {image – claim 38} features selected from asymmetry index, entropy, [autocorrelation], homogeneity, and contrast, for use in a classification stage (Fischer page 12, imaging feature categories C2 and C3 are extracted from the images (select optimized features), where page 9, fig. 6 teaches that radiographic C3 features include texture analysis including homogeneity, energy, contrast and entropy, and radiographic C2 features represent the histogram of tumor volume intensity and basic statistical metrics such as standard deviation (an asymmetry index)), and select optimized features from the set of {image – claim 38}features (Fischer page 9, fig. 6d, and page 10, section 2.6, Pearson correlation computed upon the imaging features to determine significant correlations to biomarkers, thus identifying those features most optimally indicative of relevant biomarkers for disease states);
train a [fuzzy] logic-based (Fischer page 11, although not used by the disclosed system, machine learning using Fuzzy logic to predict cancer pathological stage (with citation to reference 19, which is the secondary reference Cosma cited in this rejection) prediction model [and a plurality of diagnosis-specific treatment response models] using the optimized features (Fischer fig. 1, page 3, and pages 13–14, section 3.7, prediction models, biomarkers, including those identified using the Pearson correlation to the image features, thus “using the optimized features”, are used to train three machine-learning methods), and [store the trained fuzzy logic-based prediction model and the trained diagnosis-specific treatment response models in a cloud {server – claim 20 / based storage – claim 38} platform, {wherein the fuzzy logic-based prediction model comprises a fuzzification stage, an inference stage using a rule base, a knowledge base storing rules and associated data, and a defuzzification stage – claim 20 only}]; and
receive a subject patient dataset including features corresponding to a reduced feature dataset, compare the subject patient dataset to a feature data scheme {using – claim 20 / by applying – claim 28} the [fuzzy logic-based] prediction model (Fischer page 3, fig. 1, and page 13, section 3.7, page 12, section 3.1, 30% of the data was used for testing (testing the inferencing accuracy), and thus, the testing data being received by the trained machine learning model, the data including TCGA samples of a tissue biopsy taken from unique individual (subject patient dataset), where page 10, section 2.7 teaches using the features most strongly correlated to the image features, such as the biomarkers ANPEP, mir-217, mir-592 and mir-6715b (reduced feature dataset)) to determine a diagnosis {of the subject patient – claim 38} corresponding to one of a plurality of known disorders (Fisher page 10, section 2.7, the output of the machine learning methods was a categorized pathological stage, including the T2c stage, and the T3b stage, these stages corresponding to prostate cancer (one of a plurality of known cancer disorders)), [apply a corresponding diagnosis-specific treatment response model to predict a treatment response for the subject patient], and generate [a medical report] {including – claim 20 / comprising – claim 38} a disease severity and a disease stage (Fischer page 3, section 2.1, figure 1, the output is a predicted pathological stage, including stages T3c and T3b, corresponding to severity/spread of prostate cancer disease).
While Fischer mentions on page 11 that a Fuzzy logic model can be used to predict prostate cancer stages and cites to the Cosma reference as an example, Fischer’s main disclosure uses other kinds of machine learning models. However, Cosma teaches a fuzzy logic based prediction model, wherein the fuzzy logic-based prediction model comprises a fuzzification stage, an inference stage using a rule base, a knowledge base storing rules and associated data, and a defuzzification stage (Cosma pages 6–7, fig. 1, Fuzzy logic system to predict the pathological stage of cancer (prediction model) including a Fuzzy C-Means stage detailed further on pages 8–9 (fuzzification), a Neuro-Fuzzy predictor in a predictor stage (inference) detailed on page 10 that applies Takagi-Sugeno-Kang (TSK) rules (rule base), a FIS structure (knowledge base) mapping (rules) input data to output data (associated data) detailed on page 9, and the adaptive-neuro fuzzy inference system detailed on page 9 which outputs the summation of the contribution from each rule (defuzzification)). Cosma also teaches the trained fuzzy logic-based prediction model (Cosma pages 9–10, and page 7, figure 1, during the training stage, the ANFIS module tunes FIS parameters and optimizes them for use in the prediction stage, thus once the FIS parameters are tuned, the fuzzy-logic based prediction model is trained).
Fischer further does not explicitly teach where Madabhushi teaches resizing the collected medical images to a lower pixel resolution to reduce processing time (Madabhushi ¶ 53, MRI images are downsampled for further processing which would realize a reducing in processing time as less data means less computations), cropping the collected medical images (Madabhushi fig. 4, ¶ 56, features within a lesion ROI are extracted by selecting a section of the image, shown in fig. 4 as a box in the main image that is zoomed in for a specific region of interest), transforming a red-green-blue (RGB) color space of the collected medical images to grayscale intensity to remove undesired color variations (Madabhushi ¶53, RGB pathology images are converted to grayscale), by identifying pixel characteristics and dividing each image into segments consisting of similar characteristic pixels (Madabhushi ¶69, segmentation defines a tumoral boundary using an active color technique, considering the similar color of pixels), and a plurality of diagnosis-specific treatment response models (Madabhushi ¶¶ 47–48, 74, treatment plan generated based on a classification model), apply a corresponding diagnosis-specific treatment response model to predict a treatment response for the subject patient (Madabhushi ¶¶ 48, 74, treatment plain is personalized to the particular patient using machine learning classifier) and generate a medical report (Madabhushi ¶¶ 74–78, a personalized treatment plan is displayed on a display circuit reporting on specific treatments (surgery, pharmaceutical agent dosage and schedule) that will best suit the patient), store prediction model and the trained diagnosis-specific treatment response models in a cloud server platform (Madabhushi ¶¶ 46–48, personalized treatment plan for prostate cancer (diagnosis-specific) generated from the machine learning classifier which produces classification and is trained (trained treatment response model), where ¶24 teaches a machine learning classifier also for predicting a DECIPHER risk group (prediction model), and where ¶¶ 79 and 84 teaches the disclosed computer interacting with the cloud).
While Madabhushi does not explicitly teach a plurality of classifiers, or that while the computer that trains the models interacts with the cloud, the computer is not explicitly taught such interaction to include storing models in the cloud, nonetheless, it would have been obvious to modify the teachings of Fischer to include a plurality of the diagnosis-specific treatment response models of Madabhushi and storing same in the cloud taught by Madabhushi at least as the plurality of models would be a duplication of parts (Madabhushi’s disclosed model) per MPEP §2144.04(VI)(B) – merely duplicating the classifier model disclosed by Madabhushi, with expected result of having a binary classification per classifier, rather than the multiple outputs from a single classifier. This modification would also be using a well-known ensemble learning model rather than a single classifier model. Accordingly, such a modification would also be simple substitution of one known element for another to obtain predictable results, per MPEP §2143(I)(B). Regarding Madabhushi’s interaction with the cloud including storing of models, this would be obvious to a person having ordinary skill in the art as storing data in the cloud is a known technique, and as such would be use of known technique to improve similar devices (methods, or products) in the same way, per MPEP §2143(I)(C).
Further, while Fischer does not explicitly teach storing models in the cloud, or autocorrelation as an extracted image feature, Mopur teaches storing models in the cloud (Mopur ¶¶24, 30 environments hosted on cloud servers including trained machine learning models which are then deployed to edge devices), and autocorrelation as an extracted image feature (Mopur ¶¶ 51–53, change in images as a feature detected via auto-correlation).
Therefore, taking the teachings of Fischer and Cosma together as a whole, it would have been obvious to a person having ordinary skill in the art (herein “PHOSITA”) before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the Fuzzy logic processing teachings of Cosma at least because Fischer explicitly cites to the Cosma reference on page 11 as an alternative machine learning model that can be used, and as well Cosma states that its neuro-fuzzy model is one that outperforms all other systems for predicting pathological stage. See Cosma page 23, first paragraph.
Further, taking the teachings of Fischer as modified by Cosma and Madabhushi together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the further image pre-processing, medical report generation and treatment response model disclosed in Madabhushi at least because doing so would further facilitate improved identification of cancer patients who would receive added benefit from a specific therapy amongst a set of potential candidate therapies. See Madabhushi ¶ 19.
Still further, taking the teachings of Fischer as modified by Cosma, Madabhushi and Mopur together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the model cloud storage and autocorrelation disclosed in Mopur at least because doing so would provide saving of time, bandwidth and computing power to keep ML models accurate for accurate predictions used for decision making. See Mopur ¶ 60.
Regarding claim 21, Fischer teaches wherein the [fuzzy logic-based] prediction model is employed to determine diagnoses of a plurality of known disorders indicated by individual patient datasets (Fischer page 3, section 2.1, figure 1, machine learning model predicts pathological stage of cancer, and where page 12, section 3.1 teaches datasets from individuals from tissue biopsy (individual patients)). Fischer does not teach where Cosma teaches the fuzzy-logic based model (Cosma Abstract, neuro-fuzzy model for classifying and predicting diseases) and Madabhushi teaches wherein each diagnosis-specific treatment response model corresponds to a specific diagnosis of the known disorders and is configured to use feature data to predict treatment response (Madabhushi ¶¶ 41–42, and 47–48, treatment plan is personalized to the particular patient for their diagnosis using machine learning classifier which uses features from MRI imagery to train the classifier).
The motivations to combine Fischer with Cosma and Madabhushi are the same as those previously set forth above for claim 20.
Regarding claim 25, Fischer teaches further comprising generating the feature data scheme using training data comprising patient clinical data, laboratory data, diagnosis data, and treatment response data (Fischer page 12, sections 3.1–3.2, training data from the TCGA-PRAD Cancer Imaging Archive project including tissue biopsy data taken from unique individuals including tumor samples in different stages, clinical data and genomic data), wherein the feature data scheme includes a reduced feature dataset having a lower cardinality than an extracted feature dataset (Fischer page 10, section 2.7, and page 9, figure 6, diagnostic features for predicting the pathological stages are reduced from the original feature set in the training data to be those features that strongly correlate to imaging features).
Regarding claim 26, Fischer teaches wherein generating the reduced feature dataset comprises computing standardized feature values and applying weighting functions derived from marker variation data (Fischer page 9, figure 6, feature values are determined using radiographic features C2 and C3 from basic statistical metrics (standardized values) as they correlate to biomarkers, the features derived from T2-weighted images).
Regarding claim 27, Fisher does not explicitly teach where Madabhushi teaches wherein the program instructions further cause the at least one hardware processor to analyze the collected medical images using a trained recognition model to generate a lesion recognition report indicating whether a lesion is present (Madabhushi ¶¶ 64–65 and 71–72, risk group prediction circuit for a prostate cancer pathology (lesion is present) analyzing medical images to determine a risk group, where a display circuit displays the classification (report)).
Therefore, taking the teachings of Fischer as modified above and Madabhushi together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the report generation disclosed in Madabhushi at least because doing so would further facilitate improved identification of cancer patients who would receive added benefit from a specific therapy amongst a set of potential candidate therapies. See Madabhushi ¶ 19.
Regarding claim 28, Fisher teaches wherein includes a lesion degree label determined using feature responses extracted from segmented image blocks of the collected medical images (Fisher page 3, figure 1, from imaging data, the pathological stage is predicted and output, from amongst degrees of progression T2c and T3b of a prostate tumor). Fisher does not explicitly teach where Madabhushi teaches the lesion recognition report (Madabhushi ¶¶ 71–72, analyzing medical images to determine a risk group, where a display circuit displays the classification (report)). The motivations to combine Fischer with Madabhushi are the same as those previously set forth above for claim 27.
Regarding claim 29, Fischer teaches wherein the disease stage is defined within a numeric range denoting increasing disease severity (Fischer page 2, first full paragraph, and page 3, clinical stages of prostate cancer according to a Partin tables with a numeric value 1 to 4, increasing in severity). The motivations to combine Fischer with Madabhushi are the same as those previously set forth above for claim 20.
Regarding claim 30, Fisher does not, whereas Madabhushi teaches wherein the program instructions further cause the at least one hardware processor to generate a treatment plan corresponding to the disease stage and a disease type (Madabhushi ¶¶ 74–78, a personalized treatment plan is displayed on a display circuit reporting on specific treatments (surgery, pharmaceutical agent dosage and schedule) that will best suit the patient and corresponds to the risk group (disease stage) and for prostate cancer (disease type)). The motivations to combine Fischer with Madabhushi are the same as those previously set forth above for claim 20.
Regarding claim 31, Fisher does not, whereas Madabhushi teaches wherein the treatment plan for cancer includes generating a radiotherapy dose distribution based on anatomical imaging data of a human subject (Madabhushi ¶¶ 70–71, 74–78, a personalized treatment plan is displayed on a display circuit reporting on specific treatments (surgery, pharmaceutical agent dosage and schedule) that will best suit the patient, using the machine learning classifier that is trained using the MRI images). The motivations to combine Fischer with Madabhushi are the same as those previously set forth above for claim 20.
Regarding claim 36, Fischer teaches wherein the fuzzy logic-based prediction model further evaluates quantified expression levels of at least two cancer-related proteins measured from a biological sample of a subject as prognostic indicators of cancer (Fischer page 12, biomarkers are associated with protein levels, with page 5 and fig. 3 teaching the AKT pathway associated with advanced prostate cancer stages).
Regarding claim 37, Fischer teaches wherein at least one of the cancer-related proteins is selected from AKT, p-AKT, MAPK, p-MAPK, mTOR, p-mTOR, EGFR, HER2, HER3, PTEN, or PSA; or the cancer-related proteins are associated with a carcinoma selected from breast, lung, prostate, colon, liver, thyroid, kidney, or bile duct carcinoma (Fischer page 5 and fig. 3 teaching the AKT pathway associated with advanced prostate cancer stages).
Claims 22–23 are rejected under 35 U.S.C. 103 as being unpatentable over Fischer in view of Cosma in view of Madabhushi in view of Mopur, as set forth above regarding claim 20, and further in view of Chinese Published Patent Application No. CN 111430030A by Wu Debin et al., with reference to the EPO generated machine English language translation, (herein “Debin”).
Regarding claim 22, Fischer does not explicitly teach, where Debin teaches wherein the program instructions further cause the at least one hardware processor to perform an ovarian cancer assessment comprising: obtaining a three-dimensional ovary image of a subject using clinical imaging equipment and performing image denoising and enhancement on the three-dimensional ovary image (Debin page 4, step S1: collect three-dimensional ovarian images of the subject through medical imaging equipment, and perform image denoising and enhancement processing); comparing the enhanced three-dimensional ovary image to identify a location of an ovarian tumor (Debin page 4, step S2: determine the position of the ovarian tumor according to the enhanced image); receiving concentration data for at least one small-molecule biomarker measured from an ovarian cancer tumor sample of the subject (Debin page 4, step S3: measure the concentration of at least one small molecule biomarker in the subject's ovarian cancer tumor by a medical measuring device); comparing the concentration data to a threshold value and, upon satisfaction of a threshold condition, receiving CA-125 data, HE4 data, and PA data associated with a serum sample of the subject (Debin page 4, steps S4 and S5: compare the concentration of the obtained small molecule biomarker with the control sample; and if the concentration of the small molecule biomarker exceeds or falls below the corresponding threshold, the identification procedure is used to obtain the CA125 data, HE4 data and PA data (tumor markers) of the subject's serum sample to be tested); and evaluating the concentration data, the CA-125 data, the HE4 data, and the PA data to generate an ovarian cancer evaluation report (Debin page 5, steps S6 and S7: According to the level of the biomarker and the area value under the working characteristic curve based on the CA125 data, HE4 data and PA data, the ovarian cancer condition of the subject is evaluated through the evaluation program, and an evaluation report is generated for the doctor to diagnose and select the treatment method).
Therefore, taking the teachings of Fischer as modified above and Debin together as a whole, it would have been obvious to a “PHOSITA” before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the ovarian cancer assessment teachings of Debin as doing so would well monitor changes in ovarian cancer and improve the accuracy of assessment. See Debin page 1, Summary of the Invention section.
Regarding claim 23, Fischer does not explicitly teach, where Debin teaches wherein the concentration data for the at least one small-molecule biomarker are generated using a biomarker-based ovarian cancer assessment method comprising obtaining a biological sample from the subject, detecting a small-molecule biomarker selected from hydroxy acids, adipic acid, hydroxybutyric acid, dihydroxybutyric acid, trihydroxybutyric acid, or ketone bodies, and comparing a measured concentration of the small-molecule biomarker to a reference frequency profile (Debin page 7, sections I-III: Obtain the sample from the subject, and determine the concentration of the biomarker in the sample; the sample is selected from the group consisting of blood, serum and plasma; (II) The detection of small molecule biomarkers is carried out by contacting the sample with antibodies or antigen-binding fragments that can specifically bind to small molecule biomarkers of ovarian cancer; mainly to detect the content and concentration of small molecule markers for the measurement The obtained concentration is compared with the concentration reference frequency of the reference substance; (III) Comparing the determined concentration of the biomarker with the reference frequency distribution of the biomarker concentration, and reading the deciles from the frequency distribution of the biomarker concentration, and in section III: the at least one additional small molecule biomarker whose concentration is increased is selected from the group consisting of hydroxy acid and adipic acid).
Therefore, taking the teachings of Fischer as modified above and Debin together as a whole, it would have been obvious to a “PHOSITA” before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the ovarian cancer assessment teachings of Debin as doing so would well monitor changes in ovarian cancer and improve the accuracy of assessment. See Debin page 1, Summary of the Invention section.
Claim 24 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer in view of Cosma in view of Madabhushi in view of Mopur, as set forth above regarding claim 20, and further in view of Panetta et al., US Patent Application Publication No. US 2014/0341481 A1 (herein “Panetta”).
Regarding claim 24, Fischer teaches wherein pre-processing the collected medical images further comprises filtering operations including median filtering, adaptive median filtering (Fischer page 12, MRI imaging data for two stages were subjected to median filters, where the filters would be adaptive to the data per its stage). Fischer does not whereas Panetta teaches applying image scaling (Panetta ¶¶ 13–14, multi-scale transformations for denoising), contrast enhancement (Panetta ¶11, original image processing including contrast enhancement), image restoration and Gaussian filtering (Panetta ¶55, denoising using Gaussian filtering).
Therefore, taking the teachings of Fischer as modified above and Panetta together as a whole, it would have been obvious to a “PHOSITA” before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the pre-processing teachings of Panetta to leverage the benefits of multiple enhancements to solve real life multimedia issues. See Panetta ¶5.
Claims 32–34 are rejected under 35 U.S.C. 103 as being unpatentable over Fischer in view of Cosma in view of Madabhushi in view of Mopur, as set forth above regarding claim 20, and further in view of Chinese Published Patent Application No. CN 110993095A by Ye Lin et al., with reference to the EPO generated machine English language translation, (herein “Lin”).
Regarding claim 32, with deficiencies of Fischer noted in square brackets [], Fischer teaches wherein the program instructions further cause the at least one hardware processor to predict prostate cancer occurrence and metastasis based on a [three-dimensional] image of a prostate and bladder region of a subject and clinical parameters (Fischer fig. 1, pages 3 and 12, imaging data as shown including the prostate and bladder region is used to determine the pathology of prostate cancer along with clinical data) [including age, rectal index, family genetic history, PSA value, PIRADS score, and a ratio of peripheral prostate fat area to prostate area].
Fischer does not explicitly teach a three-dimensional image, although Fischer does teach MRI images. Further, Fischer does not teach the parameters including age, rectal index, family genetic history, PSA value, PIRADS score, and a ratio of peripheral prostate fat area to prostate area.
Cosma teaches including age, and PSA value (Cosma page 8, data normalization of input data including age and PSA level).
Madabhushi teaches PIRADS score (Madabhushi ¶52, PIRADS v2 score used in analysis).
Lin teaches a three-dimensional image, rectal index, family genetic history and a ratio of peripheral prostate fat area to prostate area (Lin pages 1-2, stereoscopic images of the human prostate and bladder are analyzed, and verbatim from Lin: “the above processing device is based on age variable, digital rectal variable, family genetic history variable, prostate imaging report and data system scoring variable, PSA value variable, the ratio of peripheral fat area to prostate area ratio variable, according to the formula to calculate the initial diagnosis of prostate cancer Risk value, the output device outputs the risk value of newly diagnosed prostate cancer, the formula is as follows: Logit (P) = In (P / (1-P)) = 1.037 * Age + coefDRE + coefHistory + 1.033 * PSA + coefPIRADS + 1.066 * (PPFA / PA)”).
Therefore, taking the teachings of Fischer and Cosma together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the age and PSA value teachings in Cosma at least because doing so would provide a predictive system that outperforms all other systems for predicting pathological stage. See Cosma page 23, first paragraph.
Further, taking the teachings of Fischer as modified by Cosma and Madabhushi together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the PIRADS score disclosed in Madabhushi at least because doing so would further facilitate improved identification of cancer patients who would receive added benefit from a specific therapy amongst a set of potential candidate therapies. See Madabhushi ¶ 19.
Still further, taking the teachings of Fischer as modified by Cosma, Madabhushi and Lin together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the stereoscopic images, rectal index, family genetic history and PPFA/PA teachings disclosed in Lin at least because doing so would allow for the prediction of occurrence and metastasis of prostate cancer without puncture biopsy. See Lin page 1.
Regarding claim 33, Fischer does not explicitly teach where Madabhushi teaches wherein the program instructions further cause the at least one hardware processor to calculate a prostate cancer lymph node metastasis risk value using imaging data and pathological parameters (Madabhushi Abstract, patient prostate cancer DECIPHER risk group calculated from features extracted from MRI images and radiomic features (pathological parameters)).
Therefore, taking the teachings of Fischer as modified by Cosma and Madabhushi together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the DECIPHER risk group calculation disclosed in Madabhushi at least because doing so would further facilitate improved identification of cancer patients who would receive added benefit from a specific therapy amongst a set of potential candidate therapies. See Madabhushi ¶ 19.
Regarding claim 34, Fischer does not explicitly teach where Lin teaches wherein predicting prostate cancer occurrence comprises combining an age parameter, a rectal index parameter, a family genetic history parameter, a prostate-specific antigen (PSA) value parameter, and a prostate imaging reporting and data system (PIRADS) scoring parameter with the ratio of peripheral prostate fat area to prostate area (Lin pages 1-2, verbatim: “the above processing device is based on age variable, digital rectal variable, family genetic history variable, prostate imaging report and data system scoring variable, PSA value variable, the ratio of peripheral fat area to prostate area ratio variable, according to the formula to calculate the initial diagnosis of prostate cancer Risk value, the output device outputs the risk value of newly diagnosed prostate cancer, the formula is as follows: Logit (P) = In (P / (1-P)) = 1.037 * Age + coefDRE + coefHistory + 1.033 * PSA + coefPIRADS + 1.066 * (PPFA / PA)”).
Therefore, taking the teachings of Fischer as modified by Cosma, Madabhushi and Lin together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the image processing of Fischer to include the prediction by combining different factor teachings disclosed in Lin at least because doing so would allow for the prediction of occurrence and metastasis of prostate cancer without puncture biopsy. See Lin page 1.
Claim 35 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer in view of Cosma in view of Madabhushi in view of Mopur, as set forth above regarding claim 20, and further in view of Sottas et al., US Patent Application Publication No. US 2021/0217528 A1 (herein “Sottas”).
Regarding claim 35, Fischer does not teach but Cosma teaches the fuzzy logic-based prediction model (Cosma pages 6–7, fig. 1, Fuzzy logic system to predict the pathological stage of cancer (prediction model). Further Fischer does not explicitly teach where Sottas teaches further evaluates quantitative blood flow indicators measured before and after administration of a pharmaceutical agent to determine a physiological response (Sottas ¶¶ 72, 44, response to drug treatment is evaluated using a biological signal such as a variation in plasma volume according to a Z score, with the following verbatim passage from Sottas: “(2.3) The individual Z-scores are then combined using a weighting function derived from the known variations of each marker with plasma volume as well as from the consistency between all Z-scores. The outcome, given as a Z-score, is an estimate of the variations in plasma volume.”).
The motivation to combine Fischer and Cosma is the same as previously set forth above regarding claim 20.
Further, taking the teachings of Fischer as modified above in the rejection for claim 20 and Sottas together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the prediction model of Fischer to include the plasma volume calculation as disclosed in Sottas at least because doing so would allow for improved decisions in the diagnosis, prognosis and treatment of a disease. Sottas ¶39.
Claim 39 is rejected under 35 U.S.C. 103 as being unpatentable over Fischer in view of Cosma in view of Madabhushi in view of Mopur, as set forth above regarding claim 20, and further in view of Hengerer et al., US Patent Application Publication No. US 2010/0247438 A1 (herein “Hengerer”).
Regarding claim 39, Fischer does not explicitly teach where Madabhushi teaches using the diagnosis-specific treatment response models (Madabhushi ¶¶ 46–48, personalized treatment plan for prostate cancer (diagnosis-specific) generated from the machine learning classifier which produces classification and is trained (trained treatment response model)), and where Hengerer teaches wherein predicting the treatment response further comprises performing an in vitro tumor assessment method (Hengerer Abstract, diagnosing tumor disease using an IVD marker, the diagnosis repeated after a defined time interval (for treatment response determination)), the in vitro tumor assessment method comprising: identifying at least one in vitro diagnostic (IVD) marker or IVD marker panel having a sensitivity above a threshold value for a tumor disease using a biological sample obtained from a patient (Hengerer fig. 3, ¶34, tumor marker with a threshold to determine in an in vitro test the sensitivity of the marker); determining a modified reference range for the at least one IVD marker or IVD marker panel such that false-negative results, false-positive results, and follow-up imaging determinations are balanced to enable tumor screening (Hengerer ¶32, patients with a positive test result are now subjected to a continuative, i.e. imaging, examination in order ultimately to diagnose the possible presence of a tumor. The method according to an embodiment of the invention enriches true positive patients and so the number of patients who are unnecessarily subjected to an imaging examination can be minimized); and selectively performing a tumor-specific imaging technique based on the modified reference range to clarify at least one of a false-negative result or a false-positive result (Hengerer ¶38, the reference range of a specific tumor marker now is adapted such that the number of individuals with false negative tests, the number of individuals with false positive tests and the number of individuals ultimately needing to be subjected to imaging diagnostics to clarify false negative and false positive results are balanced in respect of one another such that tumor screening can be carried out), or repeating the identifying and determining steps after a predetermined time interval (Hengerer Abstract, ¶ 14, IVD marker testing is repeated after a defined time interval), wherein the biological sample is selected from a blood sample, a serum sample, a plasma sample, a urine sample, a fecal sample, a saliva sample, a cerebrospinal fluid sample, a nasal discharge sample, a sputum sample, a bronchoalveolar lavage sample, a semen sample, abreast discharge sample, a wound discharge sample, an ascites sample, a gastric juice sample, or a sweat sample (Hengerer ¶43, the examination material usually being whole blood, serum, plasma, ascites, liquor, urine, feces, saliva, spinal fluid, nasal discharge, sputum, BAL, semen, breast discharge, wound discharge, gastric juices, sweat, or breath condensate).
Therefore, taking the teachings of Fischer as modified above in the rejection for claim 38 and Hengerer together as a whole, it would have been obvious to a PHOSITA before the effective filing date of the claimed invention to have modified the prediction model of Fischer to include the IVD marker testing as disclosed in Hengerer at least because doing so would allow for diagnosis of a tumor disease as early as possible without harboring the risk of overlooking too large proportion of cancer patients. See Hengerer ¶ 12.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure: Sedlak et al., US PGPub No. US 2025/0327129 A1, directed towards technologies for detection of ovarian cancer.
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to MICHELLE M KOETH whose telephone number is (571)272-5908. The examiner can normally be reached Monday-Thursday, 09:00-17:00, Friday 09:00-13:00, EDT/EST.
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, Vincent Rudolph can be reached at 571-272-8243. 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.
MICHELLE M. KOETH
Primary Examiner
Art Unit 2671
/MICHELLE M KOETH/Primary Examiner, Art Unit 2671