DETAILED CORRESPONDANCE
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Status of claims
This final office action on merits is in response to the communication received on 02/02/2026. Claim 4 is cancelled. Claims 21-30 are new. Amendments to claims 1-3, 5-7, 10-14, and 20 are acknowledged and have been carefully considered. Claims 1-3, and 5-30 are pending and considered below.
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, and 5-30 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Step 1
Under step 1, the analysis is based on MPEP 2106.03, and claims 1-3, 5-10, 21-30 are drawn to a method and claims 11-19 are drawn to an apparatus, and claim 20 is drawn to a computer readable medium. Thus, each claim, on its face, is directed to one of the statutory categories (i.e., useful process, machine, manufacture, or composition of matter) of 35 U.S.C. §101.
Step 2A Prong One
Claim 11 recites the limitations of calculating a score associated with the patient using an algorithm to analyze the received health data, the score being indicative of a likelihood of one or more target diseases; and wherein: for each examination parameter, the relative criticality is calculated based on an extent to which each respective examination parameter deviates from at least one respective reference standard associated with the one or more target diseases; and the indication is a discrete indication generated by associating the calculated relative criticality to one of a plurality of discrete ranges defined by the at least one respective reference standard. These limitations, as drafted, are processes that, under their broadest reasonable interpretations, cover performance of the limitations in the mind or by using a pen and paper. Even when considering the “a processor; and a memory to store computer program instructions for providing a score associated with a patient” language, the claim encompasses a user reviewing examination parameters, comparing the examination parameters to reference standards, determining the degree of deviation from the reference standards, evaluating the likelihood of disease based on the examination parameters, categorizing the relative criticality into predefined ranges, and forming a conclusion regarding disease risk in their mind or by using a pen and paper. The mere recitation of a processor and a memory to store computer program instructions for providing a score associated with a patient does not take the claim limitations out of the mental processes grouping. Thus, the claim recites a mental process which is an abstract idea.
Independent claims 1 and 20 recite identical or nearly identical steps with respect to claim 1 (and therefore also recite limitations that fall within this subject matter grouping of abstract ideas), and these claims are therefore determined to recite an abstract idea under the same analysis.
Under Step 2A Prong Two
The claimed limitations, as per claim 11, include:
a processor; and a memory to store computer program instructions for providing a score associated with a patient, which, when executed on the processor cause the processor to perform operations comprising:
receiving health data associated with a patient, the health data including examination parameters;
calculating a score associated with the patient using an algorithm to analyze the received health data, the score being indicative of a likelihood of one or more target diseases; and
displaying a report comprising the score and an indication of a relative criticality of each of the examination parameters
wherein: for each examination parameter, the relative criticality is calculated based on an extent to which each respective examination parameter deviates from at least one respective reference standard associated with the one or more target diseases; and
the indication is a discrete indication generated by associating the calculated relative criticality to one of a plurality of discrete ranges defined by the at least one respective reference standard.
Examiner Note: underlined elements indicate additional elements of the claimed invention identified as performing the steps of the claimed invention.
The judicial exception expressed in claim 11 is not integrated into a practical application. The claim as a whole merely describes how to generally “apply” the concept of evaluating patient examination parameters to determine disease likelihood and relative criticality in a computer environment. The claimed computer components (i.e., a processor; and a memory to store computer program instructions for providing a score associated with a patient, which, when executed on the processor cause the processor to perform operations comprising) are recited at a high level of generality and are merely invoked as tools to perform an existing process of analyzing health data, comparing examination parameters to reference standards, calculating disease risk scores, and categorizing results into predefined ranges. Simply implementing the abstract idea on a generic computer is not a practical application of the abstract idea. Accordingly, alone and in combination, these additional elements do not integrate the abstract idea into a practical application.
The judicial exception expressed in claim 11 is not integrated into a practical application. The claim recites the additional elements of receiving health data associated with a patient, the health data including examination parameters; and displaying a report comprising the score and an indication of a relative criticality of each of the examination parameters. These limitations are recited at a high level of generality (i.e., as a general means of collecting data and presenting data), and amounts to merely data gathering and outputting or displaying results, which are forms of insignificant extra-solution activities. Accordingly, even in combination, these additional elements do not integrate the abstract idea into a practical application. The claim is directed to an abstract idea.
Therefore, under step 2A, the claims are directed to the abstract idea, and require further analysis under Step 2B.
Under step 2B
Claim 1 does not include additional elements that are sufficient to amount to significantly more than the judicial exception. As discussed with respect to Step 2A, the claim as a whole merely describes how to generally “apply” the concept of evaluating patient examination parameters to determine disease likelihood and relative criticality in a computer environment. Thus, even when viewed as a whole, nothing in the claim adds significantly more (i.e., an inventive concept) to the abstract idea.
For claim 11, under step 2B, the additional elements of receiving health data associated with a patient, the health data including examination parameters; and displaying a report comprising the score and an indication of a relative criticality of each of the examination parameters have been evaluated. The apparatus comprising a processor; and a memory to store computer program instructions performs a general function of receiving patient data for analysis and risk evaluation, which represents a well-understood, routine, and conventional activity in the field of medical data processing and clinical decision support systems. The specification discloses that the processor is used in its ordinary capacity as a data input device and does not describe any improvement to the computer itself or to the functioning of the overall computer system (see [0094]). Also noted in Electric Power Group, LLC v. Alstom S.A., 830 F.3d 1350, 1354, 119 USPQ2d 1739, 1742 (Fed. Cir. 2016), merely collecting information for analysis without a technological improvement does not add significantly more to an abstract idea. The use of the apparatus is no more than collecting information before performing analysis and evaluation of examination parameters to determine disease likelihood and relative criticality, and then displaying the results, and does not integrate the abstract idea into a practical application. Therefore, the claim does not recite an inventive concept and is not patent eligible.
Claims 2-3, 5-6, 12-13, 16-17, 24-27 recite no further additional elements, and only further narrow the abstract idea. The previously identified additional elements, individually and as a combination, do not integrate the narrowed abstract idea into a practical application for reasons similar to those explained above, and do not amount to significantly more than the narrowed abstract idea for reasons similar to those explained above.
Claims 7-10, 14-15, 18-19, 21-23, and 28-30 recite the additional elements of displaying a second report comprising the second score and an indication of the relative criticality of each of the examination parameters with respect to the second score (claims 7 and 14), receiving user input modifying one of the examination parameters (claims 8 and 18), receiving user input selecting one of the examination parameters; and displaying data related to the selected one of the examination parameters in response to the user input (claims 9 and 19), receiving user input requesting scores and examination parameters of other patients having health data similar to the patient; and displaying the other patients' scores and examination parameters in response to the user input (claim 10), displaying the report and the second report together and displaying an indication of the relative criticality of each of the examination parameters with respect to the score or the second score depending upon a user selection (claim 15), receiving user input associated with a conclusion regarding next steps for the patient; and storing the report including the conclusion, wherein the conclusion includes a referral to a specialist or another health care professional (claim 21), receiving user input requesting reports of other patients having scores similar to the patient; and displaying the reports of the other patients in response to the user input (claim 22), receiving user input selecting one of the reports of the other patients; and displaying a detailed report of the selected one of the reports of the other patients after receiving additional user confirmation indicating that the detailed report includes the information identifying the patient associated with the selected one of the reports of the other patients (claim 23), displaying an indicator adjacent to the modified one of the examination parameters to identify that the modified one of the examination parameters has been modified by the user; and displaying the updated score in a manner that indicates that the updated score is based on one or more examination parameter values modified by the user (claim 28), displaying a reset button adjacent to the modified one of the examination parameters (claim 29), and receiving user input selecting one of the reports of the other patients; and displaying a detailed report of the selected one of the reports of the other patients after receiving additional user confirmation indicating that the detailed report includes information identifying the patient associated with the selected one of the reports of the other patients (claim 30). However, these additional elements amount to mere data gathering, displaying information, or insignificant application (i.e., insignificant extra-solution activities). As such, these additional elements, when considered individually or in combination with the prior devices, do not integrate the abstract idea into a practical application or amount to significantly more than the abstract idea.
Thus, as the dependent claims remain directed to a judicial exception, and as the additional elements of the claims do not amount to significantly more, the dependent claims are not patent eligible.
Therefore, the claims here fail to contain any additional element(s) or combination of additional elements that can be considered as significantly more and the claims are rejected under 35 U.S.C. 101 for lacking eligible subject matter.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (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, 5-7, 11-14, 17-20, and 24-27 are rejected under 35 U.S.C. 103 as being unpatentable over Bhuiyan (U.S. Patent 12170147 B1), referred to hereinafter as Bhuiyan, in view of Zhou (U.S. Patent Publication 2011/0190657A1), referred to hereinafter as Zhou, and Leung (U.S. Patent Publication 2024/0074657A1), referred to hereinafter as Leung.
Regarding claim 1, Bhuiyan teaches a method for providing a score associated with a patient, the method comprising (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”):
calculating a score associated with the patient using an algorithm to analyze the received health data, the score being indicative of a likelihood of one or more target diseases (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”, Bhuiyan, Col. 10., lines 47-65, “FIG. 7 depicts an example glaucoma screening system 700. This glaucoma screening system 200 is a solution made by iHealthscreen™. The system 700 may be used to screen individuals at risk of developing glaucoma and/or individuals at risk of having the disease worsening within them. System 700 uses multiple indications of glaucoma such as cup-disc ratio (binary cup-disc ratio as glaucoma vs. non-glaucoma and 3-class cup-disc ratio for glaucoma and non-glaucoma probability score), rim-to-disc ratio, peripapillary atrophy, disc hemorrhage, blood vessel structure/fractal dimension, nasalness of blood vessels, retinal entire image (for binary glaucoma), to generate risk scores (FIG. 8). A probability score of having glaucoma based on each of the parameters or phenotypes is generated, and based on these scores, the best combinations of features and probabilities are selected (and probabilities were combined as a vector) through the LMT (LMT equation's produced threshold value 80% or above, established in diagnostics field of research) is considered a glaucoma subject.”, Bhuiyan, Col. 7, lines 5-22, “In various embodiments described herein, machine learning model(s), such as a neural network that may be using a deep learning architecture, may be pre-trained with a dataset of eye fundus. In particular, the neural network may be pre-trained to detect one or more abnormalities in fundus images, for example, using datasets available to perform such pre-training. For example, this pre-training may include the use of general technique 100 for the development and/or pre-training of machine learning model(s) for use in the glaucoma domain. As one example, a deep machine learning architecture named “EfficientNet B5”, was pre-trained with the “ImageNet” dataset. This dataset was implemented to train a neural network to detect the disc hemorrhages in fundus images. This dataset included 150 images with disc hemorrhages and 650 normal or without disc hemorrhages that were used to train and test the machine learning model.”).
Bhuiyan fails to explicitly teach receiving health data associated with a patient, the health data including examination parameters; displaying a report comprising an indication of a relative criticality for each of the examination parameters; wherein: for each examination parameter, the relative criticality is calculated based on an extent to which each respective examination parameter deviates from at least one respective reference standard associated with the one or more target diseases; and the indication is a discrete indication generated by associating the calculated relative criticality to one of a plurality of discrete ranges defined by the at least one respective reference standard.
Zhou teaches receiving health data associated with a patient, the health data including examination parameters (Zhou [0051] “For each modality, clinicians are required to review multiple aspects of a report. For example, while interpreting a single HFA test report, a clinician must review test reliability data, rule out measurement artifacts (droopy lids, cataract, correction lens artifacts, and learning effects, etc.), and then make diagnostic assessment following, for example, a set of guidelines for number of parameters including Glaucoma Hemifield Test (GHT), Corrected Pattern Standard Deviation (CPSD), and pattern deviation plot (D R Anderson Automated Static Perimetry St. Louis: Mosby-Year Book 1992). Similarly, interpreting a single GDx RNFL test report requires a clinician to review image quality information, rule out measurement artifacts (such as atypical scans, saturated area caused by peripapillary atrophy, etc.), and then make a diagnostic assessment based on reviewing a number of global and local parameters including summary parameters (temporal-superior-nasal-inferior-temporal (TSNIT) average, Superior average, Inferior average, etc.), machine learning classifier (NFI) result, RNFL TSNIT plot, and RNFL image deviation map.”); and
displaying a report comprising an indication of a relative criticality for each of the examination parameters ((Zhou [0128] “After performing the analysis, it is desirable to have an integrated report to simplify interpretation and to improve workflow. The report should include glaucoma test data and treatment data, provide a summary of glaucoma detection (FIGS. 7-8), and provide trend plots of stage index and treatment data to facilitate efficient assessment of individual risk for vision impairment and treatment efficacy.”, Zhou [0113] “As illustrated in FIG. 10 a, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters, or their corresponding probability values derived from the combined measurement using conversion functions. This approach may require establishment of normative limits for the combined test, and may not utilize all of the existing analyses in individual modalities.”, Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”, and Zhou [0115] “The output of the machine learning classifier could be a classification with three categories (e.g. Within Normal Limits, Borderline, and Outside Normal Limits) or a continuous index (e.g., value ranging from 0 to 100). A threshold may be set for the index according to the desired balance of specificity and sensitivity. Presumably the thresholded index has improved sensitivity at a given specificity, or improved specificity at a given sensitivity. Therefore for an individual, it can be considered as confirming (or refuting) the individual test, if a previously undetected case is now detected, or a previous false positive is now correctly identified as not having the pathology.”).
Leung teaches wherein: for each examination parameter, the relative criticality is calculated based on an extent to which each respective examination parameter deviates from at least one respective reference standard associated with the one or more target diseases (Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”); and
the indication is a discrete indication generated by associating the calculated relative criticality to one of a plurality of discrete ranges defined by the at least one respective reference standard (Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”).
Therefore, it would have been obvious to a person having ordinary skill in the art (PHOSITA) before the effective filing date of the invention to combine the glaucoma risk scoring and machine learning disease likelihood determination of Bhuiyan with the ophthalmic examination parameter analysis and integrated glaucoma reporting of Zhou, as modified by the percentile-based normative reference standards and threshold classifications of Leung, in order to provide a glaucoma assessment system capable of receiving patient examination parameters, generating disease likelihood scores, and presenting clinically meaningful abnormality indications based on deviations from normative ocular reference values. A PHOSITA would have recognized that combining these teachings would predictably improve glaucoma screening interpretation and clinical usability by allowing examination parameters to be evaluated against population normative standards and visually categorized according to severity or abnormality ranges.
Specifically, Bhuiyan teaches generating glaucoma risk scores using machine learning algorithms trained on multiple glaucoma indications and combining those scores to determine a likelihood of glaucoma. Zhou teaches reviewing and reporting multiple ophthalmic examination parameters, including RNFL thickness values, classifier outputs, and deviations from age-corrected normal values, as well as generating integrated glaucoma assessment reports including classifications such as “Within Normal Limits,” “Borderline,” and “Outside Normal Limits.” Leung further teaches deriving percentile threshold values from age and group normative distributions, comparing measured ocular parameters against those normative thresholds, and visually identifying abnormalities when measured values fall below defined percentile ranges. A PHOSITA would have understood that Leung’s percentile normative thresholding techniques could be readily incorporated into the glaucoma scoring and reporting systems of Bhuiyan and Zhou to improve identification and visualization of clinically significant abnormalities.
A PHOSITA also would have been motivated to combine these references because ophthalmic diagnostic systems routinely rely on normative databases, percentile thresholds, and categorical abnormality indicators to improve interpretation consistency and facilitate clinical decision making. Applying percentile normative thresholds, such as those taught by Leung, to the glaucoma scoring framework of Bhuiyan and the integrated reporting framework of Zhou would have been a predictable use of known ophthalmic diagnostic techniques to yield the predictable result of providing discrete abnormality indications corresponding to the extent that examination parameters deviate from reference standards associated with glaucoma or related ocular diseases.
Further, the combination applies known techniques in their expected manner. Bhuiyan already teaches algorithmic glaucoma scoring from examination features, Zhou teaches integrated ophthalmic reporting and parameter classification, and Leung teaches normative percentile comparison and threshold abnormality determination. Combining these teachings would have involved only the routine application of known diagnostic analysis and reporting methods using known ophthalmic parameters and normative distributions, yielding no unexpected result beyond improved clinical presentation and interpretation of glaucoma risk information.
Regarding claim 2, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach wherein the indication of the relative criticality of each of the examination parameters is determined based on percentiles calculated from a normative database, wherein the normative database comprises normal eye data obtained from a selected population diagnosed as free of ocular disease, and provides population-based reference values from which percentiles are calculated (Leung [0042] In some embodiments, the threshold ACA value of an angle-location is determined from a specific cut-off percentile of the age-related distribution of that particular angle-location. The age-related distribution can be determined from a cross-sectional dataset collected from a normal healthy cohort. For example, measurements of ACA values are collected from a cohort of 300 healthy individuals, which is composed of 6 age groups: 18-30 years, 31-40 years, 41-50 years, 51-60 years, 61-70 years, and >70 years. There are 50 subjects recruited for each age group, forming an age group-specific normative distribution for each ACA parameter at each angle-location.”, Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”).
Therefore, it would have been obvious to a person having ordinary skill in the art (PHOSITA) before the effective filing date of the invention to determine the indication of relative criticality of examination parameters based on percentiles calculated from a normative database comprising normal eye data obtained from a healthy population, as taught by Leung, because Leung expressly teaches generating age and group normative distributions from cohorts of healthy individuals, deriving percentile threshold values from those normative distributions, and comparing measured ocular parameters against the percentile thresholds to determine abnormality significance. A PHOSITA would have recognized that using population normative percentile distributions to classify the significance of ocular examination parameters is a known and predictable ophthalmic diagnostic technique that improves consistency and interpretability of glaucoma assessments by allowing patient measurements to be evaluated relative to reference values derived from healthy populations.
Regarding claim 3, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach wherein the at least one respective reference standard is used to determine whether an individual value of the examination parameter is within a normal range or an abnormal range (Leung [0042] In some embodiments, the threshold ACA value of an angle-location is determined from a specific cut-off percentile of the age-related distribution of that particular angle-location. The age-related distribution can be determined from a cross-sectional dataset collected from a normal healthy cohort. For example, measurements of ACA values are collected from a cohort of 300 healthy individuals, which is composed of 6 age groups: 18-30 years, 31-40 years, 41-50 years, 51-60 years, 61-70 years, and >70 years. There are 50 subjects recruited for each age group, forming an age group-specific normative distribution for each ACA parameter at each angle-location.”, Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to use a reference standard to determine whether an individual examination parameter value falls within a normal range or an abnormal range, because Leung teaches generating normative distributions from healthy patient populations and deriving percentile threshold values from those normative datasets for evaluating ocular parameters. Specifically, Leung teaches comparing measured ACA parameter values against age and group percentile thresholds, where values falling below the threshold are identified as abnormal and values not below the threshold are considered normal. A PHOSITA would have recognized that applying normative reference standards and percentile thresholds to classify examination parameter values as normal or abnormal is a well established diagnostic technique in ophthalmologic analysis and would have predictably improved consistency and objectivity in evaluating disease-related examination parameters.
Regarding claim 5, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach wherein the indication is provided by a color corresponding to one of the plurality of discrete ranges (Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”,Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to provide the indication using a color corresponding to one of a plurality of discrete ranges, because Leung teaches determining abnormality classifications based on percentile threshold ranges derived from age and group normative distributions and further teaches visually highlighting examination parameter regions using color when measured values fall below the threshold percentile. Specifically, Leung teaches that measured AOD500 values below the age and group percentile cutoff are highlighted in grey color to indicate abnormal angle conditions. A PHOSITA would have recognized that associating colors with threshold diagnostic categories or ranges is a well-known visualization technique for improving the clarity and usability of clinical interpretation, and would have been motivated to apply color coded indications to different relative criticality ranges in order to enhance physician review and diagnostic assessment of examination parameters.
Regarding claim 6, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach further comprising: selecting the algorithm from a plurality of candidate algorithms based on the examination parameters associated with the algorithm (Zhou [0072] “The preprocessing steps associated with preparing the target vectors (T) are simple and must be consistent with the input vector configuration. It starts with the sensitivity values of the 52 test locations and followed by 3 options of pre-processing: the 52-points may be divided to superior and inferior hemi-fields to train two separate models, if the same step is applied to the input vectors; the target vectors may be scaled to the range of [−1 1], if the same step is applied to the input vectors; the target vectors is converted from log scale to linear scale, if the input vectors are in linear scale. Zhou [0073] “Multiple conversion models (with different preprocessing configurations) based on GDxECC and HFA combination were developed and tested, and the preferred model identified based on converted ECC normative database distribution and results of the testing data set. Four models selected for their attributes and performance: Model 1—0—1—1—3 (smoothing over blood vessel, full field, linear scale, scaling, and spread of 3); Model 1—0—1—1—2.5 (smoothing over blood vessel, full field, linear scale, scaling, and spread of 2.5); Model 1—0—2—0—50 (smoothing over blood vessel, full field, log scale, no scaling, and spread of 50); Model 1—1—1—1—2 (smoothing over blood vessel, hemi field, linear scale, scaling, and spread of 2).”, and
Bhuiyan, Col. 2, lines 31-35, “Each of the risk scores may be generated based on an output of a separate machine learning model, such as a deep learning model, a deep neural network, or the like, with the multiple indications as input to each of the the deep learning models.”, and Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to select an algorithm from a plurality of candidate algorithms based on the examination parameters associated with the algorithm, because Bhuiyan teaches generating glaucoma risk scores using multiple machine learning models trained on different glaucoma indications and phenotypes, including different examination parameter inputs and feature sets, while Zhou teaches developing, testing, and selecting among multiple conversion models having different preprocessing configurations and parameter arrangements, including full field versus hemi-field processing, logarithmic versus linear scaling, and different spread configurations. A PHOSITA would have understood that different examination parameter sets and preprocessing conditions are better suited to different analytical models, and therefore would have been motivated to select among multiple candidate algorithms according to the associated examination parameters in order to improve diagnostic accuracy, sensitivity, and overall glaucoma detection performance. This selection and optimization of algorithms based on examination parameter characteristics represents the predictable use of prior art elements according to their established functions.
Regarding claim 7, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach further comprising: calculating a second score associated with the patient using a second algorithm to analyze the received health data in response to determining that the examination parameters are within a desired examination parameter range and the score is out of a desired score range (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”, Bhuiyan, Col. 10., lines 47-65, “FIG. 7 depicts an example glaucoma screening system 700. This glaucoma screening system 200 is a solution made by iHealthscreen™. The system 700 may be used to screen individuals at risk of developing glaucoma and/or individuals at risk of having the disease worsening within them. System 700 uses multiple indications of glaucoma such as cup-disc ratio (binary cup-disc ratio as glaucoma vs. non-glaucoma and 3-class cup-disc ratio for glaucoma and non-glaucoma probability score), rim-to-disc ratio, peripapillary atrophy, disc hemorrhage, blood vessel structure/fractal dimension, nasalness of blood vessels, retinal entire image (for binary glaucoma), to generate risk scores (FIG. 8). A probability score of having glaucoma based on each of the parameters or phenotypes is generated, and based on these scores, the best combinations of features and probabilities are selected (and probabilities were combined as a vector) through the LMT (LMT equation's produced threshold value 80% or above, established in diagnostics field of research) is considered a glaucoma subject.”;
Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”, and Zhou [0115] “The output of the machine learning classifier could be a classification with three categories (e.g. Within Normal Limits, Borderline, and Outside Normal Limits) or a continuous index (e.g., value ranging from 0 to 100). A threshold may be set for the index according to the desired balance of specificity and sensitivity. Presumably the thresholded index has improved sensitivity at a given specificity, or improved specificity at a given sensitivity. Therefore for an individual, it can be considered as confirming (or refuting) the individual test, if a previously undetected case is now detected, or a previous false positive is now correctly identified as not having the pathology.”); and
displaying a second report comprising the second score and an indication of the relative criticality of each of the examination parameters with respect to the second score (Zhou [0128] “After performing the analysis, it is desirable to have an integrated report to simplify interpretation and to improve workflow. The report should include glaucoma test data and treatment data, provide a summary of glaucoma detection (FIGS. 7-8), and provide trend plots of stage index and treatment data to facilitate efficient assessment of individual risk for vision impairment and treatment efficacy.”, Zhou [0113] “As illustrated in FIG. 10 a, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters, or their corresponding probability values derived from the combined measurement using conversion functions. This approach may require establishment of normative limits for the combined test, and may not utilize all of the existing analyses in individual modalities.”, Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”, and Zhou [0115] “The output of the machine learning classifier could be a classification with three categories (e.g. Within Normal Limits, Borderline, and Outside Normal Limits) or a continuous index (e.g., value ranging from 0 to 100). A threshold may be set for the index according to the desired balance of specificity and sensitivity. Presumably the thresholded index has improved sensitivity at a given specificity, or improved specificity at a given sensitivity. Therefore for an individual, it can be considered as confirming (or refuting) the individual test, if a previously undetected case is now detected, or a previous false positive is now correctly identified as not having the pathology.”; and
Bhuiyan, Col. 2, lines 31-42 “Each of the risk scores may be generated based on an output of a separate machine learning model, such as a deep learning model, a deep neural network, or the like, with the multiple indications as input to each of the the deep learning models. A classification model, such as a logistic model tree (LMT) may be used to combine each of the risk scores to produce a probability/likelihood risk score of glaucoma. This technique may achieve a substantially high accuracy for the detection of glaucoma. In aspects, 8 probability scores are combined as a featured vector, which the LMT classifies based on the samples of glaucoma and non-glaucoma.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to calculate a second score associated with a patient using a second algorithm when examination parameters are within a desired range but an initial score falls outside a desired score range, because Bhuiyan teaches generating multiple glaucoma risk scores using multiple machine learning models and combining probability outputs through an additional classification model to improve glaucoma detection accuracy. Bhuiyan further teaches using separate machine learning models and probability analyses derived from different glaucoma indications and combining those outputs through a logistic model tree (LMT) classifier to produce refined likelihood determinations. Zhou teaches analyzing examination parameters relative to normative values and threshold ranges, including classifications such as “Within Normal Limits,” “Borderline,” and “Outside Normal Limits,” and further teaches that threshold classifications may confirm or refute prior test results in situations where initial results may indicate false positives or previously undetected pathology. A PHOSITA would have been motivated to apply a secondary scoring algorithm when examination parameters appear within acceptable ranges but an initial score remains abnormal in order to improve diagnostic reliability, reduce false positives and false negatives, provide additional confirmation analysis, and increase confidence in clinical decision making. It also would have been obvious to display a second report including the second score and relative criticality information for the examination parameters, as Zhou teaches integrated reporting of glaucoma test data, treatment data, stage indices, and parameter-based analyses to facilitate efficient assessment of patient risk and improve clinical workflow. The combination merely applies known machine learning scoring, threshold analysis, and integrated reporting techniques according to their established functions to yield predictable diagnostic and workflow improvements.
Claim 11 is analogous to claim 1, thus claim 11 is similarly analyzed and rejected in a manner consistent with the rejection of claim 1.
Claims 12-14 are analogous to claims 5-7, thus claims 12-14 are similarly analyzed and rejected in a manner consistent with the rejection of claims 5-7.
Regarding claim 17, Bhuiyan, Zhou, and Leung teach the invention in claim 11, as discussed above, and further teach wherein the algorithm is an artificial intelligence algorithm (Bhuiyan, Col. 14. lines 13-14, “This system employs artificial intelligence, specifically machine learning, to analyze eye images.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to implement the algorithm as an artificial intelligence algorithm, because Bhuiyan teaches the use of artificial intelligence, specifically machine learning and neural network models, for analyzing eye images and generating glaucoma risk scores and classifications. Bhuiyan further teaches the use of deep learning architectures trained on ophthalmic image datasets to detect abnormalities associated with glaucoma. A PHOSITA would have recognized that utilizing artificial intelligence algorithms in the claimed scoring system would improve the automation, accuracy, and efficiency of analyzing examination parameters and predicting disease likelihood, particularly in the context of ophthalmic diagnostic systems where AI image analysis techniques were well known and commonly applied.
Claims 18-19 are analogous to claims 8-9, thus claims 18-19 are similarly analyzed and rejected in a manner consistent with the rejection of claims 8-9.
Claim 20 is analogous to claim 1, thus claim 20 is similarly analyzed and rejected in a manner consistent with the rejection of claim 1.
Regarding claim 24, Bhuiyan, Zhou, and Leung teach the invention in claim 2, as discussed above, and further teach wherein the percentiles provide an indication of where an individual value of each examination parameter falls within an overall distribution (Leung [0042] In some embodiments, the threshold ACA value of an angle-location is determined from a specific cut-off percentile of the age-related distribution of that particular angle-location. The age-related distribution can be determined from a cross-sectional dataset collected from a normal healthy cohort. For example, measurements of ACA values are collected from a cohort of 300 healthy individuals, which is composed of 6 age groups: 18-30 years, 31-40 years, 41-50 years, 51-60 years, 61-70 years, and >70 years. There are 50 subjects recruited for each age group, forming an age group-specific normative distribution for each ACA parameter at each angle-location.”, Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to determine the indication of relative criticality based on percentiles calculated from a normative database, wherein the percentiles indicate where an individual examination parameter value falls within an overall distribution, because Leung teaches generating age group normative distributions from healthy patient populations and deriving percentile thresholds from those distributions for individual examination parameters. Leung further teaches comparing measured patient values against percentile positions, including the 5th percentile, median, and 95th percentile, to determine abnormality and extent of angle closure abnormalities. A PHOSITA would have recognized that percentile comparison against normative population data is a well-known statistical technique for determining how a patient’s examination parameter relates to an overall distribution and for identifying whether the parameter falls within normal or abnormal ranges, thereby providing predictable and clinically useful evaluation of disease risk and severity.
Regarding claim 25, Bhuiyan, Zhou, and Leung teach the invention in claim 24, as discussed above, and further teach wherein the percentiles are applied to determine whether the individual value falls within a normal range or exhibits an abnormal value warranting further clinical attention (Leung [0042] In some embodiments, the threshold ACA value of an angle-location is determined from a specific cut-off percentile of the age-related distribution of that particular angle-location. The age-related distribution can be determined from a cross-sectional dataset collected from a normal healthy cohort. For example, measurements of ACA values are collected from a cohort of 300 healthy individuals, which is composed of 6 age groups: 18-30 years, 31-40 years, 41-50 years, 51-60 years, 61-70 years, and >70 years. There are 50 subjects recruited for each age group, forming an age group-specific normative distribution for each ACA parameter at each angle-location.”, Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to apply percentiles from a normative database to determine whether an individual examination parameter value falls within a normal range or exhibits an abnormal value warranting further clinical attention, because Leung teaches generating age group normative distributions from healthy patient populations and deriving percentile thresholds from those distributions for ACA parameters. Leung further teaches using percentile thresholds, such as the 5th percentile, to determine whether a measured value indicates a narrow angle abnormality and visually highlighting measurements falling below the threshold. A PHOSITA would have recognized that comparing individual patient measurements against percentile-based normative thresholds is a well known diagnostic technique for distinguishing normal from abnormal values and for identifying patients requiring further clinical evaluation, thereby providing predictable and clinically useful assessment of disease abnormalities.
Regarding claim 26, Bhuiyan, Zhou, and Leung teach the invention in claim 2, as discussed above, and further teach wherein the selected population is selected based on patient age (Leung [0042] “In some embodiments, the threshold ACA value of an angle-location is determined from a specific cut-off percentile of the age-related distribution of that particular angle-location. The age-related distribution can be determined from a cross-sectional dataset collected from a normal healthy cohort. For example, measurements of ACA values are collected from a cohort of 300 healthy individuals, which is composed of 6 age groups: 18-30 years, 31-40 years, 41-50 years, 51-60 years, 61-70 years, and >70 years. There are 50 subjects recruited for each age group, forming an age group-specific normative distribution for each ACA parameter at each angle-location.”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to select the normative population based on patient age, because Leung teaches generating age normative distributions using separate age groups of healthy individuals and determining percentile thresholds from those age populations. Leung further teaches forming age and group normative distributions for ACA parameters at different angle locations. A PHOSITA would have recognized that selecting a reference population based on patient age improves the accuracy and reliability of diagnostic comparison because many ophthalmic measurements vary with age, and age adjusted normative databases were well known in the art for distinguishing normal anatomical variations from disease abnormalities.
Regarding claim 27, Bhuiyan, Zhou, and Leung teach the invention in claim 5, as discussed above, and further teach wherein relative criticality values in the 0-1% range are shown in red, relative criticality values in the 1-5% range are shown in yellow, relative criticality values in the 5-95% range are shown in green, and relative criticality values in the 95-100% range are shown in orange (Leung [0042] In some embodiments, the threshold ACA value of an angle-location is determined from a specific cut-off percentile of the age-related distribution of that particular angle-location. The age-related distribution can be determined from a cross-sectional dataset collected from a normal healthy cohort. For example, measurements of ACA values are collected from a cohort of 300 healthy individuals, which is composed of 6 age groups: 18-30 years, 31-40 years, 41-50 years, 51-60 years, 61-70 years, and >70 years. There are 50 subjects recruited for each age group, forming an age group-specific normative distribution for each ACA parameter at each angle-location.”, Leung [0043] “In further embodiments, an age group-specific k-th percentile of a selected ACA parameter is derived from the age group-specific normative distribution for each angle-location. This k-th percentile is used to define the threshold value of the selected ACA parameter at the angle-location where the ACA is determined as narrow angle when the measured value of the selected ACA parameter is below this threshold. For example, in certain exemplary and non-limiting embodiments, the k-th percentile can be (but is not limited to) the 5th or the 10th percentile of the distribution as shown in FIGS. 6A-6E. In some embodiments, the k-th percentile used to determine the threshold ACA value at one angle-location is not necessarily the same as the percentile used to determine the threshold ACA value at other angle-locations.”, and Leung [0044] “As an illustrative example, the AOD500 measurements at 36 angle-locations from an eye detected with gonioscopic angle closure are plotted as a polar plot in FIG. 9, showing extensive ACA abnormalities. The radial axis represents AOD500 in the unit of mm and each sector represents one of the 36 measured angle-locations. Each dark solid square symbol denotes an AOD500 value measured at the respective angle-location from the example eye. On the other hand, the normative AOD500 values from the age-group of the eye are plotted in the same polar plot with each grey star symbol denoting the 95th percentile of the age group-specific normative distribution for an angle-location, each grey cross symbol denoting the median, and each grey triangle denoting the 5th percentile. FIG. 10 illustrates a magnified version of the polar plot in FIG. 9, showing only the measured AOD500 values and the age group-specific 5th percentiles of the measured angle-locations. In this example, the age group-specific 5th percentile is used as the cut-off percentile to determine if the ACA of an angle-location is a narrow angle for all angle-locations. When the measured AOD500 value is below the age group-specific 5th percentile or equal to zero at an angle-location, the respective sector of the polar plot is highlighted in grey color. The extent of ACA abnormalities is calculated based on the total number of narrow/closed angles detected from the eye. On the contrary, an example demonstrating an eye without any ACA abnormality is illustrated in FIG. 11. None of the angle-locations has the measured AOD500 value below the respective age group-specific 5th percentile in this eye.” And Zhou [0085] “The two approaches are illustrated in FIGS. 5 & 6 based on converting the spatial distribution and scale of RNFL test data from OCT and/or GDx measurements to visual field sensitivity data but the opposite conversion could be taken as well. Black color indicates initial input, blue color indicates intermediate results, red color indicates outputs, dotted lines and arrows indicate alternative or optional path”).
Therefore, it would have been obvious to a person having ordinary skill in the art before the effective filing date of the invention to select the normative population based on patient age, because Leung teaches generating age normative distributions using separate age groups of healthy individuals and determining percentile thresholds from those age populations. Leung further teaches forming age and group normative distributions for ACA parameters at different angle locations. A PHOSITA would have recognized that selecting a reference population based on patient age improves the accuracy and reliability of diagnostic comparison because many ophthalmic measurements vary with age, and age normative databases were well known in the art for distinguishing normal anatomical variations from disease-related abnormalities.
Claims 8-10, 15-16, 21-23, and 28-30 are rejected under 35 U.S.C. 103 as being unpatentable over Bhuiyan (U.S. Patent 12170147 B1), referred to hereinafter as Bhuiyan, in view of Zhou (U.S. Patent Publication 2011/0190657A1), referred to hereinafter as Zhou, and Leung (U.S. Patent Publication 2024/0074657A1), referred to hereinafter as Leung, and further in view of Linthicum (U.S. Patent Publication 2010/0131293A1), referred to hereinafter as Linthicum.
Regarding claim 8, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach further comprising: one of the examination parameters; and updating the score (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.” And
Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”).
Bhuiyan, Zhou, and Leung fails to explicitly teach receiving user input modifying one of the parameters; and in response to the user input.
Linthicum teaches receiving user input modifying one of the parameters; and in response to the user input (Linthicum [0164] “Rather than focus on pre-determined workflows, the active listener provides a user with additional information helpful to the user in certain situations where there is no known workflow or protocol. Based on historical data and/or other input, the system displays additional information and/or functionality to the user that is relevant to the user to make an informed decision. In the background of an application and/or interface, for example, the active listener can monitor activity of data elements on a displayed interface. When these data elements reach a certain threshold, the active listener places additional information on the displayed interface to help the user make an informed decision. Alternatively or additionally, the active listener can detect when the user makes a change to an application (e.g., by dragging and dropping a data element from on widget to another widget, by conducting a search, by changing a diagnosis, etc.). By combining a context of user interaction with displayed user interface content, relevant information and/or functionality can be provided to a user, for example.”).
Therefore, it would have been obvious to a PHOSITA before the effective filing date of the invention to allow a user to modify one or more examination parameters and automatically update the resulting score in response to the modification, because Linthicum teaches an interactive clinical interface capable of detecting user changes to displayed clinical data and providing responsive functionality based on those changes, including when a user changes diagnostic information or other application data. Bhuiyan teaches generating glaucoma likelihood scores using machine learning models trained on multiple examination indications and parameters, while Zhou teaches that machine learning classifier inputs may include global, regional, and local examination parameters, including probability values and deviations from normative limits. A PHOSITA would have been motivated to combine these teachings so that when a clinician modifies examination parameter values through the interactive interface of Linthicum, the machine learning scoring system of Bhuiyan and Zhou would recalculate and update the patient risk score in order to improve clinical usability, support real-time diagnostic assessment, and permit interactive evaluation of patient examination data.
Regarding claim 9, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach the examination parameters (Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”).
Bhuiyan, Zhou, and Leung fails to explicitly teach further comprising: receiving user input selecting one of the parameters; and displaying data related to the selected one of the parameters in response to the user input.
Linthicum teaches further comprising: receiving user input selecting one of the parameters; and displaying data related to the selected one of the parameters in response to the user input (Linthicum [0083] “As shown, for example, in FIG. 2, a user can manipulate a cursor 280 to select a widget and position the widget at a location 285. Thus, a user can select widgets for display and then arrange their layout in the widget display area 215 of the UI 200. Alternatively and/or in addition, the user can reposition widgets in the widget display area 215 to modify the UI 200 layout. For example, using the cursor 280, the user can place the reason for visit widget 260 in a certain spot 285 on the widget display area 215.”, Linthicum [0085] “Certain embodiments allow healthcare information systems to find and make use of relevant information across a timeline of patient care. For example, a search-driven, role-based interface allows an end user to access, input, and search medical information seamlessly across a healthcare network. An adaptive user interface provides capabilities through a work-centered interface tailored to individual needs and responsive to changes in a work domain, for example. Semantic technology can be leveraged to model domain concepts, user roles and tasks, and information relationships. The semantic models enable applications to find, organize and present information to users more effectively based on contextual information about the user and task. Components forming a framework for query and result generation include user interface frameworks/components for building applications; server components to enable more efficient retrieval, aggregation, and composition of information based on semantic information and context; and data access mechanisms for connecting to heterogeneous information sources in a distributed environment.”, Linthicum [0164] “Rather than focus on pre-determined workflows, the active listener provides a user with additional information helpful to the user in certain situations where there is no known workflow or protocol. Based on historical data and/or other input, the system displays additional information and/or functionality to the user that is relevant to the user to make an informed decision. In the background of an application and/or interface, for example, the active listener can monitor activity of data elements on a displayed interface. When these data elements reach a certain threshold, the active listener places additional information on the displayed interface to help the user make an informed decision. Alternatively or additionally, the active listener can detect when the user makes a change to an application (e.g., by dragging and dropping a data element from on widget to another widget, by conducting a search, by changing a diagnosis, etc.). By combining a context of user interaction with displayed user interface content, relevant information and/or functionality can be provided to a user, for example.”).
Therefore, it would have been obvious to a PHOSITA before the effective filing date of the invention to allow a user to select one of the examination parameters and display data related to the selected examination parameter in response to the user input, because Linthicum teaches interactive healthcare user interfaces in which users may select and arrange widgets, search and retrieve patient information, and receive contextually relevant information and functionality based on user interaction with displayed data elements. Linthicum further teaches adaptive interfaces that organize and present medical information responsive to user actions, including selecting, modifying, or interacting with displayed clinical data. Zhou teaches that machine learning classifiers for glaucoma analysis utilize multiple examination parameters, including global, regional, and local parameters obtained from diagnostic modalities and compared against normative values. A PHOSITA would have been motivated to combine the interactive user interface functionality of Linthicum with the examination parameter diagnostic analysis of Zhou so that selection of a particular examination parameter would cause related diagnostic data associated with that parameter to be displayed, thereby improving clinician workflow, facilitating interpretation of patient specific examination data, and enabling more efficient review of glaucoma diagnostic information.
Regarding claim 10, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above, and further teach further comprising: requesting scores and examination parameters; and patients' scores and examination parameters (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”, and
Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”).
Bhuiyan, Zhou, and Leung fails to explicitly teach receiving user input requesting parameters of other patients having health data similar to the patient; and displaying the other patients' parameters in response to the user input.
Linthicum teaches receiving user input requesting parameters of other patients having health data similar to the patient; and displaying the other patients' parameters in response to the user input ((Linthicum [0085] “Certain embodiments allow healthcare information systems to find and make use of relevant information across a timeline of patient care. For example, a search-driven, role-based interface allows an end user to access, input, and search medical information seamlessly across a healthcare network. An adaptive user interface provides capabilities through a work-centered interface tailored to individual needs and responsive to changes in a work domain, for example. Semantic technology can be leveraged to model domain concepts, user roles and tasks, and information relationships. The semantic models enable applications to find, organize and present information to users more effectively based on contextual information about the user and task. Components forming a framework for query and result generation include user interface frameworks/components for building applications; server components to enable more efficient retrieval, aggregation, and composition of information based on semantic information and context; and data access mechanisms for connecting to heterogeneous information sources in a distributed environment.”, Linthicum [0140] “Additionally, FIG. 11 depicts an example of a longitudinal health record 1100 including three-dimensional (“3D”) spectrum representation 1110 of patient information. The spectrum 1110 can be used to represent patient data for one patient and/or for multiple patients, for example. The 3D navigable representation 1110 of patient clinical information uses a graphical representation akin to an electromagnetic radio spectrum to graphically represent different types of patient information. A “services” view 1110 shows a range of clinical information including patient vitals, laboratory results, diagnoses, etc. A “projects” view 1120 delineates different encounters and/or dates during which the data of the services view 1110 was obtained (e.g., a patient clinic visit where a physical examination was conducted and blood was drawn for testing).”).
Therefore, it would have been obvious to a PHOSITA before the effective filing date of the invention to receive user input requesting scores and examination parameters of other patients having health data similar to a current patient and to display those other patients’ scores and examination parameters in response to the request, because Linthicum teaches search driven and context aware healthcare interfaces capable of retrieving, aggregating, organizing, and presenting medical information across healthcare systems based on semantic relationships, contextual information, and user queries. Linthicum further teaches graphical representations capable of displaying clinical information for multiple patients, including patient vitals, laboratory results, diagnoses, and longitudinal patient data. Bhuiyan teaches generating glaucoma risk scores from multiple examination parameters using machine learning models, while Zhou teaches that such machine learning analysis may utilize global, regional, and local examination parameters obtained from diagnostic modalities and compared against normative values. A PHOSITA would have been motivated to combine the contextual retrieval and multi-patient display functionality of Linthicum with the examination parameter scoring systems of Bhuiyan and Zhou so that clinicians could retrieve and review scores and examination parameters from patients having similar health characteristics in order to improve diagnostic interpretation, facilitate comparative clinical analysis, and support more informed treatment decision making.
Regarding claim 15, Bhuiyan, Zhou, and Leung teach the invention in claim 14, as discussed above, and further teach the relative criticality of each of the examination parameters with respect to the score or the second score (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”, and
Zhou [0128] “After performing the analysis, it is desirable to have an integrated report to simplify interpretation and to improve workflow. The report should include glaucoma test data and treatment data, provide a summary of glaucoma detection (FIGS. 7-8), and provide trend plots of stage index and treatment data to facilitate efficient assessment of individual risk for vision impairment and treatment efficacy.”, Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”, and Zhou [0115] “The output of the machine learning classifier could be a classification with three categories (e.g. Within Normal Limits, Borderline, and Outside Normal Limits) or a continuous index (e.g., value ranging from 0 to 100). A threshold may be set for the index according to the desired balance of specificity and sensitivity. Presumably the thresholded index has improved sensitivity at a given specificity, or improved specificity at a given sensitivity. Therefore for an individual, it can be considered as confirming (or refuting) the individual test, if a previously undetected case is now detected, or a previous false positive is now correctly identified as not having the pathology.”).
Bhuiyan, Zhou, and Leung fails to explicitly teach the operations further comprising: displaying the report and the second report together and displaying an indication; and depending upon a user selection.
Linthicum teaches the operations further comprising: displaying the report and the second report together and displaying an indication; and depending upon a user selection (Linthicum [0085] “Certain embodiments allow healthcare information systems to find and make use of relevant information across a timeline of patient care. For example, a search-driven, role-based interface allows an end user to access, input, and search medical information seamlessly across a healthcare network. An adaptive user interface provides capabilities through a work-centered interface tailored to individual needs and responsive to changes in a work domain, for example. Semantic technology can be leveraged to model domain concepts, user roles and tasks, and information relationships. The semantic models enable applications to find, organize and present information to users more effectively based on contextual information about the user and task. Components forming a framework for query and result generation include user interface frameworks/components for building applications; server components to enable more efficient retrieval, aggregation, and composition of information based on semantic information and context; and data access mechanisms for connecting to heterogeneous information sources in a distributed environment.”, Linthicum [0164] “Rather than focus on pre-determined workflows, the active listener provides a user with additional information helpful to the user in certain situations where there is no known workflow or protocol. Based on historical data and/or other input, the system displays additional information and/or functionality to the user that is relevant to the user to make an informed decision. In the background of an application and/or interface, for example, the active listener can monitor activity of data elements on a displayed interface. When these data elements reach a certain threshold, the active listener places additional information on the displayed interface to help the user make an informed decision. Alternatively or additionally, the active listener can detect when the user makes a change to an application (e.g., by dragging and dropping a data element from on widget to another widget, by conducting a search, by changing a diagnosis, etc.). By combining a context of user interaction with displayed user interface content, relevant information and/or functionality can be provided to a user, for example.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to modify the glaucoma risk assessment and reporting system of Bhuiyan with the adaptive healthcare user interface and contextual information presentation techniques of Linthicum, as well as the integrated reporting and examination parameter analysis teachings of Zhou, in order to display multiple reports together and present relative criticality information for examination parameters responsive to user selection. Bhuiyan teaches generating multiple glaucoma risk scores using machine learning models and combining the scores to produce a likelihood determination for glaucoma. Zhou further teaches integrated reports including glaucoma test data, treatment data, and examination parameter information compared against normative limits and classified according to threshold categories indicative of clinical significance. Linthicum teaches adaptive healthcare interfaces that aggregate and compose medical information from multiple sources and dynamically display relevant information and functionality in response to user interaction and contextual selections. One of ordinary skill in the art would have been motivated to combine these teachings to improve clinical workflow, facilitate comparative review of multiple reports and scores, and provide more efficient and context-sensitive presentation of examination parameter criticality to assist clinicians in diagnostic evaluation and treatment decision making, yielding predictable results consistent with the intended purposes of the combined references.
Regarding claim 16, Bhuiyan, Zhou, and Leung teach the invention in claim 14, as discussed above, and further teach the operations further comprising: consolidating the score and the second score (Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.” and Bhuiyan, Col. 2, lines 31-42 “Each of the risk scores may be generated based on an output of a separate machine learning model, such as a deep learning model, a deep neural network, or the like, with the multiple indications as input to each of the the deep learning models. A classification model, such as a logistic model tree (LMT) may be used to combine each of the risk scores to produce a probability/likelihood risk score of glaucoma. This technique may achieve a substantially high accuracy for the detection of glaucoma. In aspects, 8 probability scores are combined as a featured vector, which the LMT classifies based on the samples of glaucoma and non-glaucoma.”).
Bhuiyan, Zhou, and Leung fails to explicitly teach for display in an image.
Linthicum teaches for display in an image (Linthicum [0079] “Widget 240 provides one or more imaging studies for review by the user. The imaging studies widget 240 includes one or more images 244 along with an imaging type 246 and an evaluation 248. For example, as shown in FIG. 2, the widget 240 includes a head CT evaluated as normal and a fetal ultrasound image evaluated as normal.”).
Therefore, it would have been obvious to a PHOSITA before the effective filing date of the invention to consolidate a first score and a second score for display in an image, because Bhuiyan teaches generating multiple glaucoma-related risk scores from separate machine learning models and combining those scores using a classification model, such as a logistic model tree (LMT), to produce a consolidated glaucoma probability or likelihood score. Bhuiyan further teaches combining multiple probability scores into a feature vector for classification and diagnostic assessment. Linthicum teaches displaying imaging studies and associated evaluations within a clinical user interface, including presentation of medical images together with corresponding diagnostic information and image evaluations. A PHOSITA would have been motivated to combine the consolidated multi-score diagnostic analysis of Bhuiyan with the image display techniques of Linthicum so that multiple diagnostic scores could be displayed together in conjunction with medical imaging data in order to improve diagnostic interpretation, streamline clinician review of patient information, and provide a more integrated visual assessment of patient condition.
Regarding claim 21, Bhuiyan, Zhou, and Leung teach the invention in claim 1, as discussed above.
Bhuiyan, Zhou, and Leung fails to explicitly teach further comprising: receiving user input associated with a conclusion regarding next steps for the patient; and storing the report including the conclusion, wherein the conclusion includes a referral to a specialist or another health care professional.
Linthicum teaches further comprising: receiving user input associated with a conclusion regarding next steps for the patient; and storing the report including the conclusion, wherein the conclusion includes a referral to a specialist or another health care professional (Linthicum [0140] “Additionally, FIG. 11 depicts an example of a longitudinal health record 1100 including three-dimensional (“3D”) spectrum representation 1110 of patient information. The spectrum 1110 can be used to represent patient data for one patient and/or for multiple patients, for example. The 3D navigable representation 1110 of patient clinical information uses a graphical representation akin to an electromagnetic radio spectrum to graphically represent different types of patient information. A “services” view 1110 shows a range of clinical information including patient vitals, laboratory results, diagnoses, etc. A “projects” view 1120 delineates different encounters and/or dates during which the data of the services view 1110 was obtained (e.g., a patient clinic visit where a physical examination was conducted and blood was drawn for testing).”Linthicum [0091] “At her 34-week appointment, however, Patricia's obstetrician/gynecologist becomes somewhat concerned at her blood pressure, which is high compared to previous readings, at 145/95. Dr. Amanda Miller orders an electrocardiogram (“EKG”) and a urinalysis (“UA”) test. Although Patricia's EKG shows a normal sinus rhythm, her UA comes back with trace amounts of Albumin, suggestive of pre-eclampsia. Dr. Miller asks Patricia to set up her next appointment for one week from today to monitor her blood pressure and kidney function.” Linthicum [0092] “The following week, Patricia's blood pressure is higher than the previous value (150/98) and Dr. Miller orders another urinalysis. The UA comes back positive again, but at about the same level as before. Dr. Miller feels it's prudent to continue the weekly visits until her blood pressure comes down to normal levels. She also mentions to Patricia that one warning sign of eclampsia is a sudden, severe headache, and, if she experiences one, she should go directly to the Emergency Department for care.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to modify the glaucoma assessment and reporting system to receive user input associated with a conclusion regarding next steps for a patient and to store the report including the conclusion, including referrals or recommendations for additional medical care, because Linthicum teaches maintaining longitudinal patient health records containing diagnoses, clinical encounters, laboratory results, and other patient information within an adaptive healthcare interface. Linthicum further teaches clinician-driven evaluation and follow up decision making, including ordering additional testing, scheduling continued monitoring visits, and directing a patient to seek emergency department care upon occurrence of certain symptoms. A PHOSITA would have been motivated to incorporate such clinician conclusions and follow up recommendations into stored patient reports in order to improve continuity of care, facilitate longitudinal patient management, preserve clinical decision documentation, and provide healthcare professionals with more comprehensive patient records for subsequent diagnosis and treatment, yielding predictable results consistent with the combined teachings of the references.
Regarding claim 22, Bhuiyan, Zhou, Leung, and Linthicum and teach the invention in claim 21, as discussed above, and further teach further comprising: receiving user input requesting reports of other patients having scores similar to the patient; and displaying the reports of the other patients in response to the user input, wherein the reports of the other patients include at least scores and conclusions regarding next steps, and information identifying the other patients is hidden when displaying reports of the other patients (Linthicum [0085] “Certain embodiments allow healthcare information systems to find and make use of relevant information across a timeline of patient care. For example, a search-driven, role-based interface allows an end user to access, input, and search medical information seamlessly across a healthcare network. An adaptive user interface provides capabilities through a work-centered interface tailored to individual needs and responsive to changes in a work domain, for example. Semantic technology can be leveraged to model domain concepts, user roles and tasks, and information relationships. The semantic models enable applications to find, organize and present information to users more effectively based on contextual information about the user and task. Components forming a framework for query and result generation include user interface frameworks/components for building applications; server components to enable more efficient retrieval, aggregation, and composition of information based on semantic information and context; and data access mechanisms for connecting to heterogeneous information sources in a distributed environment.”, Linthicum [0140] “Additionally, FIG. 11 depicts an example of a longitudinal health record 1100 including three-dimensional (“3D”) spectrum representation 1110 of patient information. The spectrum 1110 can be used to represent patient data for one patient and/or for multiple patients, for example. The 3D navigable representation 1110 of patient clinical information uses a graphical representation akin to an electromagnetic radio spectrum to graphically represent different types of patient information. A “services” view 1110 shows a range of clinical information including patient vitals, laboratory results, diagnoses, etc. A “projects” view 1120 delineates different encounters and/or dates during which the data of the services view 1110 was obtained (e.g., a patient clinic visit where a physical examination was conducted and blood was drawn for testing).”, and Linthicum [0130] “Certain embodiments can be used to provide an integrated solution for application execution and/or information retrieval based on rules and context sharing, for example. For example, context sharing allows information and/or configuration options/settings, for example, to be shared between system environments. Rules, for example, can be defined dynamically and/or loaded from a library to filter and/or process information generated from an information system and/or an application.”, and
Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to modify the patient assessment and reporting system to receive user requests for reports of other patients having similar health data and to display corresponding reports including scores and conclusions while filtering identifying information, because Linthicum teaches a search driven, role healthcare interface capable of retrieving, organizing, aggregating, and presenting patient information across a healthcare network using contextual and semantic relationships between medical data. Linthicum further teaches longitudinal patient health records containing clinical information for one or more patients, including diagnoses, laboratory data, and encounter information, as well as rule filtering and context sharing mechanisms for controlling information retrieval and presentation. Bhuiyan teaches generating glaucoma risk scores and likelihood determinations using machine learning models trained on patient examination data. A PHOSITA would have been motivated to combine the patient data retrieval and contextual presentation capabilities of Linthicum with the predictive scoring techniques of Bhuiyan in order to allow clinicians to review reports and scores of similar patients for comparative diagnostic evaluation and treatment planning while controlling displayed information according to system rules and contextual access requirements, thereby improving clinical decision support, workflow efficiency, and patient data management in a predictable manner.
Regarding claim 23, Bhuiyan, Zhou, Leung, and Linthicum and teach the invention in claim 22, as discussed above, and further teach further comprising: receiving user input selecting one of the reports of the other patients; and displaying a detailed report of the selected one of the reports of the other patients after receiving additional user confirmation indicating that the detailed report includes the information identifying the patient associated with the selected one of the reports of the other patients (Linthicum [0085] “Certain embodiments allow healthcare information systems to find and make use of relevant information across a timeline of patient care. For example, a search-driven, role-based interface allows an end user to access, input, and search medical information seamlessly across a healthcare network. An adaptive user interface provides capabilities through a work-centered interface tailored to individual needs and responsive to changes in a work domain, for example. Semantic technology can be leveraged to model domain concepts, user roles and tasks, and information relationships. The semantic models enable applications to find, organize and present information to users more effectively based on contextual information about the user and task. Components forming a framework for query and result generation include user interface frameworks/components for building applications; server components to enable more efficient retrieval, aggregation, and composition of information based on semantic information and context; and data access mechanisms for connecting to heterogeneous information sources in a distributed environment.”, Linthicum [0140] “Additionally, FIG. 11 depicts an example of a longitudinal health record 1100 including three-dimensional (“3D”) spectrum representation 1110 of patient information. The spectrum 1110 can be used to represent patient data for one patient and/or for multiple patients, for example. The 3D navigable representation 1110 of patient clinical information uses a graphical representation akin to an electromagnetic radio spectrum to graphically represent different types of patient information. A “services” view 1110 shows a range of clinical information including patient vitals, laboratory results, diagnoses, etc. A “projects” view 1120 delineates different encounters and/or dates during which the data of the services view 1110 was obtained (e.g., a patient clinic visit where a physical examination was conducted and blood was drawn for testing).”, and Linthicum [0130] “Certain embodiments can be used to provide an integrated solution for application execution and/or information retrieval based on rules and context sharing, for example. For example, context sharing allows information and/or configuration options/settings, for example, to be shared between system environments. Rules, for example, can be defined dynamically and/or loaded from a library to filter and/or process information generated from an information system and/or an application.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to modify the healthcare information retrieval and reporting system to receive user input selecting one of multiple patient reports and to display a more detailed report after additional confirmation when the detailed report includes identifying patient information, because Linthicum teaches a search driven, role healthcare interface that retrieves, organizes, and presents patient information based on contextual relationships and user interaction. Linthicum further teaches longitudinal patient records containing detailed patient-specific clinical information for one or more patients, including diagnoses, laboratory results, and encounter data, as well as rules-based filtering and context sharing mechanisms for controlling information retrieval and processing. A PHOSITA would have been motivated to implement confirmation or controlled access mechanisms before displaying detailed patient identifying information in order to improve patient privacy protection, maintain appropriate access control to sensitive medical records, and ensure that detailed patient data is only revealed following an intentional user action consistent with healthcare information security practices, yielding predictable results consistent with the combined teachings of the references.
Regarding claim 28, Bhuiyan, Zhou, Leung, and Linthicum and teach the invention in claim 8, as discussed above, and further teach further comprising: displaying an indicator adjacent to the modified one of the examination parameters to identify that the modified one of the examination parameters has been modified by the user; and displaying the updated score in a manner that indicates that the updated score is based on one or more examination parameter values modified by the user (Linthicum [0164] “Rather than focus on pre-determined workflows, the active listener provides a user with additional information helpful to the user in certain situations where there is no known workflow or protocol. Based on historical data and/or other input, the system displays additional information and/or functionality to the user that is relevant to the user to make an informed decision. In the background of an application and/or interface, for example, the active listener can monitor activity of data elements on a displayed interface. When these data elements reach a certain threshold, the active listener places additional information on the displayed interface to help the user make an informed decision. Alternatively or additionally, the active listener can detect when the user makes a change to an application (e.g., by dragging and dropping a data element from on widget to another widget, by conducting a search, by changing a diagnosis, etc.). By combining a context of user interaction with displayed user interface content, relevant information and/or functionality can be provided to a user, for example.” and
Bhuiyan, Col. 1, lines 34-48, “According to some embodiments of the present disclosure, methods of and computer program products for predicting and detecting the onset of glaucoma are provided. In various embodiments, a method of detecting glaucoma is provided. At least one neural network model of a plurality of neural network models may be pre-trained using a small data classifier. The plurality of neural network models may be trained based on a plurality of indications of glaucoma. A risk score associated with each of the plurality of indications may be simultaneously generated based on the trained plurality of neural network models. The risk score associated with each of the plurality of indications may be combined based on a classification model to produce a likelihood of glaucoma. A determination of whether glaucoma is present may be made based on the likelihood of glaucoma.” and
Zhou [0114] “Alternatively as shown in FIG. 10 b, the input parameters (feature set) for the machine learning classifier may consist of global, regional, and local parameters directly obtained from individual modalities in their own measurement units (e.g. sensitivity values or RNFL thickness values), in deviations from age-corrected normal values, or in probability values based on comparison with their respective normative limits.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to modify the patient assessment interface to display an indicator identifying examination parameters modified by a user and to display an updated score reflecting the modified parameter values, because Linthicum teaches an adaptive healthcare user interface capable of detecting when a user changes information within an application and dynamically providing updated information and functionality responsive to those user modifications. Zhou further teaches the use of examination parameter values, including modality measurements and deviations from normative values, as inputs to machine learning classification and scoring operations. Bhuiyan teaches generating glaucoma risk scores from examination inputs using machine learning models to produce updated likelihood determinations. A PHOSITA would have been motivated to combine these teachings so that modifications to examination parameter values within the user interface would result in recalculated and visibly updated scores, while also visually indicating modified parameters to improve user awareness, reduce clinician confusion, enhance traceability of manually adjusted values, and improve diagnostic workflow efficiency, yielding predictable results consistent with well known user interface and clinical decision support design principles.
Regarding claim 29, Bhuiyan, Zhou, and Leung teach the invention in claim 26, as discussed above.
Bhuiyan, Zhou, and Leung fails to explicitly teach further comprising: displaying a reset button adjacent to the modified one of the examination parameters; and resetting the modified one of the examination parameters to its value prior to modification in response to user input selecting the reset button.
Linthicum teaches further comprising: displaying a reset button adjacent to the modified one of the examination parameters; and resetting the modified one of the examination parameters to its value prior to modification in response to user input selecting the reset button ((Linthicum [0164] “Rather than focus on pre-determined workflows, the active listener provides a user with additional information helpful to the user in certain situations where there is no known workflow or protocol. Based on historical data and/or other input, the system displays additional information and/or functionality to the user that is relevant to the user to make an informed decision. In the background of an application and/or interface, for example, the active listener can monitor activity of data elements on a displayed interface. When these data elements reach a certain threshold, the active listener places additional information on the displayed interface to help the user make an informed decision. Alternatively or additionally, the active listener can detect when the user makes a change to an application (e.g., by dragging and dropping a data element from on widget to another widget, by conducting a search, by changing a diagnosis, etc.). By combining a context of user interaction with displayed user interface content, relevant information and/or functionality can be provided to a user, for example.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to provide a reset button adjacent to a modified examination parameter and to reset the modified parameter to its prior value in response to user selection of the reset button, because Linthicum teaches an adaptive healthcare user interface capable of detecting user changes within an application and dynamically providing relevant information and functionality responsive to those modifications. Linthicum further teaches monitoring user interactions with displayed data elements and responding to user-initiated changes within the interface. A PHOSITA would have recognized that, once examination parameters are modifiable by a user within a clinical assessment interface, providing a reset control to restore original parameter values would have been a predictable and common graphical user interface implementation to improve usability, minimize user input errors, permit efficient reversal of manual modifications, preserve accuracy of patient assessment workflows, and enhance clinician efficiency during diagnostic review and decision making.
Regarding claim 30, Bhuiyan, Zhou, Leung, and Linthicum and teach the invention in claim 10, as discussed above, and further teach further comprising: receiving user input selecting one of the reports of the other patients; and displaying a detailed report of the selected one of the reports of the other patients after receiving additional user confirmation indicating that the detailed report includes information identifying the patient associated with the selected one of the reports of the other patients (Linthicum [0085] “Certain embodiments allow healthcare information systems to find and make use of relevant information across a timeline of patient care. For example, a search-driven, role-based interface allows an end user to access, input, and search medical information seamlessly across a healthcare network. An adaptive user interface provides capabilities through a work-centered interface tailored to individual needs and responsive to changes in a work domain, for example. Semantic technology can be leveraged to model domain concepts, user roles and tasks, and information relationships. The semantic models enable applications to find, organize and present information to users more effectively based on contextual information about the user and task. Components forming a framework for query and result generation include user interface frameworks/components for building applications; server components to enable more efficient retrieval, aggregation, and composition of information based on semantic information and context; and data access mechanisms for connecting to heterogeneous information sources in a distributed environment.”, Linthicum [0140] “Additionally, FIG. 11 depicts an example of a longitudinal health record 1100 including three-dimensional (“3D”) spectrum representation 1110 of patient information. The spectrum 1110 can be used to represent patient data for one patient and/or for multiple patients, for example. The 3D navigable representation 1110 of patient clinical information uses a graphical representation akin to an electromagnetic radio spectrum to graphically represent different types of patient information. A “services” view 1110 shows a range of clinical information including patient vitals, laboratory results, diagnoses, etc. A “projects” view 1120 delineates different encounters and/or dates during which the data of the services view 1110 was obtained (e.g., a patient clinic visit where a physical examination was conducted and blood was drawn for testing).”, Linthicum [0130] “Certain embodiments can be used to provide an integrated solution for application execution and/or information retrieval based on rules and context sharing, for example. For example, context sharing allows information and/or configuration options/settings, for example, to be shared between system environments. Rules, for example, can be defined dynamically and/or loaded from a library to filter and/or process information generated from an information system and/or an application.”).
It would have been obvious to one of ordinary skill in the art at the time of the invention to allow a user to select one of multiple patient reports and display a more detailed report after receiving additional user confirmation indicating that the detailed report includes identifying patient information, because Linthicum teaches a search-driven, role healthcare interface capable of retrieving, aggregating, and presenting patient information across a healthcare network based on contextual user interaction. Linthicum further teaches displaying clinical information for one or multiple patients through navigable patient records containing vitals, laboratory results, diagnoses, and encounter data. Additionally, Linthicum teaches applying dynamically defined rules and context sharing to filter and process information presented through the interface. A PHOSITA would have been motivated to incorporate a confirmation step before displaying detailed reports containing patient identifying information in order to improve privacy protection, comply with healthcare confidentiality requirements, prevent inadvertent disclosure of sensitive patient information, and provide controlled access to detailed patient records, yielding predictable results consistent with well known healthcare information system security and user interface practices.
Response to Arguments
Applicant’s arguments and amendments, see Remarks/Amendments submitted on 02/02/2026 with respect to the rejection of the claims have been carefully considered and is addressed below.
Claim Rejections - 35 USC § 101
Applicant's arguments have been fully considered but are not persuasive. Applicant states that the claims are not directed to a mental process because the score is based on analysis of received health data rather than mental evaluation. However, the claims recite limitations including “calculating a score,” determining an “extent to which each respective examination parameter deviates from at least one respective reference standard,” and associating calculated relative criticality to “one of a plurality of discrete ranges.” Under the broadest reasonable interpretation, these limitations encompass mental observations, evaluations, judgments, and classifications that can practically be performed in the human mind or with pen and paper. For example, a user could review examination parameters, compare the examination parameters to reference standards, determine the degree of deviation from the standards, evaluate disease likelihood based on the examination parameters, and categorize the relative criticality into predefined ranges. Although the claim recites that the analysis is performed “using an algorithm” and involves health data does not remove the claim from the mental processes grouping, especially where the claims do not recite a specific technological implementation or other technical mechanism for performing the analysis.
Applicant additionally states that the claims are not directed to a mental process because the human mind cannot “display” a report and allegedly cannot perform the claimed analysis. However, the recited display limitations are output functions and are not the basis of the abstract idea determination. The abstract idea is directed to evaluating examination parameters, comparing the parameters to reference standards, determining deviations from the standards, categorizing the results into predefined ranges, and determining disease likelihood, all of which can practically be performed mentally or with pen and paper under the broadest reasonable interpretation of the claims.
Applicant's arguments under Step 2A, Prong Two are also unpersuasive. Applicant states that the claims improve computerized patient care and therefore integrate the alleged abstract idea into a practical application. However, the claims do not recite any improvement to computer functionality, image acquisition technology or machine learning technology itself. Rather, the claims use generic computer components as tools to collect patient information, analyze the information, calculate disease-risk scores, categorize the results, and display the information to users. Improvements to the abstract process of medical evaluation or clinical decision making do not constitute improvements to computer technology or another technical field. The claims are therefore more analogous to the claims found ineligible in Electric Power Group, LLC v. Alstom S.A., which were directed to collecting, analyzing, and displaying information, than to the claims in Enfish or McRO, which recited specific improvements to computer functionality.
Applicant's citation to Ex parte Desjardins is not persuasive because the present claims do not recite a specific technological improvement to computer operation or software architecture. Instead, the claims recite functional results such as calculating scores, determining relative criticality, and displaying reports using generic computing components. The specification confirms that the processor and memory are conventional computer components operating in their ordinary capacities. Therefore, when considered individually and as an ordered combination, the additional elements amount to no more than well-understood, routine, and conventional activities involving data gathering and display of results. Accordingly, the claims do not amount to significantly more than the recited abstract idea and remain ineligible under 35 U.S.C. § 101.
Claim Rejections - 35 USC § 103
Applicant’s arguments traversing the prior art rejection in the previous Office Action have been fully considered. However, those arguments are rendered moot because the present rejection under 35 U.S.C. §103 relies on a different set of prior art references (Bhuiyan, Zhou, and Leung or Bhuiyan, Zhou, Leung, and Linthicum), which teach or suggest the limitations of the claims. Accordingly, Applicant’s prior arguments are not responsive to the current grounds of rejection. The rejection of claims 1-3, and 5-30 under 35 U.S.C. §103 is therefore maintained.
Conclusion
The prior art made of record and not relied upon is considered pertinent to Applicant's disclosure.
Biedermann et al. (U.S. Patent Publication 2009/0119337 A1) teaches tools and methods for assessing disease activity and classifying complex diseases using basic clinical data and examinations to support personalized disease management.
Peri et al. (U.S. Patent 12274503 B1) teaches a system using multimodal data including ocular images, clinical information, and external factors, processed with a pre-trained machine learning model to generate holistic health embedded for detection, prediction, and prognosis of myopia likelihood, severity, onset, and progression.
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 KYRA R LAGOY whose telephone number is (703)756-1773. The examiner can normally be reached Monday - Friday, 8:00 am - 5:00 pm 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, Kambiz Abdi can be reached at (571)272-6702. 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.
/K.R.L./Examiner, Art Unit 3685
/KAMBIZ ABDI/Supervisory Patent Examiner, Art Unit 3685
Prosecution Insights