Notice of Pre-AIA or AIA Status
This action is made in response to the amendments/remarks filed on April 6, 2026. This action is made FINAL.
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Response to Amendment
The amendment filed April 6, 2026 has been entered. Claims 1-20 remain pending in the application.
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 therefore, subject to the conditions and requirements of this title.
Claims 15-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to non- statutory subject matter. The claims are directed to a computer program product which does not fall within at least one of the four categories of patent eligible subject matter because the claims are directed to software per se. The claims do not comprise of a product that has a physical or tangible form and also does not include any structural recitations; thus, the computer program product of claims 15- 20 is directed to software per se and the claim is directed to non-statutory embodiments that are not eligible for patent protection. See MPEP 2106.03(I), "As the courts' definitions of machines, manufactures and compositions of matter indicate, a product must have a physical or tangible form in order to fall within one of these statutory categories. Digitech, 758 F.3d at 1348, 111 USPQ2d at 1719. Thus, the Federal Circuit has held that a product claim to an intangible collection of information, even if created by human effort, does not fall within any statutory category. Digitech, 758 F.3d at 1350, 111 USPQ2d at 1720 (claimed "device profile" comprising two sets of data did not meet any of the categories because it was neither a process nor a tangible product). Similarly, software expressed as code or a set of instructions detached from any medium is an idea without physical embodiment. See Microsoft Corp. V. AT&T Corp., 550 U.S. 437, 449, 82 USPQ2d 1400, 1407 (2007); see also Benson, 409 U.S. 67, 175 USPQ2d 675 (An "idea" is not patent eligible). Thus, a product claim to a software program that does not also contain at least one structural limitation (such as a "means plus function" limitation) has no physical or tangible form, and thus does not fall within any statutory category".
Claims 15-20 are further rejected under 35 U.S.C. 101 because the claimed invention is directed to non-statutory subject matter. The claims are directed to a computer program product comprising one or more computer readable storage media, which does not fall within at least one of the four categories of patent eligible subject matter because the claims are directed to signals per se. Applicant’s specification para. 14 discloses “A “storage device” is any tangible device that can retain and store instructions for use by a computer processor. Without limitation, the computer readable storage medium may be an electronic storage medium, a magnetic storage medium, an optical storage medium, an electromagnetic storage medium, a semiconductor storage medium, a mechanical storage medium, or any suitable combination of the foregoing.” Although applicant states the computer readable storage medium is not to be construed as a storage in the form of transitory signals per se, the claim language does not exclude all transitory media and encompasses transitory forms of signal transmission, for example a carrier wave. See MPEP 2106.03(II), “a claim to a computer readable medium that can be a compact disc or a carrier wave covers a non-statutory embodiment and therefore should be rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter. See, e.g., Mentor Graphics v. EVE-USA, Inc., 851 F.3d at 1294-95, 112 USPQ2d at 1134 (claims to a "machine-readable medium" were non-statutory, because their scope encompassed both statutory random-access memory and non-statutory carrier waves).”
Claims 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more.
Independent Claims
Step 1 analysis:
Claim 1 is drawn to a method (i.e., process) and Claim 8 is drawn to a system, which are all within the four statutory categories. (Step 1 – Yes, the claim falls into one of the statutory categories). Although Claim 15 is indicated as directed to software per se and signals per se, claim 15 includes similar limitations to claims 1 and 8 and the 101 analysis applies to claim 15 as well.
Step 2A analysis – Prong One:
Claim 1 recites:
A computer-implemented method of assessing a medical procedure, the computer-implemented method comprising:
receiving input data corresponding to a patient having the medical procedure;
determining, based on the received input data and using a deep learning clustering model, a plurality of user actions to be performed during the medical procedure, wherein the determining of the plurality of user actions comprises:
extracting, from the received input data, one or more features associated with the medical procedure;
identifying mappings between patient data and the plurality of user actions based on the extracting of the one or more features; and
predicting the plurality of user actions based on the mappings;
executing a deep learning classification model based on the predicting of the plurality of user actions, wherein the executing of the deep learning classification model comprises:
comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions; and
determining, based on the comparing, one or more deviations in the medical procedure, wherein the one or more deviations correspond to one or more user actions of the predicted plurality of user actions not being performed during the medical procedure; and
alerting a user to the one or more deviations.
The series of steps as recited above, excluding the underlined portions, describes managing personal behavior or relationships or interactions between people including following rules or instructions, and therefore fall within the scope of certain methods of organizing human activity. Fundamentally, the method is that of a person gathering data corresponding to a patient, determining actions to be performed by a user during a medical procedure, determining deviations in the medical procedure, and alerting a person when the deviation occurs in the procedure, which encompasses tasks a provider would do in a healthcare setting during a procedure following rules and instructions, which describes a person interacting with another individual including following rules or instructions. Accordingly, the claim recites an abstract idea of managing interactions between people.
The steps of determining user actions to be performed, extracting, from the input data, features associated with the medical procedure, identifying mappings between patient data and user actions, predicting the plurality of user actions, comparing user actions performed during the medical procedure to the predicted user actions, and determining deviations in the medical procedure also falls within the “mental processes” grouping of abstract ideas, and describes concepts that can be performed in the human mind through observation, evaluation, judgement, and opinion. Each limitation can be performed in the human mind, with or without the use of a physical aid. Therefore, the claim recites an abstract idea of a mental process.
Claims 8 and 15 recite/describe nearly identical steps as 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.
Step 2A analysis – Prong 2:
This judicial exception is not integrated into a practical application. Independent claims 1, 8 and 15 recite the following additional elements beyond the abstract idea: using a deep learning clustering model, executing a deep learning classification model, one or more memories, and at least one processor. These limitations are recited at a high level of generality and amount to no more than mere instructions to apply the exception using generic computer components. The use of the models to carry out the abstract idea amounts to using a mathematical algorithm to apply the abstract idea, which amounts to mere instructions to apply the exception, as per MPEP 2106.05(f)(2). The limitations do not impose any meaningful limits on practicing the abstract idea, and therefore do not integrate the abstract idea into a practical application (see MPEP 2106.05(f)).
Specifically, the processor may be of any type now known or to be developed in the future (see Applicant’s specification para. 17) and the memory is any type of volatile memory now known or to be developed in the future. Examples include dynamic type random access memory (RAM) or static type RAM (see specification para. 20).
The clustering model may be a k-means clustering algorithm, a density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm, a gaussian mixture model algorithm, a balance iterative reducing and clustering using hierarchies (BIRCH) algorithm, an affinity propagation clustering algorithm, a means-shift clustering algorithm, an ordering points to identify the clustering structure (OPTICS) algorithm, an agglomerative hierarchy clustering algorithm, or another conventional or other clustering algorithm (see specification para. 53), which are known algorithms. The classification model may include a logistic regression model, a decision tree model, a random forest model, a support vector machine, a k-nearest neighbor model, a naive bayes classifier, or any other conventional or other deep learning model. The classification model may include any conventional or other machine learning models (e.g., mathematical/statistical, classifiers, feed-forward, recurrent, convolutional, deep learning, or other neural networks, etc.) (see specification para. 58). In addition, the use of the models to carry out the abstract idea amounts to using a mathematical algorithm to apply the abstract idea, which amounts to mere instructions to apply the exception, as per MPEP 2106.05(f)(2).
The limitation “receiving input data corresponding to a patient having the medical procedure” is mere data gathering recited at a high level of generality, and thus is insignificant extra-solution activity. See MPEP 2106.05(g) (“whether the limitation is significant”). In addition, all uses of the recited judicial exception require such data gathering and output, and, as such, these limitations do not impose any meaningful limits on the claim. These limitations amount to necessary data gathering and outputting. See MPEP 2106.05.
The additional elements do not show an improvement to the functioning of a computer or to any other technology, rather the additional elements perform general computing functions and do not indicate how the particular combination improves any technology or provides a technical solution to a technical problem. Accordingly, these additional elements, when considered separately and as an ordered combination, do not integrate the abstract idea into a practical application because they do not impose any meaningful limits on practicing the abstract idea. Therefore, Claims 1, 8, and 15 are directed to an abstract idea without practical application. (Step 2A – Prong 2: No, the additional elements are not integrated into a practical application).
Step 2B analysis:
As discussed above in “Step 2A analysis – Prong 2”, the identified additional elements in Independent Claims 1, 8, and 15 are equivalent to adding the words “apply it” on a generic computer, and/or generally link the use of the judicial exception to a particular technological environment or field of use. Therefore, the claims as a whole do not amount to significantly more than the judicial exception itself.
For the role of a computer in a computer implemented invention to be deemed meaningful in the context of this analysis, it must involve more than performance of “well- understood, routine, [and] conventional activities previously known to the industry.” Further, “the mere recitation of a generic computer cannot transform a patent ineligible abstract idea into a patent-eligible invention.”
The applicant’s specification discloses: the processor may be of any type now known or to be developed in the future (see Applicant’s specification para. 17) and the memory is any type of volatile memory now known or to be developed in the future. Examples include dynamic type random access memory (RAM) or static type RAM (see specification para. 20).
The clustering model may be a k-means clustering algorithm, a density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm, a gaussian mixture model algorithm, a balance iterative reducing and clustering using hierarchies (BIRCH) algorithm, an affinity propagation clustering algorithm, a means-shift clustering algorithm, an ordering points to identify the clustering structure (OPTICS) algorithm, an agglomerative hierarchy clustering algorithm, or another conventional or other clustering algorithm (see specification para. 53), which are known algorithms. The classification model may include a logistic regression model, a decision tree model, a random forest model, a support vector machine, a k-nearest neighbor model, a naive bayes classifier, or any other conventional or other deep learning model. The classification model may include any conventional or other machine learning models (e.g., mathematical/statistical, classifiers, feed-forward, recurrent, convolutional, deep learning, or other neural networks, etc.) (see specification para. 58). In addition, the use of the models to carry out the abstract idea amounts to using a mathematical algorithm to apply the abstract idea, which amounts to mere instructions to apply the exception, as per MPEP 2106.05(f)(2).
The limitation “receiving input data corresponding to a patient having the medical procedure” was found to be insignificant extra-solution activity in Step 2A, Prong Two, because it was determined to be an insignificant limitation as necessary data gathering. However, a conclusion that an additional element is insignificant extra-solution activity in Step 2A, Prong Two should be re-evaluated in Step 2B. See MPEP 2106.05, subsection I.A. At Step 2B, the evaluation of the insignificant extra-solution activity consideration takes into account whether or not the extra-solution activity is well understood, routine, and conventional in the field. See MPEP 2106.05(g).
Generic computer components recited as performing generic computer functions that are well-understood, routine and conventional activities amount to no more than implementing the abstract idea with a computerized system. Here, the claim limitations are similar to receiving and sending information over a network (Symantec, 838 F.3d at 1321, 120 USPQ2d at 1362 (utilizing an intermediary computer to forward information); OJP Techs., Inc., v. Amazon.com, Inc., 788 F.3d 1359, 1363, 115 USPQ2d 1090, 1093 (Fed. Cir. 2015) (sending messages over a network); buySAFE, Inc. v. Google, Inc., 765 F.3d 1350, 1355, 112 USPQ2d 1093, 1096 (Fed. Cir. 2014) (computer receives and sends information over a network); See MPEP 2106.05(d)(ll)(i)).
The claims do not include additional elements that are sufficient to amount to significantly more than the judicial exception because, as discussed above with respect to integration of the abstract idea into a practical application, using the additional elements to perform the steps for assessing a medical procedure amount to no more than using computer related devices to implement the abstract idea.
The use of a computer or processor to merely automate or implement the abstract idea cannot provide significantly more than the abstract idea itself. (See MPEP 2106.05(f) where mere instructions to apply an exception does not render an abstract idea patent eligible). There is no indication that the additional limitations alone or in combination improves the functioning of a computer or any other technology, improves another technology or technical field, or effects a transformation or reduction of a particular article to a different state or thing. Therefore, the claims are not patent eligible. The Examiner has therefore determined that no additional element, or combination of additional claims elements is/are sufficient to ensure the claims amount to significantly more than the abstract idea identified above (Step 2B: Independent claims - NO).
Dependent Claims
Dependent Claims 2-7, 9-14, and 16-20 are directed towards elements used to describe the medical procedures and user actions being performed. These elements include: (Claim 2, 9, and 16) the deep learning clustering model is trained based on the received input data; (Claim 3, 10, and 17) updating the deep learning clustering model based on user feedback relating to the medical procedure and user interactions performed during the medical procedure; (Claim 4, 11, and 18) the determining of the one or more deviations is based on a training of the deep learning classification model; (Claim 5, 12, and 19) updating the deep learning classification model based on user feedback relating to the medical procedure and user interactions performed during the medical procedure; (Claim 6, 13, and 20) the plurality of actual user actions is extracted from information selected from the group consisting of a log generated by a medical device and image data of the medical procedure; (Claim 7) the medical procedure is a radiological image analysis procedure. The elements described in the dependent claims recited above describe managing personal behavior or relationships or interactions between people including following rules or instructions, and therefore fall within the same scope of certain methods of organizing human activity as the independent claims. The type of training utilized by the claimed invention in claims 2-5, 9-12, and 16-29 is not described by the Applicant. As such the Examiner is required to analyze the training step given the broadest reasonable interpretation. The training of the machine learning is considered to be part of the abstract idea because they fall under data manipulations that humans perform and thus are part of the rules or instructions.
The elements as recited above also falls within the “mental processes” grouping of abstract
ideas, and describes concepts that can be performed in the human mind through observation,
evaluation, judgement, and opinion. Extracting data from a log or an image and performing an image analysis as the medical procedure are all tasks that can be performed in the human mind. Therefore, the dependent claims recite an abstract idea of a mental process.
This judicial exception is not integrated into a practical application. Specifically, the dependent
claims recite the following additional elements beyond the abstract idea: a deep learning clustering model and a deep learning classification model. The use of the models to carry out the abstract idea amounts to using a mathematical algorithm to apply the abstract idea, which amounts to mere instructions to apply the exception, as per MPEP 2106.05(f)(2). The limitations do not impose any meaningful limits on practicing the abstract idea, and therefore do not integrate the abstract idea into a practical application (see MPEP 2106.05(f)).
Specifically, the clustering model may be k-means clustering algorithm, a density-based spatial clustering of applications with noise (DBSCAN) clustering algorithm, a gaussian mixture model algorithm, a balance iterative reducing and clustering using hierarchies (BIRCH) algorithm, an affinity propagation clustering algorithm, a means-shift clustering algorithm, an ordering points to identify the clustering structure (OPTICS) algorithm, an agglomerative hierarchy clustering algorithm, or another conventional or other clustering algorithm (see specification para. 53). The classification model may be a logistic regression model, a decision tree model, a random forest model, a support vector machine, a k-nearest neighbor model, a naive bayes classifier, or any other conventional or other deep learning model (see specification para. 58).
The additional elements do not show an improvement to the functioning of a computer or to
any other technology, rather the additional elements perform general computing functions and do not
indicate how the particular combination improves any technology or provides a technical solution to a
technical problem. Accordingly, these additional elements, when considered separately and as an
ordered combination, do not integrate the abstract idea into a practical application because they do not
impose any meaningful limits on practicing the abstract idea. Therefore, the dependent claims are
directed to an abstract idea without practical application. (Step 2A – Prong 2: No, the additional
elements are not integrated into a practical application).
As discussed above, the identified additional elements in the dependent claims are equivalent to adding the words “apply it” on a generic computer, and/or generally link the use of the judicial exception to a particular technological environment or field of use. Therefore, the claims as a whole do not amount to significantly more than the judicial exception itself.
The use of a computer or processor to merely automate or implement the abstract idea cannot
provide significantly more than the abstract idea itself. (See MPEP 2106.05(f) where mere instructions to apply an exception does not render an abstract idea patent eligible). There is no indication that the
additional limitations alone or in combination improves the functioning of a computer or any other
technology, improves another technology or technical field, or effects a transformation or reduction of a
particular article to a different state or thing. Therefore, the claims are not patent eligible.
The Examiner has therefore determined that no additional element, or combination of
additional claims elements is/are sufficient to ensure the claims amount to significantly more than the
abstract idea identified above (Step 2B: Dependent claims - NO).
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.
This application currently names joint inventors. In considering patentability of the claims the examiner presumes that the subject matter of the various claims was commonly owned as of the effective filing date of the claimed invention(s) absent any evidence to the contrary. Applicant is advised of the obligation under 37 CFR 1.56 to point out the inventor and effective filing dates of each claim that was not commonly owned as of the effective filing date of the later invention in order for the examiner to consider the applicability of 35 U.S.C. 102(b)(2)(C) for any potential 35 U.S.C. 102(a)(2) prior art against the later invention.
Claims 1-20 are rejected under 35 U.S.C. 103 as being unpatentable over Bernard et al. (US Patent No. US 10140421 B1) (hereinafter Bernard), in view of Kumar et al. (US 2021/0027883 A1), in further view of Mojahed (WO 2020/178196 A1).
Regarding Claim 1, Bernard teaches the following:
A computer-implemented method of assessing a medical procedure, the computer-implemented method comprising (Col. 87, lines 14-16: a method for execution by a medical scan diagnosing system that includes a processor):
receiving input data corresponding to a patient having the medical procedure (Col. 49, lines 53-55 and Col. 57, lines 19-24: the medical scan image analysis function takes medical scan image data assigned to a medical scan as input and the medical scan diagnosing system is operable to receive a medical scan. The first medical scan is transmitted to a first client device associated with a user of the medical scan diagnosing system);
using a deep learning clustering model (Col. 77, lines 54-61: the learning model can additionally or alternatively include one or more of a Bayesian model, a support vector machine model, a cluster analysis model, or other supervised or unsupervised learning model. the model parameter data can be utilized to determine the corresponding medical scan image analysis functions.)
However, Bernard does not explicitly teach the following that is met by Kumar:
determining, based on the received input data and [using a deep learning clustering model], a plurality of predicted user actions to be performed during the medical procedure ([0049]-[0050] The workflow inference model is configured to infer a workflow that is most suited for a diagnostic exam retrieved by and performed by a clinician such as a radiologist. The workflow inference model may suggest/predict an order and identity of action items to be performed as part of a protocol for a diagnostic exam. The training data that may be included as the input segment of the input-output pairs to the training algorithm include, for example, clinical data (such as blood test results, blood pressure and body temperature readings, etc.), prior reports (such as conditions diagnosed from prior scans or other diagnostic tests), and image data (such as images from earlier scans).), wherein the determining of the plurality of user actions comprises:
extracting, from the received input data, one or more features associated with the medical procedure ([0065] and Fig. 5: the workflow, toolsets, and/or priors may be selected based on the identified parameters and one or more workflow models (e.g., a workflow inference model, a prior selection model, and a toolset prediction model as explained above with respect to FIG. 3). For example, the identified parameters (e.g., imaging modality used to acquire the images, diagnostic goal of the exam, anatomical features imaged, and clinician conducting the exam) may be entered as input into the one or more models, and the one or more models may output suggestions or predictions for the workflow, toolsets, and priors, which may then be saved as part of the exam. Fig. 5 shows parameters are associated with an exam from received diagnostic images. The parameters identified are imaging modality, body part/region, diagnostic goal, and clinician. The workflow is then generated based on the identified parameters and models. The examiner interprets the identified parameters from the input data to be the extracted features associated with the medical procedure);
identifying mappings between patient data and the plurality of user actions based on the extracting of the one or more features ([0064], Fig. 5, Claim 7: at 504, method 500 includes identifying parameters associated with the exam. the one or more workflow models are trained to correlate the one or more exam parameters with the set of suggested workflow parameters based on training datasets that include, for each of a plurality of prior diagnostic exams associated with different respective patients, patient clinical data, prior reports, image data, and diagnostic exam workflow usage data.); and
predicting the plurality of user actions based on the mappings ([0054], [0064], [0065], Fig. 3, and Fig. 5: The workflow inference model 320 may suggest/predict an order and identity of action items to be performed as part of a protocol for a diagnostic exam. The predictions may be based on parameters of the current exam. Accordingly, at 504, method 500 includes identifying parameters associated with the exam. The identified parameters may include, for example, an imaging modality 505 used to acquire the images to be analyzed in the exam, such as whether the images were acquired via CT, ultrasound, MRI, etc. As another example, the body part or region 506 that was scanned is identified (e.g., breast, head, thorax, abdomen, etc.). As yet another example, the diagnostic goal 507 of the exam is determined (e.g., mammogram, lesion detection, tumor progression, concussion, etc.). As a further example, the clinician performing the exam is determined at 508. The clinician may be determined based on information obtained from an electronic medical record of the patient, the exam may be automatically assigned to the clinician (e.g., based on preset rules), or the clinician may be determined when the exam is actually conducted. A workflow is generated with respect to FIG. 3, and the workflow may include selection of and/or ordering of action items to be conducted in order to carry out the exam);
executing a deep learning classification model based on the predicting of the plurality of user actions, wherein the executing of the deep learning classification model ([0056], [0069] the information regarding the tracked user interactions, as well as diagnostic accuracy of the workflow (e.g., based on follow-up information confirming or not confirming the initial findings the exam) and/or efficiency of the workflow may be used by the computing device (or another computing device at the same medical facility as the computing device executing method) and/or the central learning platform to retrain the models in a facility-specific manner. The plurality of models may be AI-based models (e.g., machine learning, neural networking, etc.) that are trained using information from prior exams for a plurality of patients.) comprises:
comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions ([0066] user interaction with the workflow/exam is tracked. When the predictions/suggestions output by the models are used to format/select a workflow, select priors, select toolsets, select findings, and/or select references to be displayed during an exam, user interaction may be tracked to determine if the user made any modifications to the workflow, such as accepting/rejecting findings, moving around where images/findings/references are displayed, searching for different priors or references, etc.); and
determining, based on the comparing, one or more deviations in the medical procedure ([0066] When the user/clinician performs the exam, the clinician may make adjustments to the workflow, such as moving where certain images, priors, etc., are displayed, adjusting display aspects of the images via selection of tools from the toolsets (e.g., pan, tilt, zoom, contrast, etc.), requesting additional priors or toolsets be displayed, and so forth. To ensure that the smart workflows automatically assembled according to the embodiments disclosed herein are actually assisting the clinicians to increase exam efficiency, user interaction with the workflow/exam is tracked)
It would have been obvious to one of ordinary skill in the art before the effective filing date to have combined the method of Bernard with the functions of determining predicted actions to be performed during a medical procedure, as taught by Kumar, because by predicting a workflow, toolset, or other protocol parameters, user fatigue and error in workflow selection may be reduced, while improving productivity and efficiency in diagnostic outcomes (See Kumar [0003]).
However, Bernard and Kumar do not explicitly disclose the following that is met by Mojahed:
wherein the one or more deviations correspond to one or more user actions of the predicted plurality of user actions not being performed during the medical procedure ([0048] Deviation from a particular set of parameters for a given scan protocol can be detected in real-time. The NotifyMe module can be configured to notify a technician of the protocol drift (e.g., variations from the standard protocol) by comparing a current imaging protocol to a standard or base protocol (e.g., an original or standard version of the protocol. The NotifyMe architecture can include notifications to the ordering physician or radiologist, or imaging facility management, regarding deviations from an ordered scan. For instance, a tech may restart a scan or omit a portion of an ordered scan); and
alerting a user to the one or more deviations ([0048] A tech may restart a scan or omit a portion of an ordered scan, and such an event triggers a notification message. For example, if an MR Technologist is told to run a particular scan protocol without any modifications, but they accidentally modify the echo time, NotifyMe can send an alert to the management, notifying them about a potential drift in the scan protocol).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have combined the method of assessing a medical procedure, as taught by Bernard and Kumar, with the determination of a deviation in the procedure and alerting a user to the deviation, as taught by Mojahed, because by providing real-time drift detection, facility management, service technicians, and even physicians can have access to more timely information that facilitates revising scan procedures and workflow (See Mojahed [0048]).
Regarding Claim 2, the combination of Bernard, Kumar, and Mojahed teaches the method of claim 1, and Bernard further teaches:
The computer-implemented method of claim 1, wherein the deep learning clustering model is trained based on the received input data (Col. 77, lines 47-61: Having determined the subregion training set of three-dimensional subregions corresponding to the set of full medical scans in the training set, the medical scan image analysis system can complete a training step by performing a learning algorithm on the plurality of three dimensional subregions to generate model parameter data of a corresponding learning model. the learning model can additionally or alternatively include one or more of a Bayesian model, a support vector machine model, a cluster analysis model, or other supervised or unsupervised learning model. the model parameter data can be utilized to determine the corresponding medical scan image analysis functions. Col. 49, lines 53-55 and Col. 57, lines 19-24: the medical scan image analysis function takes medical scan image data assigned to a medical scan as input and the medical scan diagnosing system is operable to receive a medical scan. The first medical scan is transmitted to a first client device associated with a user of the medical scan diagnosing system)).
Regarding Claim 3, the combination of Bernard, Kumar, and Mojahed teaches the method of claim 2, and Bernard further teaches:
The computer-implemented method of claim 2, further comprising: updating the deep learning clustering model (Col. 86, lines 1-3, Col. 87, lines 37-40, Col. 88, lines 40-43: generate an updated set of neural network parameters based on a calculated set of parameter errors and the preliminary set of neural network parameters. An updated first medical scan inference function is generated in response to determining the first review data indicates that the first diagnosis data is incorrect. The first medical scan inference function is updated in response to determining that the model quality check data compares unfavorably to the truth diagnosis data) based on information selected from the group consisting of: user feedback relating to the medical procedure (Col. 48, lines 65-66, Fig. 9B, Fig. 11B: Feedback from the expert via the interactive interface can be used to generate model accuracy data) and user interactions performed during the medical procedure (Col. 96, lines 10-15: User performance data corresponding to the fourth user in the user database is updated based on the annotation accuracy score).
Regarding Claim 4, the combination of Bernard, Kumar, and Mojahed teaches the method of claim 1, and Kumar further teaches:
The computer-implemented method of claim 1, wherein the determining of the one or more deviations in the medical procedure is based on a training of the deep learning classification model ([0066] When the user/clinician performs the exam, the clinician may make adjustments to the workflow, such as moving where certain images, priors, etc., are displayed, adjusting display aspects of the images via selection of tools from the toolsets (e.g., pan, tilt, zoom, contrast, etc.), requesting additional priors or toolsets be displayed, and so forth. To ensure that the smart workflows automatically assembled according to the embodiments disclosed herein are actually assisting the clinicians to increase exam efficiency, user interaction with the workflow/exam is tracked at. If the user modifies the workflow or otherwise indicates that the workflow, priors, and/or tool sets (or as will be explained below, the findings and/or references) are not adequate or sufficient, this information may be stored and used to re-train the models.).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the determination of the deviations being based on a training of the model, as taught by Kumar, because by leveraging AI algorithms to search through the data, multi-factorial, computation heavy analyses can be performed more accurately, enabling potential issues to be identified that would have otherwise been missed. Further, by sharing workflows at a central learning platform, best-in-class reading protocols and practices can be available to all users across a range of healthcare environments. Further, unwanted user actions can be removed, improving productivity (See Kumar [0089]).
Regarding Claim 5, the combination of Bernard, Kumar, and Mojahed teaches the method of claim 4, and Kumar further teaches:
The computer-implemented method of claim 4, further comprising: updating the deep learning classification model based on information selected from the group consisting of: user feedback relating to the medical procedure ([0053], [0084] The models may be updated based on data collected during the diagnostic exam. For example, a feedback engine may be configured to monitor the settings selected by the user when performing the exam. Multiple action items of the workflow are displayed via workflow menu. These include steps that the user may follow, and an order in which they need to be followed, for the diagnosis to be performed reliably. The action items of the workflow may include the list of steps predicted by the one or more models. Typically, the user will go through each of the steps to complete the exam. While the smart engine/inferencing service generates the sequence of steps, the menu provides the user the option of changing the sequence of steps. If this happens, feedback regarding the usage and optimize the sequence generation may be saved for later uses.) and user interactions performed during the medical procedure ([0074] the models may be updated based on received workflow usage data. For example, as explained above, when the predictions/suggestions output by the models are used to format/select a workflow, select priors, select toolsets, select findings, and/or select references to be displayed during an exam, user interaction may be tracked to determine if the user made any modifications to the work flow, such as accepting/rejecting findings, moving around where images/findings/references are displayed, searching for different priors or references, etc. Any of these interactions may indicate the predictions/suggestions were not optimal, and thus may be used to re-train the models).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the updating of the classification model based on feedback and user interaction, as taught by Kumar, because by retraining the models utilizing feedback and tracked user interactions, the workflows may be generated according to preferred medical facility guidelines/protocols, which may help each facility stay in compliance with appropriate jurisdiction-specific regulations (See Kumar [0069]).
Regarding Claim 6, the combination of Bernard, Kumar, and Mojahed teaches the method of claim 1, and Kumar further teaches:
The computer-implemented method of claim 1, wherein the plurality of actual user actions is extracted from information selected from the group consisting of: a log generated by a medical device, and image data of the medical procedure ([0071] The workflow usage data and parameters for the exam may include the hanging protocol of the images for the exam, the viewport at which each prior, reference image or literature, etc., was displayed, clinical data that was displayed during the exam, tools that were used during the exam, findings recorded in the exam, exam duration, and relevant follow-up information (e.g., confirmed patient diagnoses, confirmed misdiagnoses)).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the actual user actions being extracted from a log and image data, as taught by Kumar, because by tracking user interactions, the workflows may be generated according to preferred medical facility guidelines/protocols, which may help each facility stay in compliance with appropriate jurisdiction-specific regulations (See Kumar [0069]).
Regarding Claim 7, the combination of Bernard, Kumar, and Mojahed teaches the method of claim 1, and Bernard further teaches:
The computer-implemented method of claim 1, wherein the medical procedure is a radiological image analysis procedure (Col. 5, lines 28-31: The method involves a medical scan including imaging data corresponding to a CT scan, x-ray, or any other type of radiological scan or medical scan).
Regarding Claim 8, Bernard teaches the following:
A system for assessing a medical procedure (Col. 2, line 58: a medical scan processing system), the system comprising:
one or more memories (Col. 3, lines 19-20: one or more memory devices of one or more subsystems);
at least one processor coupled to the one or more memories, wherein the at least one processor (Col. 72, lines 31-34: medical scan image analysis system can include a processing system that includes a processor and a memory that stores executable instructions that, when executed by the processing system, facilitate performance of operations) is configured to:
receive input data corresponding to a patient having the medical procedure (Col. 49, lines 53-55 and Col. 57, lines 19-24: the medical scan image analysis function takes medical scan image data assigned to a medical scan as input and the medical scan diagnosing system is operable to receive a medical scan. The first medical scan is transmitted to a first client device associated with a user of the medical scan diagnosing system);
using a deep learning clustering model (Col. 77, lines 54-61: the learning model can additionally or alternatively include one or more of a Bayesian model, a support vector machine model, a cluster analysis model, or other supervised or unsupervised learning model. the model parameter data can be utilized to determine the corresponding medical scan image analysis functions.)
However, Bernard does not explicitly teach the following that is met by Kumar:
determine, based on the received input data and [using a deep learning clustering model], a plurality of user actions to be performed during the medical procedure, wherein the determination of the plurality of user actions comprises: ([0049]-[0050] The workflow inference model is configured to infer a workflow that is most suited for a diagnostic exam retrieved by and performed by a clinician such as a radiologist. The workflow inference model may suggest/predict an order and identity of action items to be performed as part of a protocol for a diagnostic exam. The training data that may be included as the input segment of the input-output pairs to the training algorithm include, for example, clinical data (such as blood test results, blood pressure and body temperature readings, etc.), prior reports (such as conditions diagnosed from prior scans or other diagnostic tests), and image data (such as images from earlier scans).)
extracting, from the received input data, one or more features associated with the medical procedure ([0065] and Fig. 5: the workflow, toolsets, and/or priors may be selected based on the identified parameters and one or more workflow models (e.g., a workflow inference model, a prior selection model, and a toolset prediction model as explained above with respect to FIG. 3). For example, the identified parameters (e.g., imaging modality used to acquire the images, diagnostic goal of the exam, anatomical features imaged, and clinician conducting the exam) may be entered as input into the one or more models, and the one or more models may output suggestions or predictions for the workflow, toolsets, and priors, which may then be saved as part of the exam. Fig. 5 shows parameters are associated with an exam from received diagnostic images. The parameters identified are imaging modality, body part/region, diagnostic goal, and clinician. The workflow is then generated based on the identified parameters and models. The examiner interprets the identified parameters from the input data to be the extracted features associated with the medical procedure);
identifying mappings between patient data and the plurality of user actions based on the extraction of the one or more features ([0064], Fig. 5, Claim 7: at 504, method 500 includes identifying parameters associated with the exam. the one or more workflow models are trained to correlate the one or more exam parameters with the set of suggested workflow parameters based on training datasets that include, for each of a plurality of prior diagnostic exams associated with different respective patients, patient clinical data, prior reports, image data, and diagnostic exam workflow usage data.); and
predicting the plurality of user actions based on the mappings ([0054], [0064], [0065], Fig. 3, and Fig. 5: The workflow inference model 320 may suggest/predict an order and identity of action items to be performed as part of a protocol for a diagnostic exam. The predictions may be based on parameters of the current exam. Accordingly, at 504, method 500 includes identifying parameters associated with the exam. The identified parameters may include, for example, an imaging modality 505 used to acquire the images to be analyzed in the exam, such as whether the images were acquired via CT, ultrasound, MRI, etc. As another example, the body part or region 506 that was scanned is identified (e.g., breast, head, thorax, abdomen, etc.). As yet another example, the diagnostic goal 507 of the exam is determined (e.g., mammogram, lesion detection, tumor progression, concussion, etc.). As a further example, the clinician performing the exam is determined at 508. The clinician may be determined based on information obtained from an electronic medical record of the patient, the exam may be automatically assigned to the clinician (e.g., based on preset rules), or the clinician may be determined when the exam is actually conducted. A workflow is generated with respect to FIG. 3, and the workflow may include selection of and/or ordering of action items to be conducted in order to carry out the exam);
execute a deep learning classification model based on the predicting of the plurality of user actions, wherein the execution of the deep learning classification model ([0056], [0069] the information regarding the tracked user interactions, as well as diagnostic accuracy of the workflow (e.g., based on follow-up information confirming or not confirming the initial findings the exam) and/or efficiency of the workflow may be used by the computing device (or another computing device at the same medical facility as the computing device executing method) and/or the central learning platform to retrain the models in a facility-specific manner. The plurality of models may be AI-based models (e.g., machine learning, neural networking, etc.) that are trained using information from prior exams for a plurality of patients.) comprises:
comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions ([0066] user interaction with the workflow/exam is tracked. When the predictions/suggestions output by the models are used to format/select a workflow, select priors, select toolsets, select findings, and/or select references to be displayed during an exam, user interaction may be tracked to determine if the user made any modifications to the workflow, such as accepting/rejecting findings, moving around where images/findings/references are displayed, searching for different priors or references, etc.); and
determining, based on the comparison, one or more deviations in the medical procedure ([0066] When the user/clinician performs the exam, the clinician may make adjustments to the workflow, such as moving where certain images, priors, etc., are displayed, adjusting display aspects of the images via selection of tools from the toolsets (e.g., pan, tilt, zoom, contrast, etc.), requesting additional priors or toolsets be displayed, and so forth. To ensure that the smart workflows automatically assembled according to the embodiments disclosed herein are actually assisting the clinicians to increase exam efficiency, user interaction with the workflow/exam is tracked)
It would have been obvious to one of ordinary skill in the art before the effective filing date to have combined the method of Bernard with the functions of determining predicted actions to be performed during a medical procedure, as taught by Kumar, because by predicting a workflow, toolset, or other protocol parameters, user fatigue and error in workflow selection may be reduced, while improving productivity and efficiency in diagnostic outcomes (See Kumar [0003]).
However, Bernard and Kumar do not explicitly disclose the following that is met by Mojahed:
wherein the one or more deviations correspond to one or more user actions of the predicted plurality of user actions not being performed during the medical procedure ([0048] Deviation from a particular set of parameters for a given scan protocol can be detected in real-time. The NotifyMe module can be configured to notify a technician of the protocol drift (e.g., variations from the standard protocol) by comparing a current imaging protocol to a standard or base protocol (e.g., an original or standard version of the protocol. The NotifyMe architecture can include notifications to the ordering physician or radiologist, or imaging facility management, regarding deviations from an ordered scan. For instance, a tech may restart a scan or omit a portion of an ordered scan); and
alert a user to the one or more deviations ([0048] A tech may restart a scan or omit a portion of an ordered scan, and such an event triggers a notification message. For example, if an MR Technologist is told to run a particular scan protocol without any modifications, but they accidentally modify the echo time, NotifyMe can send an alert to the management, notifying them about a potential drift in the scan protocol).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have combined the method of assessing a medical procedure, as taught by Bernard and Kumar, with the determination of a deviation in the procedure and alerting a user to the deviation, as taught by Mojahed, because by providing real-time drift detection, facility management, service technicians, and even physicians can have access to more timely information that facilitates revising scan procedures and workflow (See Mojahed [0048]).
Regarding Claim 9, the combination of Bernard, Kumar, and Mojahed teaches the system of claim 8, and Bernard further teaches:
The system of claim 8, wherein the deep learning clustering model is trained based on the received input data (Col. 77, lines 47-61: Having determined the subregion training set of three-dimensional subregions corresponding to the set of full medical scans in the training set, the medical scan image analysis system can complete a training step by performing a learning algorithm on the plurality of three dimensional subregions to generate model parameter data of a corresponding learning model. the learning model can additionally or alternatively include one or more of a Bayesian model, a support vector machine model, a cluster analysis model, or other supervised or unsupervised learning model. the model parameter data can be utilized to determine the corresponding medical scan image analysis functions. Col. 49, lines 53-55 and Col. 57, lines 19-24: the medical scan image analysis function takes medical scan image data assigned to a medical scan as input and the medical scan diagnosing system is operable to receive a medical scan. The first medical scan is transmitted to a first client device associated with a user of the medical scan diagnosing system)).
Regarding Claim 10, the combination of Bernard, Kumar, and Mojahed teaches the system of claim 9, and Bernard further teaches:
The system of claim 9, wherein the at least one processor is further configured to: update the deep learning clustering model (Col. 86, lines 1-3, Col. 87, lines 37-40, Col. 88, lines 40-43: generate an updated set of neural network parameters based on a calculated set of parameter errors and the preliminary set of neural network parameters. An updated first medical scan inference function is generated in response to determining the first review data indicates that the first diagnosis data is incorrect. The first medical scan inference function is updated in response to determining that the model quality check data compares unfavorably to the truth diagnosis data) based on information selected from the group consisting of: user feedback relating to the medical procedure (Col. 48, lines 65-66, Fig. 9B, Fig. 11B: Feedback from the expert via the interactive interface can be used to generate model accuracy data) and user interactions performed during the medical procedure (Col. 96, lines 10-15: User performance data corresponding to the fourth user in the user database is updated based on the annotation accuracy score).
Regarding Claim 11, the combination of Bernard, Kumar, and Mojahed teaches the system of claim 8, and Kumar teaches:
The system of claim 8, wherein the determination of the one or more deviations is based on a training of the deep learning classification model ([0066] When the user/clinician performs the exam, the clinician may make adjustments to the workflow, such as moving where certain images, priors, etc., are displayed, adjusting display aspects of the images via selection of tools from the toolsets (e.g., pan, tilt, zoom, contrast, etc.), requesting additional priors or toolsets be displayed, and so forth. To ensure that the smart workflows automatically assembled according to the embodiments disclosed herein are actually assisting the clinicians to increase exam efficiency, user interaction with the workflow/exam is tracked at. If the user modifies the workflow or otherwise indicates that the workflow, priors, and/or tool sets (or as will be explained below, the findings and/or references) are not adequate or sufficient, this information may be stored and used to re-train the models.).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the determination of the deviations being based on a training of the model, as taught by Kumar, because by leveraging AI algorithms to search through the data, multi-factorial, computation heavy analyses can be performed more accurately, enabling potential issues to be identified that would have otherwise been missed. Further, by sharing workflows at a central learning platform, best-in-class reading protocols and practices can be available to all users across a range of healthcare environments. Further, unwanted user actions can be removed, improving productivity (See Kumar [0089]).
Regarding Claim 12, the combination of Bernard, Kumar, and Mojahed teaches the system of claim 11, and Kumar further teaches:
The system of claim 11, wherein the at least one processor is further configured to: update the deep learning classification model based on information selected from the group consisting of: user feedback relating to the medical procedure ([0053], [0084] The models may be updated based on data collected during the diagnostic exam. For example, a feedback engine may be configured to monitor the settings selected by the user when performing the exam. Multiple action items of the workflow are displayed via workflow menu. These include steps that the user may follow, and an order in which they need to be followed, for the diagnosis to be performed reliably. The action items of the workflow may include the list of steps predicted by the one or more models. Typically, the user will go through each of the steps to complete the exam. While the smart engine/inferencing service generates the sequence of steps, the menu provides the user the option of changing the sequence of steps. If this happens, feedback regarding the usage and optimize the sequence generation may be saved for later uses.), and user interactions performed during the medical procedure ([0074] the models may be updated based on received workflow usage data. For example, as explained above, when the predictions/suggestions output by the models are used to format/select a workflow, select priors, select toolsets, select findings, and/or select references to be displayed during an exam, user interaction may be tracked to determine if the user made any modifications to the work flow, such as accepting/rejecting findings, moving around where images/findings/references are displayed, searching for different priors or references, etc. Any of these interactions may indicate the predictions/suggestions were not optimal, and thus may be used to re-train the models).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the updating of the classification model based on feedback and user interaction, as taught by Kumar, because by retraining the models utilizing feedback and tracked user interactions, the workflows may be generated according to preferred medical facility guidelines/protocols, which may help each facility stay in compliance with appropriate jurisdiction-specific regulations (See Kumar [0069]).
Regarding Claim 13, the combination of Bernard, Kumar, and Mojahed teaches the system of claim 8, and Kumar further teaches:
The system of claim 8, wherein the plurality of actual user actions is extracted from information selected from a group consisting of: a log generated by a medical device and image data of the medical procedure ([0071] The workflow usage data and parameters for the exam may include the hanging protocol of the images for the exam, the viewport at which each prior, reference image or literature, etc., was displayed, clinical data that was displayed during the exam, tools that were used during the exam, findings recorded in the exam, exam duration, and relevant follow-up information (e.g., confirmed patient diagnoses, confirmed misdiagnoses)).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the actual user actions being extracted from a log and image data, as taught by Kumar, because by tracking user interactions, the workflows may be generated according to preferred medical facility guidelines/protocols, which may help each facility stay in compliance with appropriate jurisdiction-specific regulations (See Kumar [0069]).
Regarding Claim 14, the combination of Bernard, Kumar, and Mojahed teaches the system of claim 8, and Bernard further teaches:
The system of claim 8, wherein the medical procedure is a radiological image analysis procedure (Col. 5, lines 28-31: The method involves a medical scan including imaging data corresponding to a CT scan, x-ray, or any other type of radiological scan or medical scan).
Regarding Claim 15, Bernard teaches the following:
A computer program product for assessing a medical procedure (Col. 104, lines 10-17, Col. 87, lines 14-16: a method for execution by a medical scan diagnosing system that includes a processor. "processing unit" may be a single processing device or a plurality of processing devices. Such a processing device may be a microprocessor, micro-controller, digital signal processor, graphics processing unit, microcomputer, central processing unit, field programmable gate array, programmable logic device, state machine, logic circuitry, analog circuitry, digital circuitry, and/or any device that manipulates signals), the computer program product comprising one or more computer readable storage media having program instructions collectively stored on the one or more computer readable storage media (Col. 105, lines 60-61 and Col. 106, line 1: a computer readable memory includes one or more memory elements that stores digital information), the program instruction executable by at least one processor to cause the at least one processor to ( Col. 105, lines 23-25: processors executing appropriate software and the like or any combination thereof):
receive input data corresponding to a patient having the medical procedure (Col. 49, lines 53-55 and Col. 57, lines 19-24: the medical scan image analysis function takes medical scan image data assigned to a medical scan as input and the medical scan diagnosing system is operable to receive a medical scan. The first medical scan is transmitted to a first client device associated with a user of the medical scan diagnosing system);
using a deep learning clustering model (Col. 77, lines 54-61: the learning model can additionally or alternatively include one or more of a Bayesian model, a support vector machine model, a cluster analysis model, or other supervised or unsupervised learning model. the model parameter data can be utilized to determine the corresponding medical scan image analysis functions.)
However, Bernard does not explicitly teach the following that is met by Kumar:
determine, based on the received input data and [using a deep learning clustering model], a plurality of user actions to be performed during the medical procedure, wherein the determination of the plurality of user actions comprises ([0049]-[0050] The workflow inference model is configured to infer a workflow that is most suited for a diagnostic exam retrieved by and performed by a clinician such as a radiologist. The workflow inference model may suggest/predict an order and identity of action items to be performed as part of a protocol for a diagnostic exam. The training data that may be included as the input segment of the input-output pairs to the training algorithm include, for example, clinical data (such as blood test results, blood pressure and body temperature readings, etc.), prior reports (such as conditions diagnosed from prior scans or other diagnostic tests), and image data (such as images from earlier scans).):
extracting, from the received input data, one or more features associated with the medical procedure ([0065] and Fig. 5: the workflow, toolsets, and/or priors may be selected based on the identified parameters and one or more workflow models (e.g., a workflow inference model, a prior selection model, and a toolset prediction model as explained above with respect to FIG. 3). For example, the identified parameters (e.g., imaging modality used to acquire the images, diagnostic goal of the exam, anatomical features imaged, and clinician conducting the exam) may be entered as input into the one or more models, and the one or more models may output suggestions or predictions for the workflow, toolsets, and priors, which may then be saved as part of the exam. Fig. 5 shows parameters are associated with an exam from received diagnostic images. The parameters identified are imaging modality, body part/region, diagnostic goal, and clinician. The workflow is then generated based on the identified parameters and models. The examiner interprets the identified parameters from the input data to be the extracted features associated with the medical procedure);
identifying mappings between patient data and the plurality of user actions based on the extraction of the one or more features ([0064], Fig. 5, Claim 7: at 504, method 500 includes identifying parameters associated with the exam. the one or more workflow models are trained to correlate the one or more exam parameters with the set of suggested workflow parameters based on training datasets that include, for each of a plurality of prior diagnostic exams associated with different respective patients, patient clinical data, prior reports, image data, and diagnostic exam workflow usage data.); and
predicting the plurality of user actions based on the mappings ([0054], [0064], [0065], Fig. 3, and Fig. 5: The workflow inference model 320 may suggest/predict an order and identity of action items to be performed as part of a protocol for a diagnostic exam. The predictions may be based on parameters of the current exam. Accordingly, at 504, method 500 includes identifying parameters associated with the exam. The identified parameters may include, for example, an imaging modality 505 used to acquire the images to be analyzed in the exam, such as whether the images were acquired via CT, ultrasound, MRI, etc. As another example, the body part or region 506 that was scanned is identified (e.g., breast, head, thorax, abdomen, etc.). As yet another example, the diagnostic goal 507 of the exam is determined (e.g., mammogram, lesion detection, tumor progression, concussion, etc.). As a further example, the clinician performing the exam is determined at 508. The clinician may be determined based on information obtained from an electronic medical record of the patient, the exam may be automatically assigned to the clinician (e.g., based on preset rules), or the clinician may be determined when the exam is actually conducted. A workflow is generated with respect to FIG. 3, and the workflow may include selection of and/or ordering of action items to be conducted in order to carry out the exam);
execute a deep learning classification model based on the predicting of the plurality of user actions, wherein the execution of the deep learning classification model ([0056], [0069] the information regarding the tracked user interactions, as well as diagnostic accuracy of the workflow (e.g., based on follow-up information confirming or not confirming the initial findings the exam) and/or efficiency of the workflow may be used by the computing device (or another computing device at the same medical facility as the computing device executing method) and/or the central learning platform to retrain the models in a facility-specific manner. The plurality of models may be AI-based models (e.g., machine learning, neural networking, etc.) that are trained using information from prior exams for a plurality of patients.) comprises:
comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions ([0066] user interaction with the workflow/exam is tracked. When the predictions/suggestions output by the models are used to format/select a workflow, select priors, select toolsets, select findings, and/or select references to be displayed during an exam, user interaction may be tracked to determine if the user made any modifications to the workflow, such as accepting/rejecting findings, moving around where images/findings/references are displayed, searching for different priors or references, etc.); and
determining, based on the comparison, one or more deviations in the medical procedure ([0066] When the user/clinician performs the exam, the clinician may make adjustments to the workflow, such as moving where certain images, priors, etc., are displayed, adjusting display aspects of the images via selection of tools from the toolsets (e.g., pan, tilt, zoom, contrast, etc.), requesting additional priors or toolsets be displayed, and so forth. To ensure that the smart workflows automatically assembled according to the embodiments disclosed herein are actually assisting the clinicians to increase exam efficiency, user interaction with the workflow/exam is tracked)
It would have been obvious to one of ordinary skill in the art before the effective filing date to have combined the method of Bernard with the functions of determining predicted actions to be performed during a medical procedure, as taught by Kumar, because by predicting a workflow, toolset, or other protocol parameters, user fatigue and error in workflow selection may be reduced, while improving productivity and efficiency in diagnostic outcomes (See Kumar [0003]).
However, Bernard and Kumar do not explicitly disclose the following that is met by Mojahed:
wherein the one or more deviations correspond to one or more user actions of the predicted plurality of user actions not being performed during the medical procedure ([0048] Deviation from a particular set of parameters for a given scan protocol can be detected in real-time. The NotifyMe module can be configured to notify a technician of the protocol drift (e.g., variations from the standard protocol) by comparing a current imaging protocol to a standard or base protocol (e.g., an original or standard version of the protocol. The NotifyMe architecture can include notifications to the ordering physician or radiologist, or imaging facility management, regarding deviations from an ordered scan. For instance, a tech may restart a scan or omit a portion of an ordered scan); and
alert a user to the one or more deviations ([0048] A tech may restart a scan or omit a portion of an ordered scan, and such an event triggers a notification message. For example, if an MR Technologist is told to run a particular scan protocol without any modifications, but they accidentally modify the echo time, NotifyMe can send an alert to the management, notifying them about a potential drift in the scan protocol).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have combined the method of assessing a medical procedure, as taught by Bernard and Kumar, with the determination of a deviation in the procedure and alerting a user to the deviation, as taught by Mojahed, because by providing real-time drift detection, facility management, service technicians, and even physicians can have access to more timely information that facilitates revising scan procedures and workflow (See Mojahed [0048]).
Regarding Claim 16, the combination of Bernard, Kumar, and Mojahed teaches the computer program product of claim 15, and Bernard further teaches:
The computer program product of claim 15, wherein the deep learning clustering model is trained based on the received input data (Col. 77, lines 47-61: Having determined the subregion training set of three-dimensional subregions corresponding to the set of full medical scans in the training set, the medical scan image analysis system can complete a training step by performing a learning algorithm on the plurality of three dimensional subregions to generate model parameter data of a corresponding learning model. the learning model can additionally or alternatively include one or more of a Bayesian model, a support vector machine model, a cluster analysis model, or other supervised or unsupervised learning model. the model parameter data can be utilized to determine the corresponding medical scan image analysis functions. Col. 49, lines 53-55 and Col. 57, lines 19-24: the medical scan image analysis function takes medical scan image data assigned to a medical scan as input and the medical scan diagnosing system is operable to receive a medical scan. The first medical scan is transmitted to a first client device associated with a user of the medical scan diagnosing system).
Regarding Claim 17, the combination of Bernard, Kumar, and Mojahed teaches the computer program product of claim 16, and Bernard further teaches:
The computer program product of claim 16, wherein the program instructions further cause the at least one processor to: update the deep learning clustering model (Col. 86, lines 1-3, Col. 87, lines 37-40, Col. 88, lines 40-43: generate an updated set of neural network parameters based on a calculated set of parameter errors and the preliminary set of neural network parameters. An updated first medical scan inference function is generated in response to determining the first review data indicates that the first diagnosis data is incorrect. The first medical scan inference function is updated in response to determining that the model quality check data compares unfavorably to the truth diagnosis data) based on information selected from the group consisting of: user feedback relating to the medical procedure (Col. 48, lines 65-66, Fig. 9B, Fig. 11B: Feedback from the expert via the interactive interface can be used to generate model accuracy data) and user interactions performed during the medical procedure (Col. 96, lines 10-15: User performance data corresponding to the fourth user in the user database is updated based on the annotation accuracy score).
Regarding Claim 18, the combination of Bernard, Kumar, and Mojahed teaches the computer program product of claim 15, and Kumar further teaches:
The computer program product of claim 15, wherein the determination of the one or more deviations is based on a training of the deep learning classification model ([0066] When the user/clinician performs the exam, the clinician may make adjustments to the workflow, such as moving where certain images, priors, etc., are displayed, adjusting display aspects of the images via selection of tools from the toolsets (e.g., pan, tilt, zoom, contrast, etc.), requesting additional priors or toolsets be displayed, and so forth. To ensure that the smart workflows automatically assembled according to the embodiments disclosed herein are actually assisting the clinicians to increase exam efficiency, user interaction with the workflow/exam is tracked at. If the user modifies the workflow or otherwise indicates that the workflow, priors, and/or tool sets (or as will be explained below, the findings and/or references) are not adequate or sufficient, this information may be stored and used to re-train the models.).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the determination of the deviations being based on a training of the model, as taught by Kumar, because by leveraging AI algorithms to search through the data, multi-factorial, computation heavy analyses can be performed more accurately, enabling potential issues to be identified that would have otherwise been missed. Further, by sharing workflows at a central learning platform, best-in-class reading protocols and practices can be available to all users across a range of healthcare environments. Further, unwanted user actions can be removed, improving productivity (See Kumar [0089]).
Regarding Claim 19, the combination of Bernard, Kumar, and Mojahed teaches the computer program product of claim 18, and Kumar further teaches:
The computer program product of claim 19, wherein the program instructions further cause the at least one processor to: update the deep learning classification model based on information selected from the group consisting of: user feedback relating to the medical procedure ([0053], [0084] The models may be updated based on data collected during the diagnostic exam. For example, a feedback engine may be configured to monitor the settings selected by the user when performing the exam. Multiple action items of the workflow are displayed via workflow menu. These include steps that the user may follow, and an order in which they need to be followed, for the diagnosis to be performed reliably. The action items of the workflow may include the list of steps predicted by the one or more models. Typically, the user will go through each of the steps to complete the exam. While the smart engine/inferencing service generates the sequence of steps, the menu provides the user the option of changing the sequence of steps. If this happens, feedback regarding the usage and optimize the sequence generation may be saved for later uses.), and user interactions performed during the medical procedure ([0074] the models may be updated based on received workflow usage data. For example, as explained above, when the predictions/suggestions output by the models are used to format/select a workflow, select priors, select toolsets, select findings, and/or select references to be displayed during an exam, user interaction may be tracked to determine if the user made any modifications to the work flow, such as accepting/rejecting findings, moving around where images/findings/references are displayed, searching for different priors or references, etc. Any of these interactions may indicate the predictions/suggestions were not optimal, and thus may be used to re-train the models).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the updating of the classification model based on feedback and user interaction, as taught by Kumar, because by retraining the models utilizing feedback and tracked user interactions, the workflows may be generated according to preferred medical facility guidelines/protocols, which may help each facility stay in compliance with appropriate jurisdiction-specific regulations (See Kumar [0069]).
Regarding Claim 20, the combination of Bernard, Kumar, and Mojahed teaches the computer program product of claim 15, and Kumar further teaches:
The computer program product of claim 15, wherein the plurality of actual user actions is extracted from information selected from the group consisting of: a log generated by a medical device and image data of the medical procedure ([0071] The workflow usage data and parameters for the exam may include the hanging protocol of the images for the exam, the viewport at which each prior, reference image or literature, etc., was displayed, clinical data that was displayed during the exam, tools that were used during the exam, findings recorded in the exam, exam duration, and relevant follow-up information (e.g., confirmed patient diagnoses, confirmed misdiagnoses)).
It would have been obvious to one of ordinary skill in the art before the effective filing date to have modified the combination of Bernard, Kumar, and Mojahed, to include the actual user actions being extracted from a log and image data, as taught by Kumar, because by tracking user interactions, the workflows may be generated according to preferred medical facility guidelines/protocols, which may help each facility stay in compliance with appropriate jurisdiction-specific regulations (See Kumar [0069]).
Response to Arguments
Applicant's arguments filed 04/06/2026 have been fully considered but they are not persuasive. Applicant argues claims 15-20 should not be construed as being transitory signals per se, however, the examiner respectfully disagrees. Although applicant states the computer readable storage medium is not to be construed as a storage in the form of transitory signals per se, the claim language does not exclude all transitory media and encompasses transitory forms of signal transmission, for example a carrier wave. See MPEP 2106.03(II), “a claim to a computer readable medium that can be a compact disc or a carrier wave covers a non-statutory embodiment and therefore should be rejected under 35 U.S.C. 101 as being directed to non-statutory subject matter. See, e.g., Mentor Graphics v. EVE-USA, Inc., 851 F.3d at 1294-95, 112 USPQ2d at 1134 (claims to a "machine-readable medium" were non-statutory, because their scope encompassed both statutory random-access memory and non-statutory carrier waves). Therefore, the rejection is maintained.
The Applicant argues that the claims are not directed to an abstract idea, however, the examiner respectfully disagrees. The steps of determining user actions to be performed, extracting, from the input data, features associated with the medical procedure, identifying mappings between patient data and user actions, predicting the plurality of user actions, comparing user actions performed during the medical procedure to the predicted user actions, and determining deviations in the medical procedure are all concepts that can be performed in the human mind through observation, evaluation, judgement, and opinion. A person can predict actions to be performed during a medical procedure by extracting information for input data and correlating the data with actions to be performed. The person can also compare actual actions performed to the predicted actions and determine using observations that there was a deviation in the procedure. Therefore, the claims are directed to an abstract idea. The applicant argues there is a technological improvement to the field of assessing medical procedures for completeness by integrating machine learning into a practical application that improves clinical workflows by detecting deviations in real time and alerting clinicians so that any deviations can be addressed during the procedure. However, the examiner respectfully disagrees. The use of the clustering and classification models to carry out the abstract idea amounts to using a mathematical algorithm to apply the abstract idea, which amounts to mere instructions to apply the exception. The models are used to generally apply the abstract idea without placing any limits on how the models function. Rather, these limitations only recite the outcome of “determining a plurality of user actions”, “comparing a plurality of actual user actions to the predicted plurality of user actions”, and “determining… one or more deviations”, and do not include any details about how the “determining” and “comparing” are accomplished. See MPEP 2016.05(f). Furthermore, the claims do not recite detecting deviations in real time or alerting clinicians so that any deviations can be addressed during the procedure, the claims recite “comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions” and “determining… one or more deviations in the medical procedure, wherein the one or more deviations correspond to one or more user actions of the predicted plurality of user actions not being performed during the medical procedure”. The claim language does not require that the deviations are detected in real time, only that they are detected based on actions that have been performed during the procedure. Therefore, the rejection under 35 U.S.C. 101 is maintained.
Applicant’s arguments, see pg. 16-19 of Applicant’s Remarks, filed 04/06/2026, with respect to the rejection(s) of claim(s) 1-20 under 35 U.S.C. 102 have been fully considered and are persuasive. Therefore, the rejection has been withdrawn. However, upon further consideration, a new ground(s) of rejection is made in view of Kumar et al., in further view of Mojahed. The examiner agrees that Bernard does not expressly or inherently describe the features of "extracting, from the received input data, one or more features associated with the medical procedure identifying mappings between patient data and the plurality of user actions based on the extracting of the one or more features predicting the plurality of user actions based on the mappings comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions determining, based on the comparing, one or more deviations in the medical procedure, wherein the one or more deviations correspond to one or more user actions of the predicted user actions not being performed during the medical procedure". However, the limitations of "extracting, from the received input data, one or more features associated with the medical procedure identifying mappings between patient data and the plurality of user actions based on the extracting of the one or more features predicting the plurality of user actions based on the mappings comparing a plurality of actual user actions performed during the medical procedure to the predicted plurality of user actions determining, based on the comparing, one or more deviations in the medical procedure” are taught in view of the newly found prior art, Kumar. Additionally, the limitation of “wherein the one or more deviations correspond to one or more user actions of the predicted user actions not being performed during the medical procedure" is taught in further view of the newly found prior art, Mojahed.
Conclusion
The relevant art made of record and not relied upon is considered pertinent to applicant’s disclosure.
Lyman et al. (US 2022/0005187) discloses a medical scan viewing system that generates inference data based on receiver operating characteristic parameters, which utilizes deep learning mechanisms for producing the outputs.
THIS ACTION IS MADE FINAL. 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 ALEXIS K VAN DUZER whose telephone number is (571)270-5832. The examiner can normally be reached Monday thru Thursday 8-5 CT.
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, Fonya Long can be reached at (571) 270-5096. 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.
/A.K.V./Examiner, Art Unit 3682
/EVANGELINE BARR/Primary Examiner, Art Unit 3682
Prosecution Insights