DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
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 1-20 are rejected under 35 U.S.C. § 101 because the claimed invention is directed to a judicial exception without significantly more.
Representative claim 1 is analyzed below. Independent claims 9, 18, 19, and 20 recite substantially similar subject matter and are directed to the same judicial exception. The remaining claims stand or fall therewith because they merely add insignificant extra-solution activity, data gathering limitations, or additional abstract concepts.
STEP 1
Claim 1 is directed to a machine and therefore falls within one of the statutory categories of invention under 35 U.S.C. § 101.
Accordingly, the claim is analyzed under the Alice/Mayo framework and the 2024 USPTO Subject Matter Eligibility Guidance.
STEP 2A – PRONG ONE
Claim 1 recites a judicial exception.
Specifically, claim 1 recites limitations directed to: receiving a radiology report; analyzing the radiology report; predicting one or more billing codes; identifying missing content supporting the billing codes; ranking missing billing-code-related information based on probability; determining whether content is missing; presenting suggested additions; and storing the report.
These limitations describe evaluating information, analyzing documentation, determining whether information satisfies billing requirements, predicting coding outcomes, and identifying deficiencies in documentation.
Such activities constitute commercial and administrative practices relating to billing, coding, reimbursement, and documentation review. The claimed invention manages financial and administrative aspects of healthcare services by determining billing codes and evaluating whether documentation supports reimbursement requirements.
Accordingly, the claim recites certain methods of organizing human activity, including commercial interactions and financial management practices.
Further, the limitations directed to:
• analyzing the report;
• identifying missing content;
• predicting billing codes;
• ranking probabilities;
• determining completeness; and
• selecting suggested additions
can practically be performed in the human mind or with pen and paper by a trained medical coder or billing specialist reviewing a radiology report and determining whether sufficient documentation exists to support reimbursement. Therefore, the claim also recites mental processes.
Additionally, the recited operations of: predicting billing codes; ranking based on probabilities; scoring report completeness; determining completeness thresholds; and training predictive models, constitute mathematical concepts including statistical analysis, predictive modeling, probability calculations, classification, and machine-learning operations.
Accordingly, claim 1 recites abstract ideas in the form of:
(1) Certain methods of organizing human activity;
(2) Mental processes; and
(3) Mathematical concepts.
Therefore, claim 1 recites a judicial exception.
STEP 2A – PRONG TWO
The claim as a whole does not integrate the judicial exception into a practical application.
The additional elements recited in claim 1 include:
• a radiology workstation;
• at least one display device;
• at least one user input device;
• a processor;
• a database;
• radiology images; and
• an AI component.
These elements are recited at a high level of generality and merely serve as tools for collecting, processing, displaying, and storing information.
The displaying of radiology images constitutes insignificant data gathering and presentation activity because the claim does not alter image acquisition, image processing, image reconstruction, image compression, image rendering, or image-display technology.
The claim does not recite any improvement to:
• radiology imaging technology;
• image processing technology;
• PACS technology;
• workstation functionality;
• database functionality;
• networking technology;
• display technology; or
• computer functionality itself.
Rather, the radiology workstation merely serves as an environment in which the abstract billing-code analysis is performed.
The AI component is recited functionally and merely applies predictive analytics to evaluate report content and billing information. The claim does not recite any improvement to artificial intelligence technology itself.
Accordingly, the claim merely uses generic computer components as tools to automate billing-code prediction and documentation review.
The claim therefore does not integrate the judicial exception into a practical application.
STEP 2B
Claim 1 does not recite an inventive concept sufficient to transform the abstract idea into patent-eligible subject matter.
The additional elements recited in claim 1, individually and as an ordered combination, perform well-understood, routine, and conventional functions, including: receiving reports; displaying images; storing records; processing information; analyzing text; predicting classifications; ranking outputs; presenting suggestions; and storing results.
The processor, display device, user input device, database, and workstation are generic computer components performing their ordinary functions.
The recited AI component is used as a generic tool to perform predictive analysis. Merely applying machine-learning techniques to an abstract business or administrative problem does not provide an inventive concept absent a specific technological improvement.
Viewed as an ordered combination, the claim merely automates the longstanding practice of reviewing radiology reports, determining applicable billing codes, identifying missing documentation, and recommending additional information needed to support reimbursement.
Accordingly, claim 1 does not amount to significantly more than the judicial exception itself.
Claim 2 recites displaying indications of billing codes. Displaying information constitutes insignificant post-solution activity and does not provide an inventive concept.
Claim 3 recites receiving authorization and automatically adding suggested content. This limitation merely automates editing of documentation and remains within the abstract idea.
Claim 4 recites predicting and displaying top-K candidate billing codes and receiving user selections. Ranking and selecting candidates based on predictive scores constitutes additional mathematical analysis and mental evaluation.
Claim 5 recites training the AI component on historical radiology reports annotated with billing codes and completeness information. Training predictive models using historical examples constitutes a mathematical concept.
Claim 6 recites that the AI component comprises a BERT language model. Merely specifying a known machine-learning architecture as the predictive tool does not provide an inventive concept because the model is used according to its ordinary purpose.
Claims 7 and 16: These claims recite scoring the report as to completeness. Scoring information based on predictive analysis constitutes a mathematical concept.
Claims 8 and 17: These claims recite repeating completeness evaluation until a threshold is exceeded. Repeated application of mathematical analysis until a threshold condition is met remains a mathematical concept.
Claims 10 and 11: These claims recite displaying billing-code indications and adding suggested content. These limitations constitute insignificant extra-solution activity.
Claim 12 recites transmitting images and reports between a hospital and a teleradiology service via the Internet. The transmission of information using generic networking technology constitutes insignificant data transmission activity and does not improve networking technology.
These claims recite candidate billing-code selection and training on historical data, which remain mathematical concepts and mental processes.
Claim 15 recites use of a BERT model and therefore does not add significantly more than the abstract idea.
Claim 18 recites predicting billing codes, predicting missing content, and scoring report completeness. These limitations remain directed to mathematical analysis and administrative review.
Claim 19 recites substantially the same limitations as claim 1 in method form and therefore is directed to the same abstract idea.
Claim 20 recites training an AI model using datasets, feature vectors, model updating, and outputting whether missing content exists. These limitations are directed to mathematical concepts, machine-learning training, and predictive analytics and do not integrate the judicial exception into a practical application.
Claims 1-20 are directed to abstract ideas including medical billing analysis, reimbursement support, documentation review, predictive coding, report-completeness evaluation, mathematical modeling, and mental processes. The claims do not integrate the judicial exception into a practical application and do not recite additional elements sufficient to amount to significantly more than the judicial exception. Therefore, claims 1-20 are rejected under 35 U.S.C. § 101.
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.
Claim(s) 1-20 is/are rejected under 35 U.S.C. 103 as being unpatentable over Heinze et al. (US 6915254), in view of Syeda-Mahmood et al. (US 11,244,755).
With respect to claims 1, 18, 19 and 20, Heinze et al. (US 6,915,254 B1) disclose a computer-implemented medical coding system including workstations, processors, databases, and user interfaces configured to receive physician notes, process the notes using natural language processing, and automatically assign diagnosis, procedure, and billing-related medical codes. Heinze teaches that the system automatically determines and assigns medical codes from computer-readable physician notes for reimbursement purposes (Abstract; Fig. 1; col. 1, ll. 45-67).
More specifically, Heinze discloses:
a workstation including display and user interface components for review and modification of coding results (Fig. 1, coder review workstations 85, 99; Fig. 7; Fig. 8);
receiving a physician note/report as input data for processing (host site input data including transcribed physician notes and clinical information);
analyzing the report using NLP processing including segmentation, parsing, semantic analysis, vector processing, and code assignment to generate diagnosis, procedure, and E/M billing codes (Fig. 2; Fig. 3; col. 1, ll. 45-67; parsing discussion).
generating associated codes and presenting the generated codes and supporting information through a graphical interface for review and/or modification.
Accordingly, Heinze teaches:
“provide a radiology examination reading environment configured to display images of a radiology examination on the at least one display device”
insofar as Heinze teaches workstation-based review environments displaying medical coding information and physician report data for review. (Fig. 1, Fig. 7, Fig. 8).
Heinze further teaches:
“receive a radiology report for the radiology examination which is entered using the at least one user input device”
through receipt of transcribed physician notes and clinical information for automated coding.
Heinze additionally teaches:
“analyze the radiology report to predict one or more billing codes for the radiology examination”
because Heinze automatically assigns diagnosis codes, procedure codes, and E/M billing codes using NLP analysis of medical reports.
However, Heinze does not expressly disclose:
analyzing the radiology report using an AI component to identify missing content supporting predicted billing codes and ranking suggested missing content based on probability.
For these features, Syeda-Mahmood et al. (US 11,244,755 B1) teach machine-learning and deep-learning systems that process radiology reports and medical images using trained ML/DL models to generate report findings and identify report content based upon learned radiology-report patterns. The reference teaches training AI models on large corpora of radiology reports and generating report content from learned finding labels. (Abstract; Fig. 11; ML/DL training tool 1140; trained ML/DL models 1106).
Syeda-Mahmood further teaches:
generating radiology report content using trained machine-learning models;
extracting findings from radiology reports and associating report content with clinically relevant findings;
training AI models using historical radiology reports and labeled findings to determine report completeness and generate missing report language.
Therefore, Syeda-Mahmood teaches or at least suggests:
“analyze the radiology report to identify any missing content for supporting the one or more predicted billing codes that is missing from the radiology report using an artificial intelligence (AI) component”
by employing trained ML/DL models that infer report findings and generate report content from learned radiology-report patterns.
Syeda-Mahmood additionally suggests ranking candidate findings and report content outputs according to model prediction outputs and confidence values generated by trained ML models. The finding-label prediction output vectors and matching operations inherently provide probabilistic ranking of candidate report content. (Abstract; Fig. 11).
Accordingly, Syeda-Mahmood teaches or renders obvious:
“responsive to identifying more than one billing code from the radiology reports, ranking such missing billing codes based on a probability”
and
“display an indication of the missing content based on the ranking.”
The combined teachings of Heinze and Syeda-Mahmood therefore disclose all limitations of claim 1.
Therefore, it would have been obvious to one of ordinary skill in the art at the time of the invention to modify the automated medical coding system of Heinze with the machine-learning radiology-report generation and inference techniques of Syeda-Mahmood in order to improve coding accuracy, identify omissions in clinical documentation, reduce manual coder intervention, and improve completeness of reports supporting billing and reimbursement. Heinze expressly recognizes situations where documentation is incomplete and additional information is needed for accurate coding and even flags reports for further review or requests additional information.
A skilled artisan would have found it obvious to replace or augment Heinze’s manual review and additional-information prompting mechanisms with the trained AI inference techniques of Syeda-Mahmood to automatically predict missing report content and suggest additions, thereby improving efficiency and reducing human review burden. Such modification would merely apply known machine-learning techniques to improve a known medical coding system and would have yielded the predictable result of more complete and accurate coding support documentation.
With respect to claim 2, Heinze discloses that the generated diagnosis, procedure, and billing codes are presented to coder review workstations through a graphical user interface for review and modification (Fig. 1, coder review workstations 85, 99; Fig. 7; Fig. 8; generated codes displayed through GUI for review).
Accordingly, Heinze teaches:
"in response to identifying missing content, displaying an indication of the one or more billing codes."
The indication of predicted billing codes presented to the reviewer constitutes display of the billing codes associated with the report.
Therefore claim 2 would have been obvious over Heinze in view of Syeda-Mahmood.
With respect to claim 3, Heinze discloses coder review workstations permitting review and modification of generated coding results before final storage and billing submission (Fig. 7; Fig. 8; coded results exported after review).
Heinze therefore teaches:
"receiving an authorization to add the suggested addition via the at least one user input device"
through user review and approval of coding recommendations.
Heinze further teaches modification of coding information prior to final output and storage, thereby suggesting:
"adding the suggested addition to the radiology report to generate a complete radiology report."
Accordingly claim 3 would have been obvious over Heinze in view of Syeda-Mahmood.
With respect to claim 4, Heinze discloses generation of candidate diagnosis and procedure codes from physician notes and presentation of coding results to reviewers for selection and validation. (Fig. 4; vector matching and code resolution operations; coder review workstations).
Heinze therefore teaches or at least suggests:
"predicting a top-K number of candidate billing codes"
because multiple candidate diagnosis and procedure codes are generated during vector matching and code resolution.
Heinze further teaches:
"presenting, on the display device, the top-K candidate billing codes"
through coder review interfaces.
Heinze additionally teaches user interaction selecting appropriate codes before billing submission.
Therefore claim 4 would have been obvious.
Claim 5 recites:
"the predicting is performed by an artificial intelligence component trained on historical radiology reports annotated with billing codes and annotated as to completeness."
Heinze teaches prediction of billing codes from historical physician documentation using NLP processing.
However, Heinze does not expressly disclose machine-learning training on historical radiology reports.
Syeda-Mahmood teaches ML/DL model training using large corpora of historical radiology reports and labeled findings, where the reports are used to train predictive models that generate report content and findings (Fig. 11; ML/DL Training Tool 1140; Medical Imaging Reports Corpus 1150).
It would have been obvious to train Heinze's code-prediction system using historical radiology reports and associated coding information to improve prediction accuracy.
Therefore claim 5 would have been obvious.
Claim 6 recites:
"the AI component comprises a BERT language model."
Syeda-Mahmood teaches deep-learning architectures for processing radiology reports and generating findings from report corpora.
Although BERT is not expressly disclosed in the 2022 patent, transformer-based language models were well known in the radiology NLP art at the time of filing.
Selection of a BERT model represents a predictable implementation choice for the disclosed report-analysis AI system.
Accordingly claim 6 would have been obvious.
Claim 7 recites:
"scoring the radiology report as to a degree of completeness."
Heinze teaches assessing whether sufficient information exists to support assignment of diagnosis and procedure codes and flags documents when coding knowledge is incomplete or additional information is required.
Syeda-Mahmood teaches generation and evaluation of report findings using ML prediction outputs.
Together the references suggest determining a completeness score indicating whether sufficient report information exists to support coding.
Therefore claim 7 would have been obvious.
Claim 8 recites:
"determining a degree of completeness ... repeated until the determined degree exceeds a threshold."
Heinze teaches iterative review of flagged reports requiring additional information before coding completion.
Syeda-Mahmood teaches iterative machine-learning prediction and evaluation of report findings.
It would have been obvious to repeat completeness evaluation until a threshold confidence level is achieved.
Therefore claim 8 would have been obvious.
Claims 9-18 recite substantially the same subject matter as claims 1-8 in computer-readable-medium form.
The limitations of claims 9-18 are met by the same teachings discussed above for claims 1-8 because merely reciting the method on a non-transitory computer-readable medium does not patentably distinguish over the underlying obvious method.
It would have been obvious to one of ordinary skill in the art to modify the automated NLP-based billing code generation system of Heinze with the machine-learning radiology-report generation and inference techniques of Syeda-Mahmood in order to improve coding accuracy, automatically identify omissions in radiology reports, reduce manual coder workload, improve report completeness, and provide more reliable billing-support documentation. Heinze expressly recognizes incomplete documentation and the need for additional information to support coding decisions, while Syeda-Mahmood provides AI mechanisms for inferring and generating missing report content. Combining these known techniques would have yielded the predictable result of a more accurate and efficient report-completion and billing-support system.
Response to Arguments
Applicant’s arguments with respect to claim(s) 1-20 have been considered but are moot because the new ground of rejection does not rely on any reference applied in the prior rejection of record for any teaching or matter specifically challenged in the argument.
Conclusion
Any inquiry concerning this communication or earlier communications from the examiner should be directed to ROKIB MASUD whose telephone number is (571)270-5390. The examiner can normally be reached Mon-Fri 8:00-5:00.
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, Fahd Obeid can be reached at 571-270-3324. 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.
/ROKIB MASUD/ Primary Examiner, Art Unit 3627
Prosecution Insights