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 .
Priority
Receipt is acknowledged of certified copies of papers required by 37 CFR 1.55.
Information Disclosure Statement
The information disclosure statement filed May 27, 2026 fails to comply with 37 CFR 1.98(a)(2), which requires a legible copy of each cited foreign patent document; each non-patent literature publication or that portion which caused it to be listed; and all other information or that portion which caused it to be listed.
No copy of “RandECG: Data Automation for Deep Neural Network based ECG classification (2021)” has been received. It has been placed in the application file, but the information referred to therein has not been considered.
Response to Amendment
The amendment filed May 26, 2026 has been entered. Claims 1-20 remain pending in the application. Applicant’s amendments to the Specification, Drawings, and Claims have overcome each and every objection, and 112 rejection previously set forth in the Non-Final Office Action mailed Feb. 26, 2026.
Response to Arguments
Specification:
Applicant amended specification and addressed all previous objections and the previous objections have been withdrawn.
Claim Objections:
Applicant amended claims and addressed all previous objections and the previous objections have been withdrawn.
35 USC § 112:
Applicant amended claims and addressed all previous 35 USC 112 rejections and the previous rejections have been withdrawn.
New 112 rejections are set forth below.
35 USC § 101:
Applicant amended claim 14 and overcomes the 35 USC 101 rejection directed towards non-statutory subject matter. This rejection has been withdrawn.
Applicant's arguments on pages 8-14 filed May 26, 2026 have been fully considered but they are not persuasive.
On pages 10-13 of Applicant’s response, applicant argues that “The Federal Circuit has recognized that claims directed to improved cardiac monitoring and analysis may be patent-eligible when they use specific technological techniques to identify clinically relevant conditions. See CardioNet, LLC v. InfoBionic, Inc., 955 F.3d 1358 (Fed. Cir. 2020).” Examiner disagrees that this is pertinent to the instant application. To overcome the 101 rejection, the inventive concept would need to encompass structural components that are significantly more:
In our view, the claims “focus on a specific means or method that improves” cardiac monitoring technology; they are not “directed to a result or effect that itself is the abstract idea and merely invoke generic processes and machinery.” and “[W]hile the specification may help illuminate the true focus of a claim, when analyzing patent eligibility, reliance on the specification must always yield to the claim language in identifying that focus.” See CardioNet, LLC v. InfoBionic, Inc., 955 F.3d 1358 (Fed. Cir. 2020).
Further, Applicant argues that the claims provide significantly more than the abstract idea.
First, it is noted that while ECG data is being used as part of the evaluation using the neural network, ECG parts are not part of the claimed inventions. Further, MPEP 2106.05(a)(II) states the following:
However, it is important to keep in mind that an improvement in the abstract idea itself (e.g. a recited fundamental economic concept) is not an improvement in technology.
In this case, the stated improvement is in the use of an algorithm implemented on a computer and it appears that the improvement is in the abstract idea, which would not be considered an improvement in technology. Additionally, the additional non abstract components of the claims amount to no more than mere instructions to apply the judicial exception using a generic computer component.
Applicant argues that the claims are not limited to visible waveform review. The specification does not define visible waveforms and the device would require some sort of quantitative difference in order for a change to exist. Therefore, a human would be able to look at the quantitative change and make a diagnosis.
35 USC § 103:
Applicant amended claims and overcomes the previous independent claim rejections with respect to Lee, Deepika, and Yang. Examiner agrees that all of the amended claim limitations are not taught by the cited references. A new 35 USC 103 rejection below is being applied using Lee, Deepika, and Yang in light of the amendments to the claims.
Claim Rejections - 35 USC § 112
The following is a quotation of the first paragraph of 35 U.S.C. 112(a):
(a) IN GENERAL.—The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor or joint inventor of carrying out the invention.
The following is a quotation of the first paragraph of pre-AIA 35 U.S.C. 112:
The specification shall contain a written description of the invention, and of the manner and process of making and using it, in such full, clear, concise, and exact terms as to enable any person skilled in the art to which it pertains, or with which it is most nearly connected, to make and use the same, and shall set forth the best mode contemplated by the inventor of carrying out his invention.
Claim 1-20 are rejected under 35 U.S.C. 112(a) or 35 U.S.C. 112 (pre-AIA ), first paragraph, as failing to comply with the written description requirement. The claim(s) contains subject matter which was not described in the specification in such a way as to reasonably convey to one skilled in the relevant art that the inventor or a joint inventor, or for applications subject to pre-AIA 35 U.S.C. 112, the inventor(s), at the time the application was filed, had possession of the claimed invention.
Claims 1, 14, and 15 recites the amended limitations:
“Visibly interpretable and non-visibly interpretable changes”. As-filed specification does not provide any description of “visibly interpretable and non-visibly interpretable changes”. As such, the specification does not provide an adequate written description of the claimed invention since as-filed specification does not describe any species that would be included in the claimed genus, “visibly interpretable and non-visibly interpretable.”
“Changes in one more of a shape”. As-filed specification discloses shapes of several waves in Fig. 1 and describes “normal shapes” in the specification. However, as-filed specification does not provide any description of using changes in the shape of the wave to train the neural network. As such, the specification does not provide an adequate written description of the claimed invention since as-filed specification disclosure fails to sufficiently identify how the function is performed or the result is achieved using changes in shape of the waveform.
“Changes in one or more of… magnitude”. As-filed specification does not provide any description of “magnitude”. As such, the specification does not provide an adequate written description of the claimed invention since as-filed specification does not describe any species that would be included in the claimed genus, “magnitude.”
Applicant’s amendment, filed 5/26/2026, asserts that no new matter has been added. However, the specification as-filed does not provide a written description or set forth the metes and bounds of the amended limitations above. The specification does not provide blazemarks nor direction for the instant methods encompassing the above-mentioned "limitations" as they are currently recited. The instant claims now recite limitations which were not clearly disclosed in the specification as-filed, and now change the scope of the instant disclosure as-filed. Such limitations recited in the present claims, which did not appear in the specification, as-filed, introduce new concepts (See MPEP 2163.06). Applicant is required to cancel the new matter in the response to this Office action. Alternatively, applicant is invited to provide sufficient written support for the “limitations” indicated above.
Claims 2-13, and 16-20 depend from rejected claims 1, 14, and 15, and are therefore rejected for the same reasons as above.
The following is a quotation of 35 U.S.C. 112(b):
(b) CONCLUSION.—The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the inventor or a joint inventor regards as the invention.
The following is a quotation of 35 U.S.C. 112 (pre-AIA ), second paragraph:
The specification shall conclude with one or more claims particularly pointing out and distinctly claiming the subject matter which the applicant regards as his invention.
Claims 1-20 are rejected under 35 U.S.C. 112(b) or 35 U.S.C. 112 (pre-AIA ), second paragraph, as being indefinite for failing to particularly point out and distinctly claim the subject matter which the inventor or a joint inventor (or for applications subject to pre-AIA 35 U.S.C. 112, the applicant), regards as the invention.
Regarding claims 1, 14, and 15, it is unclear what “non-visibly interpretable change” is defined as. In order for the morphological change to be interpretable, there would need to be some kind of identifiable change, even if the change is miniscule, a quantitative representation of the data would be visibly changing. For the purposes of examination, “non-visibly interpretable change” will be interpreted to include any quantitative measurement that a machine is able to detect.
Regarding claims 1, 14, and 15, it is unclear what changes in “shape” is defined as. The specification discloses the concept of a shape but does not elaborate what a change in shape consists of. Changing wave magnitude and duration would both satisfy a change in shape.
Claims 2-13 and 16-20 are dependent on claims 1, 14, and 15 and are therefore rejected for the same reasons as above.
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.
Claim 1-20 are rejected under 35 U.S.C. 101 because the claimed invention is directed to the abstract ideas of “aiding diagnosis of thyroid dysfunction of a subject based on analysis of morphological changes in electrocardiogram data”, “assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device”, “inputting the single-lead electrocardiogram data into a pre-trained neural network model to determine a probability of occurrence of thyroid dysfunction for the subject”, and “wherein the pre-trained neural network model is trained based on correlations between thyroid function and morphological changes in electrocardiograms wherein the morphological changes in the plurality of single-lead electrocardiograms comprise visibly interpretable and non-visibly interpretable changes in one or more of a shape, magnitude or duration of a wave, and wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects” without significantly more.
Step 1:
Claims 1-13 recite a method and claims 15 and 19-20 recites a computing device, a machine. Claims 14 and 16-18 recite a system. Therefore, the claims fall within the statutory categories.
Step 2A, Prong 1:
Claims 1, 14, and 15 recite limitations of aiding diagnosis of thyroid dysfunction based on analysis of morphological changes in electrocardiogram data, assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device, inputting the single-lead electrocardiogram data into a pre-trained neural network model to determine a probability of occurrence of thyroid dysfunction for a subject, and wherein the pre-trained neural network model is trained based on correlations between thyroid function and changes in electrocardiograms wherein the morphological changes in the plurality of single-lead electrocardiograms comprise visibly interpretable and non-visibly interpretable changes in one or more of a shape, magnitude or duration of a wave, and wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects. The limitations, as drafted, describe a process that, under its broadest reasonable interpretation, includes performance of the limitation in the mind and using mathematical equations except for the claim 1 recitation of “computing device including at least one processor”, the claim 14 recitation of “one or more non-transitory readable media storing instructions” and “one or more processors”, and the claim 15 recitation of “, “at least one processor including at least one core”, “at least one operably coupled to the at least one processor” and “computer-executable instructions stored in the at least one memory”. That is, other than reciting that a system is performing these tasks, nothing in the claims precludes the steps from practically being performed in the human mind. MPEP 2106.04(a)(2)(III) states that the courts consider a mental process (thinking) that "can be performed in the human mind, or by a human using a pen and paper" to be an abstract idea. For example, aside from the recitation of “processor”, and “memory” language, the claims encompasses the user obtaining ECG waveform data, comparing it to patterns in thyroid functions against changes in ECG characteristics, and determining the likelihood of a thyroid dysfunction. These limitations are a mental process. Further, the recitation of training a neural network to determine a probability of occurrence of thyroid dysfunction and training the neural network based on correlations are recitations of math as they fundamentally require statistical analysis.
Step 2A, Prong 2:
The claims recite “computing device including at least one processor”, “one or more non-transitory readable media storing instructions” “one or more processors”, “at least one processor including at least one core”, “at least one operably coupled to the at least one processor” and “computer-executable instructions stored in the at least one memory” to perform the abstract steps. These components read on a computer implemented system and are recited at a high level of generality, i.e., as a generic processor, performing a generic computer function of processing data (see in ¶[0045], ¶[0049]-¶[0051], and ¶[0055] of the specification). This generic processor limitation is no more than mere instructions to apply the exception using a generic computer component. Accordingly, this additional limitation does not integrate the abstract idea into a practical application because it does not impose any meaningful limits on practicing the abstract idea.
Step 2B:
As discussed with respect to Step 2A Prong Two, the additional elements in the claim amount to no more than mere instructions to apply the exception using a generic computer component. The same analysis applies here in 2B, i.e., mere instructions to apply an exception on a generic computer cannot integrate a judicial except into a practical application at Step 2A or provide an inventive concept in Step 2B.
Under 2019 PEG, a conclusion that an additional element is insignificant extra-solution activity in Step 2A should be re-evaluated in Step 2B to determine if it is more than what is well- understood, routine, conventional activity in the field. The specification in ¶[0045], ¶[0049]-¶[0051], and ¶[0055] does not provide any indication that the computer and computer programming is anything other than a generic, off-the-shelf computer component. Court decisions cited in MPEP 2106.05(d)(II) indicate that computer‐ implemented processes not to be significantly more than an abstract idea (and thus ineligible) where the claim, as a whole, amounts to nothing more than generic computer functions merely used to implement an abstract idea, such as an idea that could be done by a human analog (i.e., by hand or by merely thinking). Accordingly, a conclusion that the generic computer functions merely being used to implement an abstract idea is well-understood, routine, conventional activity is supported under Berkheimer Option 2.
Dependent claims 2-13 and 16-20 further limit the process of aiding diagnosis of thyroid dysfunction of a subject based on analysis of morphological changes in electrocardiogram data, assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device, inputting the single-lead electrocardiogram data into a pre-trained neural network model to determine a probability of occurrence of thyroid dysfunction for the subject, and wherein the pre-trained neural network model is trained based on correlations between thyroid function and changes in electrocardiograms wherein the morphological changes in the plurality of single-lead electrocardiograms comprise visibly interpretable and non-visibly interpretable changes in one or more of a shape, magnitude or duration of a wave, and wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects. Therefore, these claims further limit the abstract idea already indicated in independent claim 1 and they are ineligible for the same reasons provided for claim 1 above.
For these reasons, there is no inventive concept in the claims and thus they are ineligible.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
The factual inquiries for establishing a background for determining obviousness under 35 U.S.C. 103 are summarized as follows:
1. Determining the scope and contents of the prior art.
2. Ascertaining the differences between the prior art and the claims at issue.
3. Resolving the level of ordinary skill in the pertinent art.
4. Considering objective evidence present in the application indicating obviousness or nonobviousness.
Claims 1-2, 8, 13-14, and 16-18 are rejected under 35 U.S.C. 103 as being unpatentable over Lee et al. (KR1020060117546, published Nov. 17, 2006, hereinafter referred to as “Lee”), Deepika et al. (EAI Endorsed Transactions on Energy Web, Aug. 26, 2020, Vol. 8, Issue 32, cited in IDS filed on Mar. 15, 2024 and hereinafter referred to as “Deepika”), Yang et al. (US 20190090774 A1, published Mar. 28, 2019, hereinafter referred to as “Yang”), Witvliet et al. (Journal of Electrocardiology, 2021, Vol. 66, pg 33-37, available online on Mar. 4, 2021), and Attia et al. (WO 2019070978 A1, published Apr. 11, 2019, hereinafter referred to as “Attia”).
Regarding claims 1 and 14, Lee teaches a method of aiding diagnosis of a disease of a subject based on analysis of morphological changes in electrocardiogram data (Fig. 2 “A control unit (500) that determines whether a user has a disease and the type of disease based on the user's electrocardiogram signal” in ¶[0021] and “the features for disease judgment are extracted using the morphological features of the electrocardiogram signal” ¶[0017]), the method comprising: assessing electrocardiogram data (Fig. 2, “and input unit (100) that receives the user’s electrocardiogram signal” in ¶[0021]); wherein the pre-trained neural network model is trained based on the disease and electrocardiograms (Fig. 2 “The feature values extracted from the above feature extraction unit (300) are trained through a neural network to determine a disease diagnosis for the input electrocardiogram signal” ¶[0021]).
Lee does not disclose the method of aiding diagnosis of thyroid dysfunction based on analysis of electrocardiogram data, the method being performed by a computing device including at least one processor, assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device and inputting the single-lead electrocardiogram data into a pre-trained neural network model to determine a probability of occurrence of thyroid dysfunction for the subject; wherein the neural network model is trained based on correlations between thyroid function and changes in electrocardiogram characteristics. Further, Lee does not explicitly teach the pre-trained neural network model is trained based on the disease and morphological changes in a plurality of electrocardiograms wherein the morphological changes in the plurality of single-lead electrocardiograms comprise visibly interpretable and non-visibly interpretable changes in one or more of a shape, magnitude or duration of a wave, and wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Deepika teaches ECG parameters in relation to thyroid dysfunction and this analysis is used to assess the relationship among outright thyroxine (T4) and thyrotropin (TSH) levels with ECG parameters (beat, PR intervals, QRS term, QT between times, and JT between times) (pg. 2). The process involved training using TSH and FT4 levels as well as the parameters in Table 2 which correspond to electrocardiogram characteristics (pg. 4 section 3.3.3). The classification technique is used to categorize the hypothyroid and hyperthyroid illnesses. The efficiency of classifiers is evaluated using the confusion matrix on the subjects of accuracy, sensitivity, and specificity. The C4.5 algorithm gives 93.58% which is providing better accuracy than determination stump tree accuracy and also C4.5 Algorithm provides very minimum error rate than decision stump (pg. 9). Deepika also teaches that people with clear hypothyroidism shows variations in ECG like sinus bradycardia, moo plentifulness QRS buildings, QT interval prolongation, and the changes in T wave morphology. Strikingly, both deferred and condensed QT between times has been represented in hyperthyroidism (pg. 2). Quick pulse detection is one
of the most widely recognized indications of hyperthyroidism (pg. 4).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to aid in diagnosing thyroid dysfunction based on the electrocardiogram, estimate the probability of occurrence of the thyroid dysfunction using the pre-trained neural network, and train the neural network based on correlations between thyroid function and changes in ECG as taught by Deepika in the method of Lee. This is obvious because the device in Lee is capable of learning the material in Deepika and it is known that thyroid disorders affect ECGs. Further, the system is not 100% accurate so it would be obvious to provide a probability of occurrence since the medical diagnosis is uncertain and this allows for optimization of medical resources.
Even further, it would have been obvious to a person having ordinary skill in the art at the time of filing to pre-train the neural network based on changes in morphology including changes in duration (inherently both visibly and non-visibly interpretable) as taught in the method of Lee because patients with hypothyroidism and hyperthyroidism show changes in T wave morphology and differences in QT interval prolongation as taught by Deepika.
Lee and Deepika do not disclose the method being performed by a computing device including at least one processor and assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device. Further, Lee and Deepika do not teach wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Yang’s invention relates to systems and methods for localizing activity in cardiac tissue using data acquired non-invasively via electrocardiography (¶[0003]). The computing device 710 in Fig. 7 may include one or more processors and memory for storing instructions that are executed by the one or more processors to provide the functionality being discussed (¶[0062]). The device consists of a processor and memory having instructions that, when executed by the processor, are configured to: use the electrocardiography device to record electrical activity in the heart of a subject; feed the electrical data to one or more neural networks; and receive from the one or more neural networks an identification of a segment of the heart at which an arrhythmia originates (¶[0013]).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to implement the method of Lee and Deepika onto a computing device including at least one processor as taught by Yang. This is obvious because it provides a medium for the neural network to work.
Lee and Deepika do not teach the claim 14 limitation of one or more non-transitory computer readable media storing instructions, performing operations based on an electrocardiogram when executed on one or more processors.
Yang teaches computerized functions may be implemented as instructions stored using one or more machine-readable media, which may be read and executed by one or more processors. A machine-readable medium may include any mechanism for storing or transmitting information in a form readable by a machine (¶[0067]). The device consists of a processor and memory having instructions that, when executed by the processor, are configured to: use the electrocardiography device to record electrical activity in the heart of a subject; feed the electrical data to one or more neural networks; and receive from the one or more neural networks an identification of a segment of the heart at which an arrhythmia originates (¶[0013]).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to implement the device of Lee and Deepika with a computer program stored in a computer-readable storage medium as taught by Yang. This is obvious because it provides a medium for the neural network to work.
Lee, Deepika, and Yang do not disclose the method assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device. Further, Lee, Deepika, and Yang do not teach wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Witvliet’s study relates to an overview of the usefulness and potential pitfalls when implementing 1 L-ECGs into everyday clinical practice. The study teaches that single-lead electrocardiograms are increasingly used in (pre)clinical settings for the detection and monitoring of a range of rhythm and conduction disorders (abstract). The intended use of 1 L-ECGs has been detecting AF and it has been validated for this purpose. 1 L-ECGs can be a less time-consuming alternative for 12 L-ECG (pg. 33-34).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to train the neural network based on ECG data measured with single leads as taught by Witvliet in the method of Lee, Deepika, and Yang because one lead ECG is capable of detecting Afib and is becoming a more common and less time consuming alternative for 12 lead ECG.
Further, Lee, Deepika, Yang, and Witvliet do not teach wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Attia’s invention relates to computer-based technology for analyzing physiological electrical data (e.g., electrocardiogram data) (¶[2]). For the purposes of this specification, an electrocardiogram may include the traditional 12 leads, additional leads, or any number of leads to as few as a single lead (¶[6]). Although a number of examples are described herein for detecting using data from a standard 12-lead ECG, in other implementations fewer leads— including even a single-lead— ECG, can be used to effectively detect instances or individuals who are susceptible to an increased risk within a future period of time (¶[73]). The study obtained data from the MAYO CLINIC digital data vault. 163,892 adult patients (18 years or older) were identified with at least one digital, standard 10-second 12-lead ECG acquired in the supine position between January 1994 and February 2017 and at least one transthoracic echocardiogram (TTE) obtained within 14 days of the index ECG (Figure 13). For patients with multiple ECG and TTE data sets meeting these criteria, the earliest pair was used for network creation, validation, or testing, and subsequent TTE data used for analysis of follow up (¶[101]). The study involved developing a convolutional neural network (CNN). CNNs, which have been applied to images (or videos), operate such that the convolutions can be used to extract very subtle patterns in a data set. Each 12-lead ECG was considered a 12 x 5000 (i.e., 12 leads by 10 seconds duration sampled at 500 Hz) "image". The network was composed of N single lead convolutional layers, each of which was followed by a non-linear "Relu" activation function, a batch- normalization layer and a max pooling layer. The features extracted from each raw, digital signal ECG lead were fused in another convolutional layer that had access to all leads simultaneously. Following the last convolutional layer, the data were fed to a fully connected network with two hidden layers with dropout layers to avoid overfitting and an output layer that was activated using the "Softmax" function (¶[104]).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to train the neural network with a plurality of single-lead electrocardiograms that are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects as taught by Attia in the method of Lee, Deepika, Yang, and Witvliet as it is a results-effective method of training a neural network that produces accurate results and can be used to interpret ECG data from any number of leads.
Regarding claim 2, Lee discloses wherein the pre-trained neural network model includes a first sub-neural network that has been trained based on electrocardiograms measured with 12 leads (¶[0006-0008] and neural networks are composed of sub-neural networks).
Deepika further teaches the 12-lead electrocardiograph as being the standard measurement (pg. 2).
Regarding claim 8, Lee teaches wherein the electrocardiogram characteristics include at least one of a length of a QT interval, and QRS duration (“Five features, including the heart rate, QRS duration, PR interval, QT interval, and T wave type, are input into the neural network” in ¶[0056]).
Regarding claims 13 and 17, Lee does not teach wherein the pre-trained neural network model is further trained on biological data including at least one of age and gender.
Deepika teaches that Subclinical thyroid dysfunction is defined as a typical condition, where serum thyrotrophic hormone (TSH) levels is found to be below (hyperthyroidism) or above (hypothyroidism) the reference interval with traditional free thyroxin (FT4) levels. It is commonly prevailing in older (age) girls (gender) (pg. 1).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to include age or gender as additional inputs in the neural network as taught by Deepika in the method of Lee since thyroid dysfunction is more common in older women and can then be used to use to strengthen the probability measurement.
Regarding claim 16, Deepika teaches wherein the operations further comprise estimating a progress of thyroid dysfunction based on an output of the pre-trained neural network model (fig. 8 “data classification are into two groups for the dataset namely hypothyroid and hyperthyroid. The blue colour star shows the hyperthyroid and black colour circle shows the hypothyroid data set using trust region method for the optimization algorithm.” Pg. 9).
Regarding claim 18, Deepika teaches wherein the thyroid dysfunction is hyperthyroidism (“The classification technique is used to categorize the hypothyroid and hyperthyroid illnesses” pg. 9).
Claims 3-4 are rejected under 35 U.S.C. 103 as being unpatentable over Lee, Deepika, Yang, Witvliet, and Attia (hereinafter “modified Lee”) as applied to claim 2 above, in further view of Wellens (The New England Journal of Medicine, Feb. 4, 1999, Vol. 340, No. 5., pg. 381, hereinafter referred to as “Wellens”).
Regarding claim 3, modified Lee teaches the method of claim 2.
Although Lee teaches six limb leads and six precordial leads, modified Lee does not disclose wherein the pre-trained neural network model includes a second sub-neural network that has been trained based on electrocardiograms measured with at least six limb leads or six precordial leads.
Wellens’ article discusses the value of precordial leads in an electrocardiogram. The article teaches that traditionally, 12 leads are recorded, 6 leads on the extremities (limb leads) and 6 on the precordium. This has been the standard approach for almost half a century. The extremity leads give a more distant image of the electrical activity of the heart. For example, leads II and III record electrical activity from the inferior wall. The precordial leads, because they are unipolar and closer to the heart, primarily reflect the cardiac electrical activity directly beneath the electrode. Therefore, because of their position on the chest wall leads V2 to V6 primarily give information on the process of activation and repolarization of the anterior and lateral aspects of the left ventricle. Lead V1 provides information on the interventricular septum and the superior part of the right ventricle (pg. 381).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to train the neural network model with data measured with at least six limb leads or six precordial leads on as taught by Wellens in the method of modified Lee in order to obtain signals related to different positions of the heart.
Regarding claim 4, modified Lee does not disclose wherein the pre-trained neural network model includes a third sub-neural network that has been trained based on electrocardiograms measured with single leads.
Witvliet’s study teaches that single-lead electrocardiograms are increasingly used in (pre)clinical settings for the detection and monitoring of a range of rhythm and conduction disorders (abstract). The intended use of 1 L-ECGs has been detecting AF and it has been validated for this purpose. 1 L-ECGs can be a less time-consuming alternative for 12 L-ECG (pg. 33-34).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to train the neural network based on ECG data measured with single leads as taught by Witvliet in the method of modified Lee because one lead ECG is capable of detecting Afib and is becoming a more common and less time consuming alternative for 12 lead ECG.
Claims 5-6 are rejected under 35 U.S.C. 103 as being unpatentable over modified Lee as applied to claim 1 above, in further view of Oommen (Towards Data Science, Nov. 28, 2020, hereinafter referred to as “Oommen”) and the Institute of Medicine (Medicare Coverage of Routine Screening for Thyroid Dysfunction, 2003, Washington, DC: The National Academies Press. Pg. 21, hereinafter “Institute of Medicine”).
Regarding claim 5, modified Lee discloses the method of claim 1.
Although Deepika’s study assesses the relationship among outright thyroxine (T4) and thyrotropin (TSH) levels with ECG parameters in order to classify subclinical thyroid dysfunction (pg.1 and 3), modified Lee does not disclose wherein: the pre-trained neural network model includes a neural network including a plurality of residual blocks, and wherein the neural network including the residual blocks receives the electrocardiogram data and outputs a probability of occurrence of overt hyperthyroidism.
Oommen teaches the benefits of residual blocks in deep neural networks. Residual blocks create an identity mapping to activations earlier in the network to thwart the performance degradation problem associated with deep neural architectures (Oommen).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to have a neural network comprising residual blocks as taught by Oommen in the method of modified Lee in order to combat performance degradation seen in deep neural networks.
Modified Lee and Oommen (hereinafter “modified Lee”) do not disclose the neural network including the residual blocks receives the electrocardiogram data and outputs a probability of occurrence of overt hyperthyroidism.
Institute of medicine teaches that overt hyperthyroidism is defined by a low serum TSH concentration and a high serum free T4 concentration (Institute of Medicine pg. 22).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to output the probability of occurrence of overt hyperthyroidism in the method of modified Lee because modified Lee’s device is capable of diagnosing using the TSH and T4 concentration conditions of subclinical hyperthyroidism and would then be able to apply the same concept to overt hyperthyroidism.
Regarding claim 6, modified Lee and Oommen inherently teach wherein the overt hyperthyroidism corresponds to a case where a free thyroxine level is higher than a predetermined reference range or a case where a thyroid-stimulating hormone level is lower than a reference range.
Overt hyperthyroidism is defined by a low serum TSH concentration (thyroid-stimulating hormone level is lower than a reference range) and a high serum free T4 concentration (free thyroxine level is higher than a predetermined reference range) (Institute of Medicine pg. 22).
Claim 7 is rejected under 35 U.S.C. 103 as being unpatentable over modified Lee as applied to claim 1 above, in further view of Zhang et al. (Zhang, J. et al. “MLBF-Net: A Multi-Lead-Branch Fusion Network for Muti-Class Arrythmia Classification Using 12-Lead ECG”. Cardiovascular Devices and Systems. IEEE Journal of Transitional Engineering in Health and Medicine. 15 Mar. 2021. Vol.9, 2021. Doi: 10.1109/JTEHM.2021.3064675, hereinafter referred to as “Zhang”).
Regarding claim 7, modified Lee teaches the method of claim 1 including the probability of occurrence of thyroid dysfunction as shown in claim 1.
Modified Lee does not disclose wherein the pre-trained neural network model includes a plurality of sub-neural networks corresponding to a plurality of respective leads of the electrocardiogram data, and wherein outputs of the plurality of sub-neural networks are concatenated into one output to derive the probability of occurrence of thyroid dysfunction.
Zhang’s study relates to deep neural networks and automatic arrhythmia detection using 12-lead electrocardiograms. The study teaches that in the previous studies on automatic arrhythmia detection, most methods concatenated 12 leads of ECG into a matrix (neural networks corresponding to a plurality of respective leads of the electrocardiogram data; and outputs of the sub-neural networks are concatenated, neural networks inherently have sub-neural networks), and then input the matrix to a variety of feature extractors or deep neural networks for extracting useful information (abstract).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to include neural networks corresponding to a plurality of respective leads as taught by Zhang in the method of modified Lee in order to extract useful information.
Claims 9-10 are rejected under 35 U.S.C. 103 as being unpatentable over modified Lee as applied to claim 8 above, in further view of Klein et al. (Klein, I. et al. “Thyroid Disease and the Heart”. Circulation. 09 Oct. 2007. Vol.116:15, p.1725-1734. https://doi.org/10.1161/CIRCULATIONAHA.106.678326 , hereinafter referred to as “Klein”).
Regarding claim 9, modified Lee teaches the method of claim 8.
Although Lee teaches sinus tachycardia may be caused by hyperthyroidism (¶[0023]), modified Lee does not disclose wherein the probability of occurrence of thyroid dysfunction increases as the frequency of tachycardia increases.
Klein’s study relates to cardiovascular signals in relation to thyroid disease. The study teaches sinus tachycardia is the most common rhythm disturbance and is recorded in almost all patients with hyperthyroidism. An increase in resting heart rate is characteristic of this disease (pg. 1730).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to increase the probability of occurrence of thyroid dysfunction as the frequency of tachycardia increases as taught by Klein in the method of modified Lee because tachycardia is recorded in almost all patients with hyperthyroidism so the increase in tachycardia frequency may signify a thyroid dysfunction.
Regarding claim 10, modified Lee teaches the method of claim 8.
Although Lee teaches the method including extracting information based on the QT interval and using it for disease classification (¶[0016] and claim 11), modified Lee does not teach wherein the probability of occurrence of thyroid dysfunction increases as the length of the QT interval increases.
Klein teaches a variety of case reports have demonstrated that hypothyroidism may cause a prolongation of the QT interval that predisposes the patient to ventricular irritability (pg. 1731).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to increase the probability of occurrence of thyroid dysfunction as the length of the QT interval increases as taught by Klein in the method of modified Lee because hypothyroidism may cause a prolongation of the QT interval so the presence of a longer QT interval may signify a thyroid dysfunction.
Claim 12 is rejected under 35 U.S.C. 103 as being unpatentable over modified Lee as applied to claim 8 above, in further view of Tayal et al. (Tayal, B. et al. “Thyroid dysfunction and electrocardiographic changes in subjects without arrhythmias: a cross-sectional study of primary healthcare subjects from Copenhagen”. BMJ Open. 2019;9:e023854. doi:10.1136/ bmjopen-2018-023854, hereinafter referred to as “Tayal”).
Regarding claim 12, modified Lee teaches the method of claim 8.
Modified Lee does not disclose wherein the probability of occurrence of thyroid dysfunction increases as the QRS duration becomes shorter.
Tayal’s study relates to investigating associations of both overt and subclinical thyroid dysfunction with common ECG parameters in a large primary healthcare population. In the study, overt hyperthyroid subjects had a shorter QRS duration (88.5±9.6) in comparison with the euthyroid (93.5±10.5 ms) subjects (pg. 4).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to increase the probability of occurrence of thyroid dysfunction as the QRS duration becomes shorter as taught by Tayal in the method of modified Lee because overt hyperthyroid subjects had a shorter QRS duration so the shortening of the QRS duration may indicate a thyroid dysfunction.
Claim 15 is rejected under 35 U.S.C. 103 as being unpatentable over modified Lee, as evidenced by Silberschatz et al. (Silberschatz, A. et al. “Operating System Concepts” 4 May 2018. ISBN 9781119124894, pg. 18, hereinafter referred to as “Silberschatz”).
Regarding claim 15, Lee teaches a system for aiding diagnosis of a disease based on analysis of morphological changes in electrocardiogram data (Fig. 2 “A control unit (500) that determines whether a user has a disease and the type of disease based on the user's electrocardiogram signal” in ¶[0021] and “the features for disease judgment are extracted using the morphological features of the electrocardiogram signal” ¶[0017]), and accessing electrocardiogram data (Fig. 2, “and input unit (100) that receives the user’s electrocardiogram signal” in ¶[0021]); wherein the pre-trained neural network model is trained based on the disease and electrocardiograms Fig. 2 “The feature values extracted from the above feature extraction unit (300) are trained through a neural network to determine a disease diagnosis for the input electrocardiogram signal” ¶[0021]).
Lee does not disclose a system for aiding diagnosis of thyroid dysfunction based on analysis of morphological changes in electrocardiogram data, the system comprising: at least one processor including at least one core; at least one memory operably coupled to the at least one processor; and computer-executable instructions stored in the at least one memory, wherein the computer- executable instructions, when executed by the at least one processor cause the system to perform operations; assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device, inputting the single-lead electrocardiogram data into a pre-trained neural network model, and estimating a probability of occurrence of thyroid dysfunction for the subject based on an output of the pre-trained neural network model. Further, Lee does not explicitly teach wherein the pre-trained neural network model is trained on correlations between thyroid function and morphological changes in a plurality of single-lead electrocardiograms, wherein the morphological changes in the plurality of single-lead electrocardiograms comprise visibly interpretable and non-visibly interpretable changes in one or more of a shape, magnitude or duration of a wave, and wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Deepika teaches ECG parameters in relation to thyroid dysfunction and this analysis is used to assess the relationship among outright thyroxine (T4) and thyrotropin (TSH) levels with ECG parameters (beat, PR intervals, QRS term, QT between times, and JT between times) (pg. 2). The process involved training using TSH and FT4 levels as well as the parameters in Table 2 which correspond to electrocardiogram characteristics (pg. 4 section 3.3.3). The classification technique is used to categorize the hypothyroid and hyperthyroid illnesses. The efficiency of classifiers is evaluated using the confusion matrix on the subjects of accuracy, sensitivity, and specificity. The C4.5 algorithm gives 93.58% which is providing better accuracy than determination stump tree accuracy and also C4.5 Algorithm provides very minimum error rate than decision stump (pg. 9). Deepika also teaches that people with clear hypothyroidism shows variations in ECG like sinus bradycardia, moo plentifulness QRS buildings, QT interval prolongation, and the changes in T wave morphology. Strikingly, both deferred and condensed QT between times has been represented in hyperthyroidism (pg. 2). Quick pulse detection is one
of the most widely recognized indications of hyperthyroidism (pg. 4).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to diagnose thyroid dysfunction based on the electrocardiogram, estimate the probability of occurrence of the thyroid dysfunction using the pre-trained neural network, and train the neural network based on correlations between thyroid function and changes in ECG as taught by Deepika in the method of Lee. This is obvious because device in Lee is capable of learning the material in Deepika and it is known that thyroid disorders affect ECGs. Further, the system is not 100% accurate so it would be obvious to provide a probability of occurrence since the medical diagnosis is uncertain and this allows for optimization of medical resources.
Even further, it would have been obvious to a person having ordinary skill in the art at the time of filing to pre-train the neural network based on changes in morphology including changes in duration (inherently both visibly and non-visibly interpretable) as taught in the device of Lee because patients with hypothyroidism and hyperthyroidism show changes in T wave morphology and differences in QT interval prolongation as taught by Deepika.
Lee and Deepika do not disclose the computing device comprising: a processor including at least one core; and memory including program codes that are executable on the processor and assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device. Further, Lee and Deepika do not teach wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Yang teaches the computing device 710 in Fig. 7 may include one or more processors and memory for storing instructions that are executed by the one or more processors to provide the functionality being discussed (¶[0062]). The device consists of a processor and memory having instructions that, when executed by the processor, are configured to: use the electrocardiography device to record electrical activity in the heart of a subject; feed the electrical data to one or more neural networks; and receive from the one or more neural networks an identification of a segment of the heart at which an arrhythmia originates (¶[0013]). Further, processors inherently contain at least one core as evidenced in Silberschatz “processor – a physical chip that contains one or more CPUs” and “core – the basic computation unit of the CPU” (Sillerschatz pg. 18).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to implement the method of Lee and Deepika onto a computing device including at least one processor with at least one core and a memory with program codes as taught by Yang and evidenced by Sillerschatz. This is obvious because it provides a medium for the neural network to work.
Lee and Deepika do not disclose assessing single-lead electrocardiogram data measured by a single-lead electrocardiogram measurement device. Further, Lee and Deepika do not teach wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Witvliet’s study relates to an overview of the usefulness and potential pitfalls when implementing 1 L-ECGs into everyday clinical practice. The study teaches that single-lead electrocardiograms are increasingly used in (pre)clinical settings for the detection and monitoring of a range of rhythm and conduction disorders (abstract). The intended use of 1 L-ECGs has been detecting AF and it has been validated for this purpose. 1 L-ECGs can be a less time-consuming alternative for 12 L-ECG (pg. 33-34).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to train the neural network based on ECG data measured with single leads as taught by Witvliet in the device of Lee, Deepika, and Yang because one lead ECG is capable of detecting Afib and is becoming a more common and less time consuming alternative for 12 lead ECG.
Further, Lee, Deepika, Yang, and Witvliet do not teach wherein the plurality of single-lead electrocardiograms are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects.
Attia’s invention relates to computer-based technology for analyzing physiological electrical data (e.g., electrocardiogram data) (¶[2]). For the purposes of this specification, an electrocardiogram may include the traditional 12 leads, additional leads, or any number of leads to as few as a single lead (¶[6]). Although a number of examples are described herein for detecting using data from a standard 12-lead ECG, in other implementations fewer leads— including even a single-lead— ECG, can be used to effectively detect instances or individuals who are susceptible to an increased risk within a future period of time (¶[73]). The study obtained data from the MAYO CLINIC digital data vault. 163,892 adult patients (18 years or older) were identified with at least one digital, standard 10-second 12-lead ECG acquired in the supine position between January 1994 and February 2017 and at least one transthoracic echocardiogram (TTE) obtained within 14 days of the index ECG (Figure 13). For patients with multiple ECG and TTE data sets meeting these criteria, the earliest pair was used for network creation, validation, or testing, and subsequent TTE data used for analysis of follow up (¶[101]). The study involved developing a convolutional neural network (CNN). CNNs, which have been applied to images (or videos), operate such that the convolutions can be used to extract very subtle patterns in a data set. Each 12-lead ECG was considered a 12 x 5000 (i.e., 12 leads by 10 seconds duration sampled at 500 Hz) "image". The network was composed of N single lead convolutional layers, each of which was followed by a non-linear "Relu" activation function, a batch- normalization layer and a max pooling layer. The features extracted from each raw, digital signal ECG lead were fused in another convolutional layer that had access to all leads simultaneously. Following the last convolutional layer, the data were fed to a fully connected network with two hidden layers with dropout layers to avoid overfitting and an output layer that was activated using the "Softmax" function (¶[104]).
Therefore, it would have been obvious to a person having ordinary skill in the art at the time of filing to train the neural network with a plurality of single-lead electrocardiograms that are extracted from a plurality of 12-lead electrocardiograms obtained from a plurality of other subjects as taught by Attia in the device of Lee, Deepika, Yang, and Witvliet as it is a results-effective method of training a neural network that produces accurate results and can be used to interpret ECG data from any number of leads.
Claim 19 is rejected under 35 U.S.C. 103 as being unpatentable over modified Lee, as evidenced by Silberschatz, as applied to claim 15, and further in view of Wellens.
Regarding claim 19, modified Lee as evidenced by Silberschatz teaches the system of claim 15.
Modified Lee and Wellens teach wherein the plurality of sub-neural networks includes a first sub-neural network that has been trained based on electrocardiograms measured with 12 leads, a second sub-neural network that has been trained based on electrocardiograms measured with at least six limb leads or six precordial leads, and a third sub-neural network model that has been trained based on electrocardiograms measured with single leads (see 35 USC 103 rejection for claims 2-4 above).
Claim 20 is rejected under 35 U.S.C. 103 as being unpatentable over modified Lee, as evidenced by Silberschatz, as applied to claim 15, and further in view of Zhang.
Regarding claim 20, modified Lee as evidenced by Silberschatz teaches the system of claim 15.
Modified Lee and Zhang teach wherein outputs of the plurality of sub-neural networks are concatenated into one to derive the probability of occurrence of thyroid dysfunction (see 35 USC 103 rejection for claim 7 above).
Conclusion
Applicant's amendment necessitated the new ground(s) of rejection presented in this Office action. Accordingly, THIS ACTION IS MADE FINAL. See MPEP § 706.07(a). Applicant is reminded of the extension of time policy as set forth in 37 CFR 1.136(a).
A shortened statutory period for reply to this final action is set to expire THREE MONTHS from the mailing date of this action. In the event a first reply is filed within TWO MONTHS of the mailing date of this final action and the advisory action is not mailed until after the end of the THREE-MONTH shortened statutory period, then the shortened statutory period will expire on the date the advisory action is mailed, and any nonprovisional extension fee (37 CFR 1.17(a)) pursuant to 37 CFR 1.136(a) will be calculated from the mailing date of the advisory action. In no event, however, will the statutory period for reply expire later than SIX MONTHS from the mailing date of this final action.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Emily N Cirulnick whose telephone number is (571)272-9734. The examiner can normally be reached M-Th 8-5:30 and every other F 8-4:30ET.
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, Unsu Jung can be reached at (571) 272-8506. 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.
/E.N.C./ Patent Examiner, Art Unit 3792
/UNSU JUNG/ Supervisory Patent Examiner, Art Unit 3792
Prosecution Insights