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
This application is a continuation of US Application no. 17/054,130, now US Patent no. 12,097,035, filed 9 November 2020, which is the national stage entry of PCT/US2019/031322 filed 8 May 2019, which claims the benefit of domestic priority from US Provisional Application no. 62/668,691 filed 8 May 2018.
Claim Rejections - 35 USC § 102
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 the appropriate paragraphs of 35 U.S.C. 102 that form the basis for the rejections under this section made in this Office action:
A person shall be entitled to a patent unless –
(a)(1) the claimed invention was patented, described in a printed publication, or in public use, on sale, or otherwise available to the public before the effective filing date of the claimed invention.
Claim(s) 1, 5-8, 12-15, 19, and 20 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Rapin et al. (WO 2017/072250).
In regard to claim 1, Rapin et al. disclose a method comprising:
inputting each of a set of training electrocardiogram (EKG) measurements into a machine learning model (pp 16-17, training begins by “taking one cardiac signal from a dataset containing cardiac signals and their known delineations,” followed by “computing the output of the network”. Particularly see page 16 line 20 thru page 17 line 5), to generate for each of the set of training EKG measurements, a corresponding set of outputs (page 17 lines 1-5) comprising an estimated start time of a first feature and an estimated end time of a second feature (page 16 lines 27-30 and page 20 line 22 thru page 21 line 4, the annotations are expressed as wave start and end points (e.g., P, QRS, T waves) and after processing the network provides the onset and offset of each identified wave);
for each of the training EKG measurements:
comparing the corresponding set of outputs to a first feature start time label and a second feature end time label of the training EKG measurement (page 17 lines 6-10, the network parameters are modified to decrease a cost function comparing the known delineation and the output of the network); and
adjusting one or more weight matrices of the machine learning model based on the comparing to train the machine learning model (page 14 lines 1-8, the neural network computes weighted sums using matrix W, and W is expressly identified as the learned parameter; page 17 lines 6-10, during training, the parameters are modified through gradient descent based on the comparison between the known delineation; page 17 lines 10-12, training is repeated over the dataset, after which the trained neural network is recovered; Rapin et al. identifies the learned parameters as the wight matrices W and then teaches modifying those parameters through a gradient step based on comparison of the network output with labeled delineations. It is considered that this comprises and adjustment of one or more weights matrices based on the comparison to train the machine learning model);
receiving EKG data of a subject, wherein the EKG data comprises at least a full beat of the subject (page 20, lines 5-9, the cardiac signal comprising the P-wave, QRS complex, or T-wave is used as input to train the neural network); and
inputting, by a processing device, the EKG data into the machine learning model (page 20 lines 7-14, the trained neural network reads each time point of the cardiac signal) to generate a first output comprising at least a start time of the first feature of the EKG data and an end time of the second feature of the EKG data (page 20 lines 16-23, the post-processing allocates each time point to a wave and provides the onset and offset of each identified wave; Rapin et al. further teaches that the delineation annotations used during training are expressed as wave start and end points, page 16 lines 27-30 and page 17 lines 1-4).
In regard to claim 2, Rapin et al. teaches the step for determining, by the processing device, whether a time difference between the start time of the first feature of the EKG data and the end time of the second feature of the EKG data is greater than a threshold. In regard to claim 5, Rapin et al. teach that the machine learning model comprises a neural network comprising one or more of a convolutional layer or a recurrent layer (page 5 line 24 – page 6 line 21, page 15 line 20 – page 16 line 2).
In regard to claim 6, Rapin et al. teach that the first feature is associated with one of a J- point, a T-wave, a P-wave, an R-wave, or an S-wave (page 19 line 17 thru page 20 line 15, while J-point is not expressly taught, it is considered implicit as this this the end point of the QRS complex).
In regard to claim 7, Rapin et al. teach that the the second feature is associated with one of a J- point, a T-wave, a P-wave, an R-wave, or an S-wave (page 19 line 17 thru page 20 line 15, while J-point is not expressly taught, it is considered implicit as this this the end point of the QRS complex).
In regard to claim 8, Rapin et al. discloses a system comprising:
a memory (page 7 lines 5-9, a memory is not expressly disclosed but is considered a necessary feature for storing the software comprising a trained neural network implemented on a computer device which is considered to comprise a system; page 20 line 29 thru page 21 line 3; the device may comprise a ECG device, mobile phone, tablet etc); and
a processing device operatively coupled to the memory, the processing device to (page 7 lines 5-9 and page 20 line 29 thru page 21 line 3, the invention comprises a computer device (i.e., a processing device) implemented software trained neural network):
input each of a set of training electrocardiogram (EKG) measurements into a machine learning model (pp 16-17, training begins by “taking one cardiac signal from a dataset containing cardiac signals and their known delineations,” followed by “computing the output of the network”. Particularly see page 16 line 20 thru page 17 line 5), to generate for each of the set of training EKG measurements, a corresponding set of outputs (page 17 lines 1-5) comprising an estimated start time of a first feature and an estimated end time of a second feature (page 16 lines 27-30 and page 20 line 22 thru page 21 line 4, the annotations are expressed as wave start and end points (e.g., P, QRS, T waves) and after processing the network provides the onset and offset of each identified wave);
for each of the training EKG measurements:
compare the corresponding set of outputs to a first feature start time label and a second feature end time label of the training EKG measurement (page 17 lines 6-10, the network parameters are modified to decrease a cost function comparing the known delineation and the output of the network); and
adjust one or more weight matrices of the machine learning model based on the comparing to train the machine learning model (page 14 lines 1-8, the neural network computes weighted sums using matrix W, and W is expressly identified as the learned parameter; page 17 lines 6-10, during training, the parameters are modified through gradient descent based on the comparison between the known delineation; page 17 lines 10-12, training is repeated over the dataset, after which the trained neural network is recovered; Rapin et al. identifies the learned parameters as the wight matrices W and then teaches modifying those parameters through a gradient step based on comparison of the network output with labeled delineations. It is considered that this comprises and adjustment of one or more weights matrices based on the comparison to train the machine learning model);
receive EKG data of a subject, wherein the EKG data comprises at least a full beat of the subject (page 20, lines 5-9, the cardiac signal comprising the P-wave, QRS complex, or T-wave is used as input to train the neural network); and
input, by a processing device, the EKG data into the machine learning model (page 20 lines 7-14, the trained neural network reads each time point of the cardiac signal) to generate a first output comprising at least a start time of the first feature of the EKG data and an end time of the second feature of the EKG data (page 20 lines 16-23, the post-processing allocates each time point to a wave and provides the onset and offset of each identified wave; Rapin et al. further teaches that the delineation annotations used during training are expressed as wave start and end points, page 16 lines 27-30 and page 17 lines 1-4).
In regard to claim 12, Rapin et al. teach that the machine learning model comprises a neural network comprising one or more of a convolutional layer or a recurrent layer (page 5 line 24 – page 6 line 21, page 15 line 20 – page 16 line 2).
In regard to claim 13, Rapin et al. teach that the first feature is associated with one of a J- point, a T-wave, a P-wave, an R-wave, or an S-wave (page 19 line 17 thru page 20 line 15, while J-point is not expressly taught, it is considered implicit as this this the end point of the QRS complex).
In regard to claim 14, Rapin et al. teach that the the second feature is associated with one of a J- point, a T-wave, a P-wave, an R-wave, or an S-wave (page 19 line 17 thru page 20 line 15, while J-point is not expressly taught, it is considered implicit as this this the end point of the QRS complex).
In regard to claim 15, Rapin et al. discloses a non-transitory computer-readable medium having instructions stored thereon which, when executed by a processing device, cause the processing device to (page 7 lines 5-9, invention pertains to software comprising a trained neural network implemented on a computer device which is considered to comprise a system; page 20 line 29 thru page 21 line 3; the invention comprises a computer device implemented software comprising the trained neural network):
input each of a set of training electrocardiogram (EKG) measurements into a machine learning model (pp 16-17, training begins by “taking one cardiac signal from a dataset containing cardiac signals and their known delineations,” followed by “computing the output of the network”. Particularly see page 16 line 20 thru page 17 line 5), to generate for each of the set of training EKG measurements, a corresponding set of outputs (page 17 lines 1-5) comprising an estimated start time of a first feature and an estimated end time of a second feature (page 16 lines 27-30 and page 20 line 22 thru page 21 line 4, the annotations are expressed as wave start and end points (e.g., P, QRS, T waves) and after processing the network provides the onset and offset of each identified wave);
for each of the training EKG measurements:
compare the corresponding set of outputs to a first feature start time label and a second feature end time label of the training EKG measurement (page 17 lines 6-10, the network parameters are modified to decrease a cost function comparing the known delineation and the output of the network); and
adjust one or more weight matrices of the machine learning model based on the comparing to train the machine learning model (page 14 lines 1-8, the neural network computes weighted sums using matrix W, and W is expressly identified as the learned parameter; page 17 lines 6-10, during training, the parameters are modified through gradient descent based on the comparison between the known delineation; page 17 lines 10-12, training is repeated over the dataset, after which the trained neural network is recovered; Rapin et al. identifies the learned parameters as the wight matrices W and then teaches modifying those parameters through a gradient step based on comparison of the network output with labeled delineations. It is considered that this comprises and adjustment of one or more weights matrices based on the comparison to train the machine learning model);
receive EKG data of a subject, wherein the EKG data comprises at least a full beat of the subject (page 20, lines 5-9, the cardiac signal comprising the P-wave, QRS complex, or T-wave is used as input to train the neural network); and
input, by a processing device, the EKG data into the machine learning model (page 20 lines 7-14, the trained neural network reads each time point of the cardiac signal) to generate a first output comprising at least a start time of the first feature of the EKG data and an end time of the second feature of the EKG data (page 20 lines 16-23, the post-processing allocates each time point to a wave and provides the onset and offset of each identified wave; Rapin et al. further teaches that the delineation annotations used during training are expressed as wave start and end points, page 16 lines 27-30 and page 17 lines 1-4).
In regard to claim 19, Rapin et al. teach that the machine learning model comprises a neural network comprising one or more of a convolutional layer or a recurrent layer (page 5 line 24 – page 6 line 21, page 15 line 20 – page 16 line 2).
In regard to claim 20, Rapin et al. teach that the first feature is associated with one of a J- point, a T-wave, a P-wave, an R-wave, or an S-wave (page 19 line 17 thru page 20 line 15, while J-point is not expressly taught, it is considered implicit as this this the end point of the QRS complex).
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) 2-4, 9-11, and 16-18 is/are rejected under 35 U.S.C. 103 as being unpatentable over Rapin et al. (WO 2017/072250) in view of Chakravarthy et al. (US Publication no. 2016/0135708).
In regard to claims 2, 9, and 16, Rapin et al. is considered to substantially describe the invention as claimed. Rapin et al. generates the onset (start time) and offset (end time) of ECG waves including P waves, QRS complexes, and T waves, from the output of trained neural networks. Rapin et al. further teaches computing interval measurements derived from those onset and offset locations, such as the PR interval, ST elevation, heart rate, QT interval, QTc, corrected JT interval, and QRS duration (page 6 lines 14-29, page 12 lines 4-10, and page 20 lines 16-25). Using the predetermined thresholds to determine whether a condition is present. Specifically, Rapin et al. teaches that a label us considered present if the score exceeds a predetermined threshold (e.g., 0.5). However this is a classification score and not necessarily a threshold of the cardiac based measurement. Nor does Rapin et al. provide an alert in the event that time difference between features is greater than the threshold. Rapin et al. thus fails to teach determining, by the processing device, whether a time difference between the start time of the first feature of the EKG data and the end time of the second feature of the EKG data is greater than a threshold.
Chakravarthy et al. describes a system for automatically monitoring QT intervals in a patient (para 18). The QT interval is measured similarly to that in Rapin et al. with respect to the onset and offset of the signals. Chakravarthy et al. teaches comparing the detected QT interval to a threshold to detect elongation of the QT interval (para 35; the QT interval being the time interval between the Q and T waves, wherein an elongated QT interval would have a longer than usual time interval between the Q and T wave such that the difference in time interval would be greater than usual). If the QT interval is determined to be longer than the threshold, an alert or notification is sent to the attending medical personnel (para 6 and 35). Modification of Rapin et al. to take an additional step of determining if the difference between the start time to the end time is greater than a threshold is considered to have been obvious to one of ordinary skill in the art at the time of the invention since it would allow Rapin et al. to recognize when abnormal cardiac rhythms are present.
In regard to claims 3, 10, and 17, Rapin et al. is considered to substantially describe the invention as claimed, however does not teach generating an alert indicating that the time difference satisfies the threshold in response to determining that the time difference is greater than the threshold. Rapin et al. further teaches highlighting measurements that fall outside of normal range and displaying delineation labels associated with Long QTc and first degree atrioventricular block on a display of the device (page 21 lines 4-14, page 22 lines 5-29, and page 29 lines 1-15). Chakravarthy et al. teaches that if the QT interval is determined to be longer than the threshold, an alert or notification is sent to the attending medical personnel (para 6 and 35). Modification of Rapin et al. generate an alert when the time difference is greater than the threshold considered obvious to one of ordinary skill in the since the technique is explicitly taught by Chakravarthy et al. for automatically determining if a patient is abnormal cardiac rhythms and alerting medical professionals to such condition.
In regard to claims 4, 11, and 18, Rapin et al. is considered to substantially describe the invention as claimed, however does not teach determining a previous interval between the first feature and the second feature for the subject and generating the threshold based at least in part on the previous interval. Chakravarthy et al. obtains a template ECG signal with QT interval, wherein it is considered that the template signal comprises the basis for the threshold of a normal QT interval. The template signal is also considered a previously obtained measurement from the subject (para 67).
It would have been obvious to one of ordinary skill in the art to modify Rapin et al. to determine the interval threshold based on a previous interval associated with the subject because individualized thresholds account for normal inter-patient variability in ECG morphology and interval duration. Using a historical interval as a baseline would have improved the accuracy of the automated ECG interpretation of Rapin et al. by reducing false positive and improving the ability to identify clinically significant changes in a patient’s cardiac rhythm.
Double Patenting
The nonstatutory double patenting rejection is based on a judicially created doctrine grounded in public policy (a policy reflected in the statute) so as to prevent the unjustified or improper timewise extension of the “right to exclude” granted by a patent and to prevent possible harassment by multiple assignees. A nonstatutory double patenting rejection is appropriate where the conflicting claims are not identical, but at least one examined application claim is not patentably distinct from the reference claim(s) because the examined application claim is either anticipated by, or would have been obvious over, the reference claim(s). See, e.g., In re Berg, 140 F.3d 1428, 46 USPQ2d 1226 (Fed. Cir. 1998); In re Goodman, 11 F.3d 1046, 29 USPQ2d 2010 (Fed. Cir. 1993); In re Longi, 759 F.2d 887, 225 USPQ 645 (Fed. Cir. 1985); In re Van Ornum, 686 F.2d 937, 214 USPQ 761 (CCPA 1982); In re Vogel, 422 F.2d 438, 164 USPQ 619 (CCPA 1970); In re Thorington, 418 F.2d 528, 163 USPQ 644 (CCPA 1969).
A timely filed terminal disclaimer in compliance with 37 CFR 1.321(c) or 1.321(d) may be used to overcome an actual or provisional rejection based on nonstatutory double patenting provided the reference application or patent either is shown to be commonly owned with the examined application, or claims an invention made as a result of activities undertaken within the scope of a joint research agreement. See MPEP § 717.02 for applications subject to examination under the first inventor to file provisions of the AIA as explained in MPEP § 2159. See MPEP § 2146 et seq. for applications not subject to examination under the first inventor to file provisions of the AIA . A terminal disclaimer must be signed in compliance with 37 CFR 1.321(b).
The filing of a terminal disclaimer by itself is not a complete reply to a nonstatutory double patenting (NSDP) rejection. A complete reply requires that the terminal disclaimer be accompanied by a reply requesting reconsideration of the prior Office action. Even where the NSDP rejection is provisional the reply must be complete. See MPEP § 804, subsection I.B.1. For a reply to a non-final Office action, see 37 CFR 1.111(a). For a reply to final Office action, see 37 CFR 1.113(c). A request for reconsideration while not provided for in 37 CFR 1.113(c) may be filed after final for consideration. See MPEP §§ 706.07(e) and 714.13.
The USPTO Internet website contains terminal disclaimer forms which may be used. Please visit www.uspto.gov/patent/patents-forms. The actual filing date of the application in which the form is filed determines what form (e.g., PTO/SB/25, PTO/SB/26, PTO/AIA /25, or PTO/AIA /26) should be used. A web-based eTerminal Disclaimer may be filled out completely online using web-screens. An eTerminal Disclaimer that meets all requirements is auto-processed and approved immediately upon submission. For more information about eTerminal Disclaimers, refer to www.uspto.gov/patents/apply/applying-online/eterminal-disclaimer.
Claims 1-20 are rejected on the ground of nonstatutory double patenting as being unpatentable over claims 1-17 of U.S. Patent No. 12,097,035. Although the claims at issue are not identical, they are not patentably distinct from each other because the ‘035 patent anticipates each and every element of the present invention. For instance, such anticipation is exhibited by the conflicting limitations between claims 1 of the ‘035 patent over claim 1 of the present invention.
Claim 1 of the ‘035 patent:
A method comprising:
inputting each of a set of training electrocardiogram (EKG) measurements into a machine learning model to generate an output comprising an estimated start time of a Q-wave and an estimated end time of a T-wave
for each of the set of training EKG measurements;
comparing each of the set of outputs to a corresponding labeled start time of a Q-wave and a labeled end time of a T-wave and
adjusting one or more weight matrices of the machine learning model based on the comparing to train the machine learning model; receiving EKG data of a subject, wherein the EKG data comprises at least a full beat of the subject; and
inputting, by a processing device, the EKG data into the machine learning model to generate a first output comprising a segmentation of the full beat, wherein the segmentation comprises at least a start time of a Q-wave and an end time of a T-wave of the EKG data.
Claim 1 of the present invention:
A method comprising:
inputting each of a set of training electrocardiogram (EKG) measurements into a machine learning model to generate,
for each of the set of training EKG measurements, a corresponding set of outputs comprising an estimated start time of a first feature and an estimated end time of a second feature;
for each of the training EKG measurements:
comparing the corresponding set of outputs to a first feature start time label and a second feature end time label of the training EKG measurement; and
adjusting one or more weight matrices of the machine learning model based on the comparing to train the machine learning model; receiving EKG data of a subject, wherein the EKG data comprises at least a full beat of the subject; and
inputting, by a processing device, the EKG data into the machine learning model to generate a first output comprising at least a start time of the first feature of the EKG data and an end time of the second feature of the EKG data.
The difference between the claim sets is that the ‘035 patent more narrowly recites the start time of the Q Wave and end time of the T-wave vs the broader recitation of “first feature start time label” and “second feature end time label” in the present invention. However, the narrower recitation reads on and anticipates the broader recitation of the present invention. Therefore, the inventions are not deemed patentably distinct.
Similar overlap is also exhibited between claims 2-17 of the ‘035 patent over claims 2-20 of the present invention.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Tarassenko et al. (US Publication no. 2010/0056939) relates to detecting cardiac signal onset and offset points so that the QT interval can be obtained. This is achieved by using two separate Hidden Markov Models, one of which detects the Qonset point and one the Toffset, the two models being independent (para 44). By using two separate models, Tarassenko et al. fails to teach or suggest training a single machine learning model by inputting each of a set of training electrocardiogram (EKG) measurements into a machine learning model to generate an output comprising an first feature start time label and second feature end time label training EKG measurements and comparing each of the set of outputs to a corresponding first feature start time label and second feature end time label for adjusting one or more weight matrices of the machine learning model based on the comparing to train the machine learning model as recited in the present invention.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to BRIAN T GEDEON whose telephone number is (571)272-3447. The examiner can normally be reached M-F 8:00 am to 5:30 PM ET.
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, David E. Hamaoui can be reached at 571-270-5625. 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.
/BRIAN T GEDEON/Primary Examiner, Art Unit 3796 14 July 2026
Prosecution Insights