Deep Learning Based Approach For OCT Image Quality Assurance

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 . Continued Examination Under 37 CFR 1.114 A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on 24 June 2026 has been entered. The Applicant’s Remarks filed 26 May 2026 have been received and considered. The claim 7 objection in the Final Office Action mailed 26 February 2026 has been withdrawn. The IDS’s filed 23 July 2026 and 04 September 2026 have been received and considered. Claims 1, 7, 15, and 19 have been amended. Claim 2 has been canceled. Claims 1 and 3 – 20, all of the remaining claims still pending in this application, have been rejected. Response to Applicant’s Remarks In view of the Applicant’s arguments filed 26 May 2026, regarding amendments to independent claims 1, 15, and 19, the previously applied prior art rejections are withdrawn. Applicant's arguments are moot in view of the new grounds of rejection set forth below in light of the amendment. 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. Claims 1, 3 – 7, 9 – 10, 12 – 13, and 15 – 20 are rejected under 35 U.S.C. 103 as being unpatentable over Non-Patent Literature “ARC –OCT: Automatic detection of lumen border in intravascular OCT images” to Cheimariotis et al. (hereinafter Cheimariotis) in view of Non-Patent Literature “Fully automated plaque characterization in intravascular OCT images using hybrid convolutional and lumen morphology features” to Lee et al. (hereinafter Lee) in further view of Rothberg et al. (US Publication No. 20190130554 A1) (hereinafter Rothberg). Claim 1 Cheimariotis teaches a method of identifying quality diagnostic medical images, the method comprising: receiving, by one more processors, diagnostic medical images of a lumen captured as a series of images during a patient imaging procedure ("Our study was based on 1812 images, obtained from 20 coronary arteries [LAD (left anterior descending artery), n = 14; LCX (left circumflex artery), n = 1; RCA (right coronary artery), n = 5; mean length = 39.7 ± 10.0 mm], of patients who underwent a clinically indicated cardiac catheterization and OCT. From these images, 308 corresponded to stented arterial segments. A representative OCT image of the elaborated set is presented in Fig. 1. Images were acquired with a Frequency Domain OCT imaging system (FD-OCT C7-XRT OCT Intravascular Imaging System, Westford, MA, USA). The pullback speed was 20 mm/s, the axial resolution was 15 μm, and the frame rate was 100 frames/s.", Section 3. Method); analyzing by one or more processors, in real time or near real time, with a trained machine learning model, the diagnostic medical images, wherein the trained machine learning model is trained on a set of annotated diagnostic medical images with respect to lumen view obstruction within the medical images ("In this study we presented ARC –OCT, a method for fully automatic detection of lumen borders in OCT images. Our major goal was to develop a method capable of segmenting real-world OCT images containing stented segments, branches, and artifacts [2,5,6].", Section 6. Discussion and conclusion; "In this paper, we present ARC –OCT, a method for fully automatic detection of the lumen border in OCT images. ARC –OCT relies on a multi-step segmentation procedure that applies on real-world OCT images…The luminal contour in a radial scan (the so-called A-line) is located at the point where the scan (starting from the catheter) meets a bright shape (i.e., a transition from black to white). Once the initial assessment of the internal contour is performed, artifacts are corrected by applying a smoothing function on the curve, which represents the internal contour. Next, side branches of the vessel are identified with statistical control of the curve slope…we compared ARC –OCT with manual segmentation (conducted by expert cardiologists) corresponding to the ground-truth.", Section 1. Introduction; "Concerning time performance, the algorithm can segment a single slice in less than a second, without applying 3D smoothing, which is suitable for real time inspection.", Section 5. Results), wherein ground truth images are labeled; Cheimariotis teaches the system being able to function on both good and poor quality images, but does not explicitly teach Identifying by one or more processors, based on the analyzing, image quality for the diagnostic medical images, wherein the image quality is based, at least in part, on one or more confidence threshold values for lumen detection due to an extent of artifacts within the lumen causing lumen view obstruction. However, Lee teaches Identifying, based on the analyzing, an extent of artifacts within the lumen causing lumen view obstruction (Figure 2; "IVOCT images were acquired with a frequency-domain OCT system (ILUMIEN OPTIS; St. Jude Medical Inc.). During image acquisition, the optical probe was automatically pulled back from distal to proximal at a speed of 36 mm/s, frame rate of 180 frames/sec, and axial resolution of 20 μm. All IVOCT image frames were initially reviewed…frames having poor quality due to luminal blood…unclear lumen…artifact…reverberation…”, Experimental Methods - Images and their annotation). This clearly indicates that one of ordinary skill in the art would want to review all images (“All IVOCT image frames”) to identify any artifacts or objects that a causing lumen view obstruction. (“The proposed method gave better classification results for the fibrolipidic class (F1 score ≈ 0.887) as compared to those in fibrocalcific class (F1 score ≈ 0.677).”, Results), wherein the the fibrolipidic class represents lipid-rich, potentially vulnerable arterial blockages and the fibrocalcific class represents hardened, calcified plaques. PNG media_image1.png 534 530 media_image1.png Greyscale It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to modify the teachings of Cheimariotis, to incorporate identifying the extent of lumen artifacts, as disclosed by Lee. The suggestion/motivation for doing so would have been to determine images having the artifacts, allowing the images to be analyzed much quicker and categorizing the said artifacts into two distinct types, if the current view allows it. Neither Cheimariotis, or Lee, or the combination teaches Identifying by one or more processors, an image quality for the diagnostic medical images, wherein the image quality is based, at least in part, on one or more confidence threshold values for lumen detection. outputting for display on a user interface, in real time or near real time, an indication of the identified image quality. However, Rothberg teaches Identifying by one or more processors, an image quality for the diagnostic medical images, wherein the image quality is based, at least in part, on one or more confidence threshold values for lumen detection and outputting for display on a user interface, in real time or near real time, an indication of the identified image quality ("FIG. 6 illustrates an example of a graphical user interface (GUI) 600 that may be displayed by a display screen of a computing device during imaging with an ultrasound device in communication with the computing device, in accordance with certain embodiments described herein. The GUI 600 displays an ultrasound image 602, a live quality indicator 604, a record option 614, and an instruction 616. The live quality indicator 604 includes a bar 606 having a first end 608 and a second end 610, a slider 618, an acceptability indicator 620, and text 612.", Paragraph [0085]; “The live quality may be calculated in real-time during imaging, as the images are collected. The process 200 begins with act 202”, Paragraph [0054]; “Additionally, while the above description has described calculating qualities for ultrasound images, the same processes may also be used to calculate the qualities of other types of medical images, such as X-ray, computerized tomography (CT), and magnetic resonance imaging (MRI) images.”, Paragraph [0080]), wherein it is known to those skilled in the art that ultrasound is also widely and effectively used in arterial imaging to evaluate blood flow, structure, and blockages in the arteries. It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee, to incorporate identifying and outputting, in real time or near real time, an indicator of image quality, which may include arterial images amongst other image types, as disclosed by Rothberg. The suggestion/motivation for doing so would have been so the practitioner can adjust to get a better view of the lumen during the procedure, whether that be for categorizing the artifacts/objects in the lumen or for a better overall view of the lumen. Claim 3 Regarding Claim 3, dependent on claim 1, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 1. Cheimariotis further teaches wherein the series of diagnostic medical images is obtained through an optical coherence tomography pullback (Rejected as applied to claim 1). Claim 4 Regarding Claim 4, dependent on claim 1, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 1. While Cheimariotis teaches utilizing a dynamic threshold selection method on the respective regions of interest of the original image, in which a histogram is divided into two classes by seeking for minimal intraclass and maximal interclass variance, the classifying the diagnostic medical images as a first classification or a second classification is not explicitly taught. However, Lee further teaches classifying the diagnostic medical images as a first classification or a second classification ("If an A-line included ≥3 pixels of either lipidous or calcified plaques, we determined which of these two classes was in the majority, and then labeled the A-line accordingly as fibrolipidic or fibrocalcific.", Experimental Methods - Images and their annotations). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate classifying the images into two classes, as disclosed by Lee. The suggestion/motivation for doing so would have been to have been using the classified types to determine a treatment plan to rid the artifacts lining the wall of the lumen. Claim 5 Regarding Claim 5, dependent on claim 4, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 4. Cheimariotis does not teach providing an alert or notification when the diagnostic medical images areclassified in the second classification. However, Lee further teaches providing an alert or notification when the diagnostic medical images areclassified in the second classification (Figure 6; "Hybrid classification followed by CRF gave good A-line classification results. The proposed method gave better classification results for the fibrolipidic class (F1 score ≈ 0.887) as compared to those in fibrocalcific class (F1 score ≈ 0.677).", Results), wherein the different colors in Figure 6 directly below notify the user of the classification since the classification is color dependent. PNG media_image2.png 434 551 media_image2.png Greyscale Claim 6 Regarding Claim 6, dependent on claim 1, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 1. Cheimariotis further teaches wherein the set of annotated diagnostic medical images comprises annotations including clear, blood, or guide catheter (Figure 1). PNG media_image3.png 409 392 media_image3.png Greyscale Claim 7 Regarding Claim 7, dependent on claim 1, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 1. Cheimariotis further teaches wherein the diagnostic medical images are (Rejected as applied to claim 7). Claim 9 Regarding Claim 9, dependent on claim 1, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 1. Neither Cheimariotis, or Lee, or the combination teach computing a probability indicative of whether the diagnostic medical images are acceptable or not acceptable. However, Rothberg teaches computing a probability indicative of whether the diagnostic medical images are acceptable or not acceptable (“In some embodiments, the method further includes generating for display, during the imaging, an indicator of an acceptability of the sequence of images, where the indicator of the acceptability of the sequence of images includes an indicator of a threshold quality such that a distance from the first end to the indicator of the threshold quality relative to the length of the bar is proportional to the threshold quality.”, Paragraph [0009]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate computing a probability indicative of the diagnostic images being acceptable or not, as disclosed by Rothberg. The suggestion/motivation for doing so would have been to determine, from the images taken, which images would be most viable for further analysis and removing those images that are prone to error or could potentially lead to misdiagnosis. Claim 10 Regarding Claim 10, dependent on claim 9, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 9. Neither Cheimariotis, or Lee, or the combination teach using a threshold method to convert the computed probability to a classification of the diagnostic medical images. However, Rothberg teaches using a threshold method to convert the computed probability to a classification of the diagnostic medical images (“If the quality is above a threshold quality (e.g., 50% on a scale of 0% to 100%), automatic measurement based on the sequence of images may proceed. If the quality is below the threshold quality, an error may be displayed indicating that automatic measurement will not proceed.”, Paragraph [0044]). Claim 12 Regarding Claim 12, dependent on claim 9, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 9. Cheimariotis does not teach using morphological classification to convert the computed probability to a classification of the diagnostic medical images. However, Lee further teaches using morphological classification to convert the computed probability to a classification of the diagnostic medical images (Figure 8; “Since A-line classification does not fully consider spatial similarities between nearby A-lines in θ and z, we employed classification noise cleaning. In particular, we implemented a fully connected CRF that defines the pairwise edge potentials of each probabilistic classification (θ, z) by a linear combination of Gaussian kernels39.”, Image Analysis Methods – Classification noise reduction using fully connected conditional random field (CRF); “Adding lumen morphological features gave statistically significant improvement (p < 0.05), as compared to classification with convolutional features alone.”, Abstract). PNG media_image4.png 245 546 media_image4.png Greyscale It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate using morphological classification to convert the computed probability to a classification of the diagnostic medical images, as disclosed by Lee. The suggestion/motivation for doing so would have been to better distinguish the artifacts in the image for a more accurate image classification. Claim 13 Regarding Claim 13, dependent on claim 9, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 9. Neither Cheimariotis, or Lee, or the combination teach wherein acceptable means that the diagnostic medical images are above a predefined threshold quality which allows for evaluation of characteristics of human tissue above a threshold level of accuracy or confidence. However, Rothberg teaches wherein acceptable means that the diagnostic medical images are above a predefined threshold quality which allows for evaluation of characteristics of human tissue above a threshold level of accuracy or confidence ('If the quality is above a threshold quality (e.g., 50% on a scale of 0% to 100%), automatic measurement based on the sequence of images may proceed. If the quality is below the threshold quality, an error may be displayed indicating that automatic measurement will not proceed.", Paragraph [0044]). Claim 17 Regarding Claim 17, dependent on claim 16, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 16. Cheimariotis does not teach wherein the instructions are configured to display a plurality of OCT images along with an indicator associated with a classification of each image of the plurality of OCT images. However, Lee further teaches wherein the instructions are configured to display a plurality of OCT images along with an indicator associated with a classification of each image of the plurality of OCT images (Figure 6), where specific colors in the image "indicate" the classification type. PNG media_image2.png 434 551 media_image2.png Greyscale It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate displaying the images with an indicator associated with the classification, as disclosed by Lee. The suggestion/motivation for doing so would have been to better distinguish the artifacts in the image. Claim 15, an independent system claim, is rejected for the same reasons as applied to claim 1. Claim 19, an independent non-transitory computer readable medium claim, is rejected for the same reason as applied to claim 1. Claims 16, 18, and 20 are rejected for the same reasons as applied to the above claims. Claim 8 is rejected under 35 U.S.C. 103 as being unpatentable over Non-Patent Literature “ARC –OCT: Automatic detection of lumen border in intravascular OCT images” to Cheimariotis et al. (hereinafter Cheimariotis) in view of Non-Patent Literature “Fully automated plaque characterization in intravascular OCT images using hybrid convolutional and lumen morphology features” to Lee et al. (hereinafter Lee) in further view of Rothberg et al. (US Publication No. 20190130554 A1) (hereinafter Rothberg) in further view of WANG et al. (US Publication No. 20240005501 A1) (hereinafter WANG). Claim 8 Regarding Claim 8, dependent on claim 1, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 1. Neither Cheimariotis, or Lee, or Rothberg, or the combination teaches classifying the diagnostic medical images as a clear medical image or a blood medical image. However, WANG teaches classifying the diagnostic medical images as a clear medical image or a blood medical image ("The inventors therefore considered the significance of tissue pathology of wound as well as the frequency of morphological appearances in typical skin OCT images, to identify a plurality of classes that could be associated with improved accuracy. They decided to define seven distinctive image sub-types within the OCT image of a skin wound, namely neoepidermis, clot, granular tissue, collagen, intact tissue, blood (liquid) and outside (also referred to as “void” or “background”).", Paragraph [0089]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate classifying the diagnostic medical images as a clear medical image or a blood medical image, as disclosed by WANG. The suggestion/motivation for doing so would have been to determine if an image contained residual blood so as to discard the image to reduce error during patient diagnosis and medical image analysis. Claim 11 is rejected under 35 U.S.C. 103 as being unpatentable over Non-Patent Literature “ARC –OCT: Automatic detection of lumen border in intravascular OCT images” to Cheimariotis et al. (hereinafter Cheimariotis) in view of Non-Patent Literature “Fully automated plaque characterization in intravascular OCT images using hybrid convolutional and lumen morphology features” to Lee et al. (hereinafter Lee) in further view of Rothberg et al. (US Publication No. 20190130554 A1) (hereinafter Rothberg) in further view of Wang et al. (US Publication No. 20160213253 A1) (hereinafter Wang). Claim 11 Regarding Claim 11, dependent on claim 9, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 9. Neither Cheimariotis, or Lee, or Rothberg, or the combination teaches using graph cuts to convert the computed probability to a classification of the diagnostic medical images. However, Wang teaches using graph cuts to convert the computed probability to a classification of the diagnostic medical images ("A final probability estimate of strut presence can then be determined based on the interpolated strut depths and the high confidence struts that have been identified. In some examples, the depth calculator 52 can use a graph cut method to localize depths of stent struts based on the OCT image data 38.", Paragraph [0031]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate using graph cuts to covert the computed probability to a classification of the diagnostic medical images, as disclosed by Wang. The suggestion/motivation for doing so would have been to allow for a more precise classification and reduce error and misdiagnosis of the captured medical images. Claim 14 is rejected under 35 U.S.C. 103 as being unpatentable over Non-Patent Literature “ARC –OCT: Automatic detection of lumen border in intravascular OCT images” to Cheimariotis et al. (hereinafter Cheimariotis) in view of Non-Patent Literature “Fully automated plaque characterization in intravascular OCT images using hybrid convolutional and lumen morphology features” to Lee et al. (hereinafter Lee) in further view of Rothberg et al. (US Publication No. 20190130554 A1) (hereinafter Rothberg) in further view of Fridley et al. (US Publication No. 20220260988 A1) (hereinafter Fridley). Claim 14 Regarding Claim 14, dependent on claim 13, Cheimariotis, in view of Lee and Rothberg, teaches the invention as claimed in claim 13. Neither Cheimariotis, or Lee, or Rothberg, or the combination teaches wherein a value for the predefined threshold quality is determined by optimizing a machine learning model. However, Fridley teaches wherein a value for the predefined threshold quality is determined by optimizing a machine learning model ("For example, the inference subsystem 130 can implement machine learning models that are configured to (i) optimize the threshold at which anomalies are predicted or detected.", Paragraph [0062]). It would have been obvious to one of ordinary skill in the art before the effective filing date of the claimed invention to further modify the teachings of Cheimariotis, in view of Lee and Rothberg, to incorporate determining a value for a predefined threshold quality by optimizing a machine learning model, as disclosed by Fridley. The suggestion/motivation for doing so would have been to train the machine learning model with various batches of training data, optimizing the threshold until a desired success rate was achieved so as to reduce error in medical image classifications. Conclusion The prior art made of record and not relied upon is considered pertinent to applicant's disclosure. Shiju Joseph, Asif Adnan, David Adlam, "Automatic segmentation of coronary morphology using transmittance-based lumen intensity-enhanced intravascular optical coherence tomography images and applying a localized level-set-based active contour method," J. Med. Imag. 3(4) 044001 (29 November 2016) https://doi.org/10.1117/1.JMI.3.4.044001, where Joseph also discloses annotated OCT images of a lumen. PNG media_image5.png 444 568 media_image5.png Greyscale PNG media_image6.png 307 553 media_image6.png Greyscale Huaizhong Zhang et al., “Automatic vessel lumen segmentation in optical coherence tomography (OCT) images”. Applied Soft Computing Journal (2020). https://doi.org/10.1016/j.asoc.2019.106042, where Zhang also discloses annotated OCT images of a lumen. PNG media_image7.png 313 656 media_image7.png Greyscale Any inquiry concerning this communication or earlier communications from the examiner should be directed to Ronde Miller whose telephone number is (703) 756-5686 The examiner can normally be reached Monday-Friday 8:00-4:00. 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 Gregory Morse can be reached on (571) 272-3838. 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. /RONDE LEE MILLER/Examiner, Art Unit 2663 /GREGORY A MORSE/Supervisory Patent Examiner, Art Unit 2698