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and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114.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 05/22/2026 has been entered.
Response to Arguments
Applicant Arguments:
Claim Rejections - 35 USC § 102:
In regards to Argument 1, With respect to independent claims 1, 19 and 20: Applicant's arguments filed 05/22/2026 have been fully considered but they are not persuasive. In the present application, applicant argues: “Applicant respectfully submits that Sorenson fails to disclose at least determining a number of clusters by grouping descriptors extracted from first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof, and automatically training a first number of data analysis tools based on the descriptors extracted from the first medical image data and the first analysis data related to the first medical image data, wherein at least one of the first number of data analysis tools is trained specifically for each determined cluster using the descriptors within the determined cluster. The present framework is directed to a framework that (1) determines clusters of medical image data by grouping images based on extracted image descriptors, and then (2) automatically trains a dedicated data analysis tool specifically for each determined cluster using the descriptors within the determined cluster. This architecture enables the system to develop specialized tools tuned to the specific characteristics of each image cluster. …Sorenson fails to disclose determining a number of clusters by grouping descriptors extracted from first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof. The Examiner maps the claimed "determining a number of clusters" to Sorenson's auto- categorization module (paragraphs 214-215) and workflow correlation (paragraph 183). This mapping is fundamentally flawed. Sorenson's auto-categorization module categorizes images based on rules, training based on user, machine learning, DICOM Headers, in-image analysis, analysis of pixel attributes, landmarks within the images, characterization methods, statistical methods, or any combination thereof, not descriptors extracted from the image content itself. See Sorenson at, e.g., paragraph [0215]. Similarly, Sorenson's machine learning module in paragraph 183 correlates image data to workflows based on in-image analysis and metadata. This is not "determining a number of clusters by grouping image descriptors." On page 6 of the Office Action, the Examiner also points to paragraph 76 of Sorenson, where "a COPD engine can machine learn based on the same COPD engine data," and interprets "the same COPD engine data" as a "cluster." This interpretation is not reasonable. In Sorenson, the COPD engine data simply refers to the set of data that was used to train or update a COPD engine - it is not the result of a clustering process applied to image descriptors. There is no disclosure in Sorenson of a step that (a) extracts image descriptors from a set of medical images and (b) groups those descriptors into clusters to thereby determine a number of clusters. Additionally, Sorenson also fails to disclose automatically training a first number of data analysis tools based on the descriptors extracted from the first medical image data and the first analysis data related to the first medical image data, wherein at least one of the first number of data analysis tools is trained specifically for each determined cluster using the descriptors within the determined cluster. The claims require that for each cluster discovered through image descriptor grouping, a data analysis tool is specifically trained using the descriptors within that specific cluster … The Examiner's reading requires treating Sorenson's entire multi-engine ecosystem as equivalent to "automatically training a first number of data analysis tools...wherein at least one is trained for each determined cluster." But this reads away the requirement that the specialization of the tools be determined by and tied to the clusters and descriptors. The claimed process is: cluster the data using descriptors -- determine from the clustering what data trains each - automatically train those tools using descriptors from within the respective cluster. Sorenson's process is: design engines clinically -- train them independently -- offer them in an app store. These are fundamentally different architectures. Consequently, Sorenson fails to disclose "each and every element" as set forth in these claims. Accordingly, Applicant respectfully asks the Examiner to withdraw the rejection of these claims. (Remark Pages 7-10)
In regards to Argument 2, With respect to claim 10, Applicant/s state/s Sorenson fails disclose Sorenson fails disclose at least wherein the clusters correspond to different pathologies. As discussed previously, Sorenson fails to disclose a prior clustering step, much less clusters corresponding to different pathologies.. (Remark Page 11)
In regards to Argument 3, With respect to claim 12, Applicant/s state/s , Sorenson fails disclose Sorenson fails disclose at least extract the descriptors by sampling the first medical image data, the third medical image data, the first analysis data related to the first medical image data, the third analysis data, or a combination thereof using a sampling model. As discussed previously, Sorenson fails to disclose determining descriptors for images in the first or third medical image data, much less extracting such descriptors by sampling image data using a sampling model. (Remark Page 11)
Claim Rejections - 35 USC § 103:
In regards to Argument 4, With respect to claim 11, For claim 11, the Examiner acknowledges that Sorenson does not disclose determine the number of clusters by performing a K-means algorithm, and relies instead on Yerebakan for this feature. However, as argued above, Sorenson fails to disclose the foundational claim elements - specifically, determining clusters by grouping image descriptors, and training tools specific to the cluster. Yerebakan fails to cure these fundamental deficiencies. Moreover, the proposed combination is improper. However, Sorenson already achieves finding recognition through its multi-engine platform - there is no motivation within the context of Sorenson to add descriptor-based K-means clustering as a prerequisite to engine training, because Sorenson's training is performed by developers on specific pre-defined disease categories, not triggered by data-driven clustering. (Remark Page 13)
In regards to Argument 5, With respect to claim 13, the asserted combination of Sorenson and Yerebakan fails disclose at least automatically train the first or second number of data analysis tools in response to the number of descriptors in any one cluster exceeds exceeding a threshold value. The most fundamental flaw in the Examiner's § 103 rejection is a misreading of what Yerebakan's threshold actually does. Yerebakan at paragraph [0052] describes a distance threshold that governs how similar two descriptors must be in order to be assigned to the same cluster during the clustering algorithm. This is a parameter that controls the granularity of clustering - it determines whether two descriptors are "similar enough" to belong to the same group. Claim 13, by contrast, recites a threshold on the number of descriptors within an already-formed cluster - and this threshold serves as a trigger for training. The two thresholds are conceptually and functionally distinct. These are not the same threshold, and Yerebakan nowhere discloses or suggests using a descriptor count within a cluster as a trigger for training a data analysis tool. (Remark Page 13)
Examiner’s Response:
In response to Argument 1, With respect to independent claims 1, 19 and 20: the Examiner respectfully disagree. Sorenson discloses A machine learning module can receive image data from a medical image data source. The machine learning module can correlate image data (read as clustering) from the medical image data source to a workflow based on in-image analysis (read as “analysis data”) and metadata ( read as “medical image data”) … the correlation of the medical data to a machine learning module or collection of machine learning can be done based on pattern extraction, feature extraction or image processing which result of a medical image classification (clusterization). … The machine learning module can associate the series with the workflow based on in-image analysis, for example, by reconstructing the image or analyzing the pixel characteristics (e.g., intensity, shading, color, etc.) to determine the anatomy or modality. This correlation can be done based on organ (liver, kidney, heart, brain . . . ), body part (head, head and neck, chest, abdomen . . . ) or general feature extraction is interpreted as “ determining a number of clusters by grouping descriptors extracted”. Therefore, Sorenson teaches “determining a number of clusters by grouping descriptors extracted from based on first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof,” The Examiner is interpreted under Broadest Reasonable Interpretation of the claim limitation “determining a number of clusters” base on only “a first image data” or “a third image data”. For the purpose of examination , “descriptors extracted” is considered as “organ (liver, kidney, heart, brain . . . ), body part (head, head and neck, chest, abdomen . . . ) or general feature extraction”, in fact the applicant does not specifically define what “descriptors extracted” is. Thus, it is suggested the applicant at least to amend to define detail about “descriptors” is for compact prosecution purpose.
Also, Sorenson discloses the medical image data can be sent via a network to the Processing Server where the auto-categorization module can categorize each of the images sent to the processing server …the auto-categorization module can categorize the images (read as “determine the number of clusters”) based on rules, training based on user, machine learning, DICOM Headers, in-image analysis, analysis of pixel attributes, landmarks within the images, characterization methods, statistical methods, or any combination thereof. The tracking module can track the images based on categories, for example, modality, orientation (e.g., axial, coronal, sagittal, off axis, short axis, 3 chamber view, or any combination thereof), anatomies (organs, vessels, bones, or any combination thereof), body section (e.g., head, next, chest, abdomen, pelvis, extremities, or any combination thereof), sorting information (e.g., 2D, 2.5D, 3D, 4D), study/series description, scanning protocol, sequences, options, flow data, or any combination thereof is interpreted as “determining a number of clusters by grouping descriptors extracted”. Therefore, Sorenson teaches “determining a number of clusters by grouping descriptors extracted from based on first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof,” (Paragraph 214-215)
Second, Sorenson teaches the machine learning module can read image data from the series list and determine that based on the metadata and in-image analysis of the image data, a specific image data should be recommended to the user for the ejection fraction workflow. This correlation will be reinforce based on end-user feedback up to the point where the machine learning module will be able to automatically select the relevant data (consider as “specifically for each determined cluster) to be processed by other machine learning module or read using a particular clinical protocol. … through numerous iterations, the machine learning module can propose specific image data by machine learning based on, for example, the user's interactions and the information from the in-image analysis, metadata and patient clinical context or exam order is consider as “at least one of the first number of data analysis tools is trained specifically for each determined cluster using the descriptors within the determined cluster.” (Paragraph 183) For the purpose of examination , The Examiner is interpreted under Broadest Reasonable Interpretation of the claim limitation “determined cluster using the descriptors within the determined cluster” is considered as “correlation will be reinforce based on end-user feedback … select the relevant data to be processed”.
Furthermore, Sorenson discloses machine learning module that is part of an artificial intelligence findings system includes an image identification engine that can extract features from a new medical image being analyzed to match this data to data present in the archive with the same characteristic (disease). Additionally, for example, the machine learning module can analyze data to extract a feature and prebuild a hanging protocol for a better viewing experience. Additionally, for example, the image identification engine within the machine learning module can analyze data to extract a feature and find similar data for the same patient to automatically present the data with its relevant prior information. For example, the similar data (read as “ cluster”) can pertain to anatomical structures (read as “descriptors”) such as body parts, anatomic anomalies and anatomical features is interpreted as “at least one of the first number of data analysis tools is trained specifically for each determined cluster using the descriptors within the determined cluster.” For the purpose of examination , “the image identification engine within the machine learning module can analyze data to extract a feature and find similar data for the same patient to automatically present the data with its relevant prior information. For example, the similar data can pertain to anatomical structures such as body parts, anatomic anomalies and anatomical features” is considered as “trained specifically for each determined cluster using the descriptors within the determined cluster”, in fact the applicant does not specifically define what “determined cluster using the descriptors” is. Thus, it is suggested the applicant at least to amend to define detail about “determined cluster using the descriptors” is for compact prosecution purpose.
The Examiner states that in light of MPEP 2111, the Examiner has interpreted the claims properly. Specifically, during patent prosecution, the pending claims must be “given their broadest reasonable interpretation assistant with the specification.” The Examiner has interpreted the claim language in reference to the specification. Because applicant has the opportunity to amend the claims during prosecution, given a claim in its broadest reasonable interpretation will reduce the possibility that the claim, once issued will be interpreted more or broadly than is justified. Although the cited reference is different from the invention disclosed, the language of Applicant's claims is sufficiently broad to reasonably read on the cited reference. A broad reading does not constitute “teaching away.”
Further, it has been held that nonpreferred embodiments failing to assert discovery beyond that known in the art does not constitute a “teaching away” unless such disclosure criticizes, discredits, or otherwise discourages the solution claimed. In re Susi, 440 F.2d 442, 169 USPQ 423 (CCPA 1971), In re Gurley, 27 F.3d 551, 554, 31 USPQ2d 1130, 1132 (Fed. Cir. 1994), In re Fulton, 391 F.3d 1195, 1201, 73 USPQ2d 1141, 1146 (Fed. Cir. 2004), (see MPEP §2124).
Disclosed examples and preferred embodiments do not constitute a teaching away from a broader disclosure or nonpreferred embodiments. In re Susi, 440 F.2d 442, 169 USPQ 423 (CCPA 1971). “A known or obvious composition does not become patentable simply because it has been described as somewhat inferior to some other product for the same use.” In re Gurley, 27 F.3d 551, 554, 31 USPQ2d 1130, 1132 (Fed. Cir. 1994) (The invention was directed to an epoxy impregnated fiber-reinforced printed circuit material. The applied prior art reference taught a printed circuit material similar to that of the claims but impregnated with polyester-imide resin instead of epoxy. The reference, however, disclosed that epoxy was known for this use, but that epoxy impregnated circuit boards have “relatively acceptable dimensional stability” and “some degree of flexibility,” but are inferior to circuit boards impregnated with polyester-imide resins. The court upheld the rejection concluding that applicant’s argument that the reference teaches away from using epoxy was insufficient to overcome the rejection since “Gurley asserted no discovery beyond what was known in the art.” 27 F.3d at 554, 31 USPQ2d at 1132.). Furthermore, “[t]he prior art’s mere disclosure of more than one alternative does not constitute a teaching away from any of these alternatives because such disclosure does not criticize, discredit, or otherwise discourage the solution claimed….” In re Fulton, 391 F.3d 1195, 1201, 73 USPQ2d 1141, 1146 (Fed. Cir. 2004). (MPEP §2124).
In response to Argument 2, With respect to dependent claim 10, the Examiner respectfully disagree. Sorenson discloses a machine learning module that is part of an artificial intelligence findings system includes an image identification engine that can extract features from a new medical image being analyzed to match this data to data present in the archive with the same characteristic (disease) … the similar data can pertain to anatomical structures such as body parts, anatomic anomalies and anatomical features is interpreted as “determining a number of clusters corresponding to different pathologies.” (Paragraph 184). Also, Sorenson discloses an engine developer can train a lung nodule detection engine to detect lung nodules (is interpreted as “pathologies) in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules … a COPD engine can machine learn based on the same COPD engine data (read as cluster), based on another COPD engine data, or any combination thereof. “determining a number of clusters corresponding to different pathologies”) (Paragraph 76). Therefore, it is clearly stated that Sorenson teaches/discloses determining a number of clusters corresponding to different pathologies.
In response to Argument 3, With respect to dependent claim 12, the Examiner respectfully disagree. Sorenson discloses FIG. 19 illustrates a DICOM Header table to map MR series to functions such as Time Volume Analysis (TVA), Flow, DE, and Perfusion by the file analysis module. The file analysis module can create a mapping function (e.g., Ax=B) such that the machine learning module can learn every time image data (e.g., a series) is loaded into the function. This process can be repeated until x converges for all of the label requirements. (read as “descriptors”) For example, in FIG. 19, A=[mr chest heart short_axis 4d no mri cardiac cont flow quantross sax sorted]; Label, B=‘TVA’; and Mapping function, x, such that: Ax=B.Such a process is known as a Bayesian Inference. All of the image information and mapping functions can be tracked in the log file. For example, a similar Bayesian approach (is consider as “Sampling model”) can be performed by the image analysis module during in-image analysis. In one embodiment, the in-image analysis module can determine information such as anatomy or modality based on analysis of the image/image volume. Such image information can be extracted and processes in a similar function is interpreted as “extract the descriptors by sampling the first medical image data, the third medical image data, the first analysis data related to the first medical image data, the third analysis data, or a combination thereof using a sampling model”. (Paragraphs 205-206)
Therefore, it is clearly stated that Sorenson teaches/discloses extract the descriptors by sampling the first medical image data, the third medical image data, the first analysis data related to the first medical image data, the third analysis data, or a combination thereof using a sampling model.
In response to Argument 4, With respect to dependent claim 11, the Examiner respectfully disagree. In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). The examiner states Sorenson, in view of Yerebakan, does teach/disclose on “determine the number of clusters by performing a K-means algorithm”. Moriya is only used to disclose “determine clusters by performing K-means algorithm”, the other limitations “determine the number of clusters” has been disclosed by Sorenson. Since they are in the same field of using image analysis in medical image data, one of ordinary skilled in the art would have been motivated to combine the references since this will improving recognizing similar findings that may be distributed throughout the medical image data.
The Examiner made a proper determination of obviousness under 35 U.S.C. §103, and also provided an appropriate supporting rationale in view of the decision by the Supreme Court in KSR International Co. v. Teleflex Inc. (KSR), 550 U.S. 398, 82 USPQ2d 1385 (2007). The Examiner’s rational are based on the Office’s current understanding of the law, and are believed to be fully consistent with the binding precedent of the Supreme Court. Furthermore, the Examiner supported the rejection under 35 U.S.C. §103 via making the clear articulation of the reason(s) why the claimed invention would have been obvious by citing the specific areas in the prior art references. Further the Examiner, clearly stating the modification of the inventions, supported the rejection under 35 U.S.C. §103 by making the analysis explicit. Last, the Examiner did not make conclusory statements. The Court quoting In re Kahn, 441 F.3d 977, 988, 78 USPQ2d 1329, 1336 (Fed. Cir. 2006), stated that “‘[R]ejections on obviousness cannot be sustained by mere conclusory statements; instead, there must be some articulated reasoning with some rational underpinning to support the legal conclusion of obviousness.’” KSR, 550 U.S. at ___, 82 USPQ2d at 1396. Therefore, the Examiner has established a proper 35 U.S.C. §103 rejection, which is disclosed in detail below.
In response to Argument 5, With respect to dependent claim 13, the Examiner respectfully disagree. the Examiner respectfully disagree. In response to applicant's arguments against the references individually, one cannot show nonobviousness by attacking references individually where the rejections are based on combinations of references. See In re Keller, 642 F.2d 413, 208 USPQ 871 (CCPA 1981); In re Merck & Co., 800 F.2d 1091, 231 USPQ 375 (Fed. Cir. 1986). The examiner states Sorenson, in view of Yerebakan, does teach/disclose “automatically train the first or second number of data analysis tools in response to the number of descriptors in any one cluster exceeds a threshold value”. Yerebakan is only used to teaches “in response to the number of descriptors in any one cluster exceeds exceeding a threshold value.” Yerebakan discloses Other clustering algorithms may be use … the number of clusters can be set by the clustering algorithm in dependence of the characteristics of the data set. Factors that may influence the number of clusters include but are not limited to: the distance threshold between two descriptors for the descriptors to be treated as similar for the purposes of clustering, … the total number of findings within the image (is consider as “number of descriptors in any one cluster exceeds exceeding a threshold value”) (Paragraph 52). Furthermore, Yerebakan is onteaches automates the complex and time-consuming task of sorting and grouping findings and additionally mitigates the need to impose a threshold on the number of medical abnormality candidates. (is consider as “number of descriptors in any one cluster exceeds exceeding a threshold value”). (Paragraph 16). the other limitations “automatically train the first or second number of data analysis tools” has been disclosed by Sorenson.
The Examiner made a proper determination of obviousness under 35 U.S.C. §103, and also provided an appropriate supporting rationale in view of the decision by the Supreme Court in KSR International Co. v. Teleflex Inc. (KSR), 550 U.S. 398, 82 USPQ2d 1385 (2007). The Examiner’s rational are based on the Office’s current understanding of the law, and are believed to be fully consistent with the binding precedent of the Supreme Court. Furthermore, the Examiner supported the rejection under 35 U.S.C. §103 via making the clear articulation of the reason(s) why the claimed invention would have been obvious by citing the specific areas in the prior art references. Further the Examiner, clearly stating the modification of the inventions, supported the rejection under 35 U.S.C. §103 by making the analysis explicit. Last, the Examiner did not make conclusory statements. The Court quoting In re Kahn, 441 F.3d 977, 988, 78 USPQ2d 1329, 1336 (Fed. Cir. 2006), stated that “‘[R]ejections on obviousness cannot be sustained by mere conclusory statements; instead, there must be some articulated reasoning with some rational underpinning to support the legal conclusion of obviousness.’” KSR, 550 U.S. at ___, 82 USPQ2d at 1396. Therefore, the Examiner has established a proper 35 U.S.C. §103 rejection, which is disclosed in detail below.
Claim Status
Claim(s) 1-10, 12 and 14-20 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Sorenson et al (U.S. 20180137244 A1; Sorenson).
Claim(s) 11 and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sorenson et al (U.S. 20180137244 A1; Sorenson), in view of Yerebakan Halid et al (EP – 3869453 A1; Yerebakan).
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.
(a)(2) the claimed invention was described in a patent issued under section 151, or in an application for patent published or deemed published under section 122(b), in which the patent or application, as the case may be, names another inventor and was effectively filed before the effective filing date of the claimed invention.
Claim(s) 1-10, 12 and 14-20 is/are rejected under 35 U.S.C. 102(a)(1) as being anticipated by Sorenson et al (U.S. 20180137244 A1; Sorenson).
Regarding claim 1, Sorenson discloses A system for medical data analysis (Fig.1: a medical data review system, Fig.2 ), comprising: a non-transitory memory for storing machine-readable instructions; and a processing circuit in communication with the non-transitory memory, the processing circuit being operative with the machine-readable instructions (Paragraph 77: “Referring to FIG. 2, image processing server 110 includes memory 201 (e.g., dynamic random access memory or DRAM) hosting one or more image processing engines 113-115, which may be installed in and loaded from persistent storage device 202 (e.g., hard disks), and executed by one or more processors (not shown). “) to perform steps including
determining a number of clusters by grouping descriptors extracted from first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof, ( Paragraph 214-215: “The medical image data can be sent via a network to the Processing Server where the auto-categorization module can categorize each of the images sent to the processing serve …the auto-categorization module can categorize the images based on rules, training based on user, machine learning, DICOM Headers, in-image analysis, … sorting information (e.g., 2D, 2.5D, 3D, 4D), study/series description, scanning protocol, sequences, options, flow data, or any combination thereof.”; it show that the each of categories is interpreted as “ clusters” data”; Paragraphs 83-85; Paragraphs 100-118: “a variety of image processing tools can be accessed by a user using the diagnostic image processing features of the medical data review system. Alternatively, such image processing tools can be implemented as image processing engines 113-115 which are then evoked in other third party systems, such as a PACS or EMR, or other clinical or information system. … Vessel Analysis tools may include a comprehensive vascular analysis package for CT and MR angiography capable of a broad range of vascular analysis tasks, … Calcium scoring tools may include identification of coronary calcium with Agatston, volume and mineral mass algorithms.”, it shows that the set of the diagnostic image processing features is input to each image processing tools as vessel analysis tools or calcium scoring tools is read as “cluster”; Paragraph 183: “A machine learning module can receive image data from a medical image data source. The machine learning module can correlate image data (read as clustering) from the medical image data source to a workflow based on in-image analysis (read as “analysis data”) and metadata ( read as “medical image data”) … the correlation of the medical data to a machine learning module or collection of machine learning can be done based on pattern extraction, feature extraction or image processing which result of a medical image classification (clusterization). … The machine learning module can associate the series with the workflow based on in-image analysis, for example, by reconstructing the image or analyzing the pixel characteristics (e.g., intensity, shading, color, etc.) to determine the anatomy or modality. This correlation can be done based on organ (liver, kidney, heart, brain . . . ), body part (head, head and neck, chest, abdomen . . . ) or general feature extraction”) and
automatically training a first number of data analysis tools (Figs. 1-2: image processing engines 113-114-115) based on the first medical image data and the first analysis data related to the first medical image data, (Paragraphs 65-66: “The engine or e-suites can detect findings (e.g., a disease, an indication, a feature, an object, a shape, a texture, a measurement, insurance fraud, or any combination thereof). The one or more engines and/or one or more e-suites can detect findings from studies (e.g., clinical reports, images, patient data, image data, metadata, or any combination thereof) based on metadata, known methods of in-image analysis, or any combination thereof. … the engines and/or the e-suites can machine learn or be trained using machine tearing algorithms based on prior findings periodically such that as the engines/e-suites process more studies, the engines/e-suites can detect findings more accurately.”; Paragraph 70: “Tools, engines, e-suites, training tools, coding tools, or any combination thereof can be displayed and used via image processing server 110 or in a 2D and/or 3D medical imaging software application, or medical data review system, … The second user or group can use the machine learning/training tools and the feedback from this usage can be applied to train the first engine to detect findings with higher accuracy. The first engine can be updated by image processing server 110 and stored in the application store 109. The processing of image data by engines and updating of the engines can occur at image processing server 110, the image processing application store 109, or any combination thereof.” Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules (e.g., geometric shapes, textures, other combination of features resulting in detection of lung nodules, or any combination thereof))
wherein at least one of the first number of data analysis tools (Figs. 1-2: image processing engines 113-114-115) is trained specifically for each determined cluster using the descriptors within the determined cluster. (Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules … a COPD engine can machine learn based on the same COPD engine data, based on another COPD engine data, or any combination thereof.”, it show that “the same COPD engine data” is interpreted as “determine cluster”; “Paragraph 183: “the machine learning module can read image data from the series list and determine that based on the metadata and in-image analysis of the image data, a specific image data should be recommended to the user for the ejection fraction workflow. This correlation will be reinforce based on end-user feedback up to the point where the machine learning module will be able to automatically select the relevant data (consider as “specifically for each determined cluster) to be processed by other machine learning module or read using a particular clinical protocol. … through numerous iterations, the machine learning module can propose specific image data by machine learning based on, for example, the user's interactions and the information from the in-image analysis, metadata and patient clinical context or exam order”); Paragraph 185: “machine learning module that is part of an artificial intelligence findings system includes an image identification engine that can extract features from a new medical image being analyzed to match this data to data present in the archive with the same characteristic (disease). Additionally, for example, the machine learning module can analyze data to extract a feature and prebuild a hanging protocol for a better viewing experience. Additionally, for example, the image identification engine within the machine learning module can analyze data to extract a feature and find similar data for the same patient to automatically present the data with its relevant prior information. For example, the similar data (read as “ cluster”) can pertain to anatomical structures (read as “descriptors”) such as body parts, anatomic anomalies and anatomical features”)
Regarding claim 2, Sorenson discloses the processing circuit is operative with the machine-readable instructions to perform: selecting at least one of the first number of data analysis tools; and executing the selected at least one data analysis tool to, based on second medical image data, output second analysis data. (Figs. 1-3: medical data source 105, image processing engine as “113. lung nodule -114. bone fracture – 115. vessel detect; Paragraph 69: “The engines or e-suites of image processing server 110 can process studies depending on which engines are selected by the user via a graphical user interface (GUI) or website (local or on the internet) of image processing server 110.” ; Paragraph 92: “the user can select engines that detect specific features of lung nodules, for example, an engine for texture, and engine for nodule shape, an engine for intensity, or any combination thereof. Such engines can be run in parallel, in series, or any combination thereof.”)
Regarding claim 3, Sorenson discloses the processing circuit is operative with the machine-readable instructions to enable selection of the at least one data analysis tool via a user interface. (Figs. 1-3: medical data source 105, image processing engine as “113. lung nodule -114. bone fracture – 115. vessel detect; Paragraph 69: “The engines or e-suites of image processing server 110 can process studies depending on which engines are selected by the user via a graphical user interface (GUI) or website (local or on the internet) of image processing server 110.”)
Regarding claim 4, Sorenson discloses the user interface displays the first medical image data, the second medical image data, third medical image data, or a combination thereof. (Figs. 1-3: medical data source 105; Paragraphs 165, 168, Paragraph 202: “The image processing server can receive image data from the medical image data source. The image processing server (e.g., an engine (not shown) that is part of the image processing server) can analyze and send the image data to the client device to be displayed over a network (e.g., WAN or LAN). The display of the image data and application settings/preferences on the client can be a default preference based on the user's default preferences or the image processing server/client application's default preference.”)
Regarding claim 5, Sorenson discloses the user interface applies a user operated data analysis tool to the first or third medical image data to generate the first analysis data or third analysis data. (Figs. 1-3: medical data source 105; image processing tools 107; Paragraph 70: “A medical imaging software application is a client application that accesses the output of the image processing tools 107 of image processing system 106. … The processing of image data by engines and updating of the engines can occur at image processing server 110, the image processing application store 109, or any combination thereof.”; Paragraph 202: “The image processing server can receive image data from the medical image data source. The image processing server (e.g., an engine (not shown) that is part of the image processing server) can analyze and send the image data to the client device to be displayed over a network (e.g., WAN or LAN).”)
Regarding claim 6, Sorenson discloses the user interface displays the first analysis data, the second analysis data, the third analysis data, or a combination thereof. (Figs. 1-3: medical data source 105; image processing tools 107; Paragraph 70: “A medical imaging software application is a client application that accesses the output of the image processing tools 107 of image processing system 106. … The processing of image data by engines and updating of the engines can occur at image processing server 110, the image processing application store 109, or any combination thereof.”; Paragraph 202: “The image processing server can receive image data from the medical image data source. The image processing server (e.g., an engine (not shown) that is part of the image processing server) can analyze and send the image data to the client device to be displayed over a network (e.g., WAN or LAN).”)
Regarding claim 7, Sorenson discloses the processing circuit is operative with the machine-readable instructions to perform: automatically generating a second number of data analysis tools based on third medical image data and third analysis data related to the third medical image data. (Figs. 1-3: medical data source 105, image processing engine as “113. lung nodule -114. bone fracture – 115. vessel detect; Paragraph 60: “each of image processing engines or modules 113-115 may be configured to perform a specific image processing operation on medical images, such as, for example, lung nodule detection, bone fracture detection, organ identification and segmentation, blood clot detection, image body part categorization, chronic obstructive pulmonary disease (COPD) detection, or soft tissue characterization. An image processing engine can perform such a detection based on the shape, texture, sphericity measurement, color, or other features obtained from the medical images or which are derived or implied by the clinical content.”)
Regarding claim 8, Sorenson discloses the processing circuit is operative with the machine-readable instructions to perform: updating the first number of data analysis tools based on third medical image data and third analysis data related to the third medical image data. (Figs. 1-3: medical data source 105; image processing tools 107; image processing engine as “113. lung nodule – 114. bone fracture- 115. vessel detect; Paragraphs 69-70: “A medical imaging software application is a client application that accesses the output of the image processing tools 107 of image processing system 106. … The processing of image data by engines and updating of the engines can occur at image processing server 110, the image processing application store 109, or any combination thereof.”; Paragraph 72: “Image processing server 110 can have a GUI for one or more users or groups to train, code, develop, upload, delete, track, purchase, update, or process data on engines or e-suites. “;Paragraphs 95-96)
Regarding claim 9, Sorenson discloses the processing circuit is operative with the machine-readable instructions to select at least one of the first and second number of data analysis tools. (Figs. 1-3: medical data source 105, image processing engine as “113. lung nodule -114. bone fracture – 115. vessel detect; Paragraph 69: “The engines or e-suites of image processing server 110 can process studies depending on which engines are selected by the user via a graphical user interface (GUI) or website (local or on the internet) of image processing server 110.”)
Regarding claim 10, Sorenson discloses the clusters correspond to different pathologies. (Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules … a COPD engine can machine learn based on the same COPD engine data based on another COPD engine data, or any combination thereof.”; Paragraphs 183-184: “ a machine learning module that is part of an artificial intelligence findings system includes an image identification engine that can extract features from a new medical image being analyzed to match this data to data present in the archive with the same characteristic (disease) … the similar data can pertain to anatomical structures such as body parts, anatomic anomalies and anatomical features”)
Regarding claim 12, Sorenson discloses the processing circuit is operative with the machine-readable instructions to extract the descriptors by sampling the first medical image data, the third medical image data, the first analysis data related to the first medical image data, the third analysis data, or a combination thereof using a sampling model (Paragraph 197: “the machine learning module can categorize the image data based on in-image analysis and/or metadata (e.g., DICOM headers or tags).The machine learning module can identify any image information from the image data such as the modality, orientation (e.g., axial, coronal, sagittal, off axis, short axis, 3 chamber view, or any combination thereof), anatomies … study/series description, scanning protocol”); Paragraphs 205-206: “FIG. 19 illustrates a DICOM Header table to map MR series to functions such as Time Volume Analysis (TVA), Flow, DE, and Perfusion by the file analysis module. The file analysis module can create a mapping function (e.g., Ax=B) such that the machine learning module can learn every time image data (e.g., a series) is loaded into the function. This process can be repeated until x converges for all of the label requirements. (read as “descriptors”) For example, in FIG. 19, A=[mr chest heart short_axis 4d no mri cardiac cont flow quantross sax sorted]; Label, B=‘TVA’; and Mapping function, x, such that: Ax=B.Such a process is known as a Bayesian Inference. All of the image information and mapping functions can be tracked in the log file. For example, a similar Bayesian approach (is consider as “Sampling model”) can be performed by the image analysis module during in-image analysis. In one embodiment, the in-image analysis module can determine information such as anatomy or modality based on analysis of the image/image volume. Such image information can be extracted and processes in a similar function”; Paragraph 214-215: “the medical image data can be sent via a network to the Processing Server where the auto-categorization module can categorize each of the images sent to the processing server …the auto-categorization module can categorize the images based on rules, training based on user, machine learning, DICOM Headers, in-image analysis, analysis of pixel attributes, landmarks within the images, characterization methods, statistical methods, or any combination thereof.”)
Regarding claim 14, Sorenson discloses the processing circuit is operative with the machine-readable instructions to generate a corresponding data analysis tool by using the descriptors, corresponding images, analysis data corresponding to the images, or a combination thereof, within one cluster. (Paragraph 83; Paragraphs 100-118: “a variety of image processing tools can be accessed by a user using the diagnostic image processing features of the medical data review system. Alternatively, such image processing tools can be implemented as image processing engines 113-115 which are then evoked in other third party systems, such as a PACS or EMR, or other clinical or information system. … Vessel Analysis tools may include a comprehensive vascular analysis package for CT and MR angiography capable of a broad range of vascular analysis tasks, … Calcium scoring tools may include identification of coronary calcium with Agatston, volume and mineral mass algorithms. … Lobular decomposition tools identify tree-like structures within a volume of interest, e.g. a scan region containing a vascular bed … Segmentation, analysis & tracking tools support analysis and characterization of masses and structures, such as solitary pulmonary nodules or other potential lesions.”, it show that each image processing tools as vessel analysis tools or calcium scoring tools is read as “cluster; Paragraph 197: “The machine learning module can identify any image information from the image data such as the modality, orientation … sorting information (e.g., 2D, 2.5D, 3D, 4D), study/series description”)
Regarding claim 15, Sorenson discloses at least one of the first or second number of data analysis tools comprises a neural network, wherein the processing circuit is operative with the machine-readable instructions to generate the corresponding data analysis tool by training the neural network with the first medical image data, the third medical image data or corresponding descriptors as input data and the first medical image data or the third analysis data as desired output data. (Paragraphs 46-47: “a supervised or unsupervised machine learned engine can be used to monitor and learn the effectiveness of various engines in various situations when a first set of medical images associated with a particular clinical study is received from a medical data source, one or more image processing engines are invoked to process (e.g., recognizing shapes, features, trends in the images or other data or measurements) the medical images (or data, used synonymously in this application) according to a predetermined or machine learned suggested order for performing the engine operations that is configured for the particular type of imaging study.”; Paragraph 208: “Other machine learning approaches for in-image analysis and metadata can be implemented such as decision tree learning, association rule learning, artificial neural networks, deep learning … convolutional neural network based on deep learning framework and naïve Bayes classifier, or any combination thereof.”)
Regarding claim 16, Sorenson discloses comprising at least a first and a second client device each connected to the processing circuit via a network, the first client device comprising a first user interface and the second client device comprising a second user interface. (Paragraph 54: “Referring to FIG. 1, medical data review system 100 includes one or more client devices 101-102 communicatively coupled to medical image processing server 110 over network 103. Client devices 101-102 can be a desktop, laptop, mobile device, workstation, etc.”)
Regarding claim 17, Sorenson discloses the first user interface applies the at least one data analysis tool to the first medical image data to generate the first analysis data. (Paragraphs 59-60: “The image processing engines 113-115 can be uploaded and listed in a Web server 109, in this example, an application store, to allow a user of clients 101-102 to purchase, select, and download one or more image processing engines as part of client applications 111-112 respectively. The selected image processing engines can be configured to a variety of configurations (e.g., in series, in parallel, or both) to perform a sequence of one or more image processing operations. … each of image processing engines or modules 113-115 may be configured to perform a specific image processing operation on medical images, such as, for example, lung nodule detection, bone fracture detection,”)
Regarding claim 18, Sorenson discloses wherein the second user interface controls the selection of the at least one data analysis tool after the first number of data analysis tools are generated, the first medical image data and the first analysis data are received by the processing circuit via the network. . (Paragraphs 59-60: “The image processing engines 113-115 can be uploaded and listed in a Web server 109, in this example, an application store, to allow a user of clients 101-102 to purchase, select, and download one or more image processing engines as part of client applications 111-112 respectively. The selected image processing engines can be configured to a variety of configurations (e.g., in series, in parallel, or both) to perform a sequence of one or more image processing operations. … each of image processing engines or modules 113-115 may be configured to perform a specific image processing operation on medical images, such as, for example, lung nodule detection, bone fracture detection,”)
Regarding claim 19, A computer-implemented method of medical data analysis, (Fig.1: a medical data review system, Paragraph 77: “Referring to FIG. 2, image processing server 110 includes memory 201 (e.g., dynamic random access memory or DRAM) hosting one or more image processing engines 113-115, which may be installed in and loaded from persistent storage device 202 (e.g., hard disks), and executed by one or more processors (not shown). “) comprising:
determining a number of clusters by grouping descriptors extracted from first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof, ( Paragraph 214-215: “The medical image data can be sent via a network to the Processing Server where the auto-categorization module can categorize each of the images sent to the processing serve …the auto-categorization module can categorize the images based on rules, training based on user, machine learning, DICOM Headers, in-image analysis, … sorting information (e.g., 2D, 2.5D, 3D, 4D), study/series description, scanning protocol, sequences, options, flow data, or any combination thereof.”; it show that the each of categories is interpreted as “ clusters” data”; Paragraphs 83-85; Paragraphs 100-118: “a variety of image processing tools can be accessed by a user using the diagnostic image processing features of the medical data review system. Alternatively, such image processing tools can be implemented as image processing engines 113-115 which are then evoked in other third party systems, such as a PACS or EMR, or other clinical or information system. … Vessel Analysis tools may include a comprehensive vascular analysis package for CT and MR angiography capable of a broad range of vascular analysis tasks, … Calcium scoring tools may include identification of coronary calcium with Agatston, volume and mineral mass algorithms.”, it shows that the set of the diagnostic image processing features is input to each image processing tools as vessel analysis tools or calcium scoring tools is read as “cluster”; Paragraph 183: “A machine learning module can receive image data from a medical image data source. The machine learning module can correlate image data (read as clustering) from the medical image data source to a workflow based on in-image analysis (read as “analysis data”) and metadata ( read as “medical image data”) … the correlation of the medical data to a machine learning module or collection of machine learning can be done based on pattern extraction, feature extraction or image processing which result of a medical image classification (clusterization). … The machine learning module can associate the series with the workflow based on in-image analysis, for example, by reconstructing the image or analyzing the pixel characteristics (e.g., intensity, shading, color, etc.) to determine the anatomy or modality. This correlation can be done based on organ (liver, kidney, heart, brain . . . ), body part (head, head and neck, chest, abdomen . . . ) or general feature extraction”) and
automatically training a first number of data analysis tools (Figs. 1-2: image processing engines 113-114-115) based on the first medical image data and the first analysis data related to the first medical image data, (Paragraphs 65-66: “The engine or e-suites can detect findings (e.g., a disease, an indication, a feature, an object, a shape, a texture, a measurement, insurance fraud, or any combination thereof). The one or more engines and/or one or more e-suites can detect findings from studies (e.g., clinical reports, images, patient data, image data, metadata, or any combination thereof) based on metadata, known methods of in-image analysis, or any combination thereof. … the engines and/or the e-suites can machine learn or be trained using machine tearing algorithms based on prior findings periodically such that as the engines/e-suites process more studies, the engines/e-suites can detect findings more accurately.”; Paragraph 70: “Tools, engines, e-suites, training tools, coding tools, or any combination thereof can be displayed and used via image processing server 110 or in a 2D and/or 3D medical imaging software application, or medical data review system, … The second user or group can use the machine learning/training tools and the feedback from this usage can be applied to train the first engine to detect findings with higher accuracy. The first engine can be updated by image processing server 110 and stored in the application store 109. The processing of image data by engines and updating of the engines can occur at image processing server 110, the image processing application store 109, or any combination thereof.” Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules (e.g., geometric shapes, textures, other combination of features resulting in detection of lung nodules, or any combination thereof))
wherein at least one of the first number of data analysis tools is trained specifically for each determined cluster using the descriptors within the determined cluster . (Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules … a COPD engine can machine learn based on the same COPD engine data, based on another COPD engine data, or any combination thereof.”, it show that “the same COPD engine data” is interpreted as “determine cluster”; Paragraph 208: “Other machine learning approaches for in-image analysis and metadata can be implemented such as decision tree learning, association rule learning, artificial neural networks, deep learning, inductive logic programming, support vector machines, clustering,”; Paragraph 183: “the machine learning module can read image data from the series list and determine that based on the metadata and in-image analysis of the image data, a specific image data should be recommended to the user for the ejection fraction workflow. This correlation will be reinforce based on end-user feedback up to the point where the machine learning module will be able to automatically select the relevant data (consider as “specifically for each determined cluster) to be processed by other machine learning module or read using a particular clinical protocol. … through numerous iterations, the machine learning module can propose specific image data by machine learning based on, for example, the user's interactions and the information from the in-image analysis, metadata and patient clinical context or exam order”); Paragraph 185: “machine learning module that is part of an artificial intelligence findings system includes an image identification engine that can extract features from a new medical image being analyzed to match this data to data present in the archive with the same characteristic (disease). Additionally, for example, the machine learning module can analyze data to extract a feature and prebuild a hanging protocol for a better viewing experience. Additionally, for example, the image identification engine within the machine learning module can analyze data to extract a feature and find similar data for the same patient to automatically present the data with its relevant prior information. For example, the similar data (read as “ cluster”) can pertain to anatomical structures (read as “descriptors”) such as body parts, anatomic anomalies and anatomical features”)
Regarding claim 20, One or more non-transitory computer-readable media comprising computer-readable instructions, that when executed by one or more processing units, (Fig.1: a medical data review system, Paragraph 77: “Referring to FIG. 2, image processing server 110 includes memory 201 (e.g., dynamic random access memory or DRAM) hosting one or more image processing engines 113-115, which may be installed in and loaded from persistent storage device 202 (e.g., hard disks), and executed by one or more processors (not shown). “) cause the one or more processing units to perform steps comprising:
determining a number of clusters by grouping descriptors extracted from first medical image data, third medical image data, first analysis data related to the first medical image data, third analysis data, or a combination thereof, ( Paragraph 214-215: “The medical image data can be sent via a network to the Processing Server where the auto-categorization module can categorize each of the images sent to the processing serve …the auto-categorization module can categorize the images based on rules, training based on user, machine learning, DICOM Headers, in-image analysis, … sorting information (e.g., 2D, 2.5D, 3D, 4D), study/series description, scanning protocol, sequences, options, flow data, or any combination thereof.”; it show that the each of categories is interpreted as “ clusters” data”; Paragraphs 83-85; Paragraphs 100-118: “a variety of image processing tools can be accessed by a user using the diagnostic image processing features of the medical data review system. Alternatively, such image processing tools can be implemented as image processing engines 113-115 which are then evoked in other third party systems, such as a PACS or EMR, or other clinical or information system. … Vessel Analysis tools may include a comprehensive vascular analysis package for CT and MR angiography capable of a broad range of vascular analysis tasks, … Calcium scoring tools may include identification of coronary calcium with Agatston, volume and mineral mass algorithms.”, it shows that the set of the diagnostic image processing features is input to each image processing tools as vessel analysis tools or calcium scoring tools is read as “cluster”; Paragraph 183: “A machine learning module can receive image data from a medical image data source. The machine learning module can correlate image data (read as clustering) from the medical image data source to a workflow based on in-image analysis (read as “analysis data”) and metadata ( read as “medical image data”) … the correlation of the medical data to a machine learning module or collection of machine learning can be done based on pattern extraction, feature extraction or image processing which result of a medical image classification (clusterization). … The machine learning module can associate the series with the workflow based on in-image analysis, for example, by reconstructing the image or analyzing the pixel characteristics (e.g., intensity, shading, color, etc.) to determine the anatomy or modality. This correlation can be done based on organ (liver, kidney, heart, brain . . . ), body part (head, head and neck, chest, abdomen . . . ) or general feature extraction”) and
automatically training a first number of data analysis tools (Figs. 1-2: image processing engines 113-114-115) based on the first medical image data and the first analysis data related to the first medical image data, (Paragraphs 65-66: “The engine or e-suites can detect findings (e.g., a disease, an indication, a feature, an object, a shape, a texture, a measurement, insurance fraud, or any combination thereof). The one or more engines and/or one or more e-suites can detect findings from studies (e.g., clinical reports, images, patient data, image data, metadata, or any combination thereof) based on metadata, known methods of in-image analysis, or any combination thereof. … the engines and/or the e-suites can machine learn or be trained using machine tearing algorithms based on prior findings periodically such that as the engines/e-suites process more studies, the engines/e-suites can detect findings more accurately.”; Paragraph 70: “Tools, engines, e-suites, training tools, coding tools, or any combination thereof can be displayed and used via image processing server 110 or in a 2D and/or 3D medical imaging software application, or medical data review system, … The second user or group can use the machine learning/training tools and the feedback from this usage can be applied to train the first engine to detect findings with higher accuracy. The first engine can be updated by image processing server 110 and stored in the application store 109. The processing of image data by engines and updating of the engines can occur at image processing server 110, the image processing application store 109, or any combination thereof.” Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules (e.g., geometric shapes, textures, other combination of features resulting in detection of lung nodules, or any combination thereof))
wherein at least one of the first number of data analysis tools is trained specifically for each determined cluster using the descriptors within the determined cluster.. (Paragraph 76: “an engine developer can train a lung nodule detection engine to detect lung nodules in studies on the developer platform of the application store 109 or a data analytic system (not shown) by training the engine based on various features of detecting lung nodules … a COPD engine can machine learn based on the same COPD engine data, based on another COPD engine data, or any combination thereof.”, it show that “the same COPD engine data” is interpreted as “determine cluster”; Paragraph 208: “Other machine learning approaches for in-image analysis and metadata can be implemented such as decision tree learning, association rule learning, artificial neural networks, deep learning, inductive logic programming, support vector machines, clustering,”; Paragraph 183: “the machine learning module can read image data from the series list and determine that based on the metadata and in-image analysis of the image data, a specific image data should be recommended to the user for the ejection fraction workflow. This correlation will be reinforce based on end-user feedback up to the point where the machine learning module will be able to automatically select the relevant data (consider as “specifically for each determined cluster) to be processed by other machine learning module or read using a particular clinical protocol. … through numerous iterations, the machine learning module can propose specific image data by machine learning based on, for example, the user's interactions and the information from the in-image analysis, metadata and patient clinical context or exam order”); Paragraph 185: “machine learning module that is part of an artificial intelligence findings system includes an image identification engine that can extract features from a new medical image being analyzed to match this data to data present in the archive with the same characteristic (disease). Additionally, for example, the machine learning module can analyze data to extract a feature and prebuild a hanging protocol for a better viewing experience. Additionally, for example, the image identification engine within the machine learning module can analyze data to extract a feature and find similar data for the same patient to automatically present the data with its relevant prior information. For example, the similar data (read as “ cluster”) can pertain to anatomical structures (read as “descriptors”) such as body parts, anatomic anomalies and anatomical features”)
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.
Claim(s) 11 and 13 is/are rejected under 35 U.S.C. 103 as being unpatentable over Sorenson et al (U.S. 20180137244 A1; Sorenson), in view of Yerebakan Halid et al (EP – 3869453; Yerebakan).
Regarding claim 11, Sorenson discloses all the claims invention except wherein the processing circuit is operative with the machine-readable instructions to determine the number of clusters (Paragraph 183: “the machine learning module can correlate image data from the medical image data source to a workflow based on in-image analysis and metadata. … The correlation of the medical data to a machine learning module or collection of machine learning can be done based on pattern extraction, feature extraction or image processing which result of a medical image classification (clusterization)”)
However, Sorenson does not discloses determine the number of clusters by performing a K-means algorithm.
Yerebakan discloses the processing circuit is operative with the machine-readable instructions to determine the number of clusters by performing a K-means algorithm. (Paragraphs 44-45 : “At step 204 of the medical image processing method 200, a clustering algorithm is applied to the plurality of sets of data generated in step 202 to generate a plurality of groups, based on a similarity between image descriptors of the plurality of sets of data. Each of the plurality of groups includes one or more of the plurality of sets of data. … the clustering algorithm uses an iterative refinement process, such as a "k-means" algorithm”)
Therefore, it would been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to modify the invention of Sorenson by the clustering processing that is taught by Yerebakan, to make the invention that processing medical image data by grouping together findings using clustering techniques; thus, one of ordinary skilled in the art would have been motivated to combine the references since this will improving recognizing similar findings that may be distributed throughout the medical image data as well as reducing time-consuming task of sorting and grouping findings in the medical image data. (Yerebakan: Paragraphs 15-16)
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filling date of the claimed invention.
Regarding claim 13, Sorenson discloses the processing circuit is operative with the machine-readable instructions to automatically train the first or second number of data analysis tools (Paragraphs 65-66: “The engine or e-suites can detect findings (e.g., a disease, an indication, a feature, an object, a shape, a texture, a measurement, insurance fraud, or any combination thereof). The one or more engines and/or one or more e-suites can detect findings from studies (e.g., clinical reports, images, patient data, image data, metadata, or any combination thereof) based on metadata, known methods of in-image analysis, or any combination thereof. … the engines and/or the e-suites can machine learn or be trained using machine tearing algorithms based on prior findings periodically such that as the engines/e-suites process more studies, the engines/e-suites can detect findings more accurately.”)
However, Sorenson does not disclose automatically train the first or second number of data analysis tools when the number of descriptors in any one cluster exceeds a threshold value.
Yerebakan discloses the processing circuit is operative with the machine-readable instructions to automatically train the first or second number of data analysis tools (Paragraphs 85-86 : “ By generating and visually indicating groups of image patches with similar features, identification of these different conditions may be facilitated … mitigate the need to cap the number of medical abnormality candidates that, for example, an automated algorithm may generate. Because image patches having similar features are grouped together, potentially irrelevant image patches (e.g. image patches that do not indicate an illness or injury) have a chance of being grouped together, and thus easily identified.”) when the number of descriptors in any one cluster exceeds a threshold value. (Paragraph 52 : “the number of clusters can be set by the clustering algorithm in dependence of the characteristics of the data set. Factors that may influence the number of clusters include but are not limited to: the distance threshold between two descriptors for the descriptors to be treated as similar for the purposes of clustering, the degree of similarity between two descriptors, and the total number of findings within the image.”; Paragraph 16: “complex and time-consuming task of sorting and grouping findings and additionally mitigates the need to impose a threshold on the number of medical abnormality candidates. Grouping findings according to similarity between image descriptors enables different types of medical abnormality candidates to be grouped together”)
Therefore, it would been obvious to one having ordinary skill in the art before the effective filling date of the claimed invention to modify the invention of Sorenson by the clustering processing that is taught by Yerebakan, to make the invention that processing medical image data by grouping together findings using clustering techniques; thus, one of ordinary skilled in the art would have been motivated to combine the references since this will improving recognizing similar findings that may be distributed throughout the medical image data as well as reducing time-consuming task of sorting and grouping findings in the medical image data. (Yerebakan: Paragraphs 15-16)
Thus, the claimed subject matter would have been obvious to a person having ordinary skill in the art before the effective filling date of the claimed invention.
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Namer Yelin et al (U.S. 20120197619 A1), “System and Method for Generating a Patient-Specific Digital Image-Based Model of An Anatomical Structure”, teaches about The method may comprise receiving medical image data and metadata of a specific patient. A patient-specific digital image-based model of an anatomical structure may be generated based on the medical image data and the metadata. A computerized simulation of an image-guided procedure may be performed using the digital image-based model and the metadata.
Campanatti , Jr et al (U.S 20140087342 A1), “Training and Testing System for Advanced Image Processing”, teaches about medical image processing training including at least one medical image associated with a medical image processing training course (MIPTC) is displayed in a first display area. An instruction is displayed in a second display area, where the instruction requests a user to perform a quantitative determination on at least a portion of a body part within the medical image displayed in the first display area. In response to a user action from the user, the requested determination is performed on the displayed medical image. It is determined automatically without user intervention at least one quantitative value representing a result of the user action. The quantitative value is compared to a predefined model answer.
Krishna et al (U.S. 20150227702 A1), “Multi-Factor Brain Analysis Via Medical Imaging Decision Support System and Methods”, teaches about A medical imaging decision support system is provided that can conduct, and help medical professionals conduct multi-factor brain analysis. Data for disparate processing modes (for example, EEG, MRI, etc.) can be input to the system, processed in parallel in a cloud environment, and the results can be rendered in a thin client (for example, browser) for a user's rapid multi-modal evaluation of a brain.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Duy A Tran whose telephone number is (571)272-4887. The examiner can normally be reached Monday-Friday 8:00 am - 5:00 pm.
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, ONEAL R MISTRY can be reached at (313)-446-4912. 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.
/DUY TRAN/Examiner, Art Unit 2674
/ONEAL R MISTRY/Supervisory Patent Examiner, Art Unit 2674
Prosecution Insights