# US 10,504,197 B2 · Clinical decision support systems and methods > Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt - Number: US 10,504,197 B2 - Document ID: US10504197B2 - Status: Granted - Kind: B2 - Application number: 13/264,657 - Filed: 2010-03-09 - Published: 2012-02-16 - Granted: 2019-12-10 - Earliest claimed: 2009-04-15 - Assignees: Koninklijke Philips N.V. - Applicants: Lilla Boroczky; Ye Xu; Jeremy Drysdale; Michael Chun-chieh Lee; Lina Arbash Meinel; Thomas Buelow - Inventors: Lilla Boroczky; Ye Xu; Jeremy Drysdale; Michael Chun-chieh Lee; Lina Arbash Meinel; Thomas Buelow - Web page: https://patentgrep.com/patent/US10504197B2 ## CPC - G16H 30/20: for handling medical images, e.g. DICOM, HL7 or PACS - G16H 40/63: for local operation - G16H 50/20: for computer-aided diagnosis, e.g. based on medical expert systems - G16H: HEALTHCARE INFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR THE HANDLING OR PROCESSING OF MEDICAL OR HEALTHCARE DATA ## Abstract A clinical decision support (CDS) system comprises: a case grouping sub-system (10) including a graphical user interface (30) that is operative to simultaneously display data representing a plurality of patient cases and further configured to enable a user to group selected patient cases represented by the simultaneously displayed data into clinically related groups (32) as selected by the user; a probative features determination sub-system (12) that is operative to determine probative features (44) that correlate with the clinically related groups; and a CDS user interface (16) that is operative to receive current patient data relating to a current patient case and to output clinical decision support information based on values of the probative features determined from the received current patient data. ## Claims ### Claim 1 A clinical decision support (CDS) system comprising: a case grouping sub-system including a graphical user interface that simultaneously displays data representing a plurality of patient cases and receives from a user a selection of patient cases represented by the simultaneously displayed data to form each of a plurality of user-defined clinically related groups; a probative features determination sub-system that determines probative features that distinguish among the clinically related group using a machine learning component that implements an algorithm or combination of algorithms selected from a group consisting of: a genetic algorithm, a k-nearest neighbor algorithm, a k-mean clustering algorithm, a simulated annealing algorithm, an exhaustive search algorithm, a feature reduction algorithm, and a floating search algorithm; a CDS user interface that receives current patient data relating to a current patient case; and a CDS inference engine that: identifies a select clinically related group based on values of the probative features determined from the received current patient data; determines clinical decision support information based on one or more treatments associated with the select clinically related group; and provides the clinical decision support information to the user for prescribing to the current patient, wherein the clinical decision support information includes a medical regime to be applied corresponding to a prior medical regime of one or more prior patient cases having values of probative features most closely related to the values of the probative features determined from the current patient data. ### Claim 2 (depends on 1) The CDS system of claim 1, wherein the case grouping sub-system provides drag-and-drop case grouping in which the user drags displayed data representing a patient case from a source visual display region to a region representing the clinically related group. ### Claim 3 (depends on 2) The CDS system of claim 2, wherein the case grouping sub-system provides a drag-and-drop grouping in which a user drags displayed data representing a patient case from a source visual display region to a new group creation region to create an other clinically related group. ### Claim 4 (depends on 1) The CDS system of claim 1, wherein the displayed data representing a patient case comprises a patient image or a thumbnail image generated from a patient image. ### Claim 5 (depends on 4) The CDS system of claim 4, wherein the patient image includes one or more of: a computed tomography (CT) image, a magnetic resonance (MR) image, a positron emission tomography (PET) image, a single photon emission computed tomography (SPECT) image, a fluoroscopy image, an ultrasound image, and an x-ray image. ### Claim 6 (depends on 1) The CDS system of claim 1, wherein the case grouping sub-system concurrently displays patient case icons comprising patient images or thumbnail images generated from patient images and enables a user to graphically group selected patient case icons to form the user-defined clinically related group. ### Claim 7 (depends on 6) The CDS system of claim 6, wherein the patient images are one or more of: computed tomography (CT) images, magnetic resonance (MR) images, positron emission tomography (PET) images, single photon emission computed tomography (SPECT) images, fluoroscopy images, ultrasound images, and x-ray images. ### Claim 8 (depends on 1) The CDS system of claim 1, wherein the machine learning component determines the probative features based on values for different patient cases that correlate with the clinically related group related to each patient case. ### Claim 9 (depends on 8) The CDS system of claim 8, wherein the machine learning component determines a distance or similarity function associating the probative features to the clinically related groups. ### Claim 10 (depends on 1) The CDS system of claim 1, wherein the machine learning component determines probative feature weights. ### Claim 11 (depends on 1) The CDS system of claim 1, wherein the probative features determination sub-system includes an image features extraction component that extracts image features from a patient image, the machine learning component operating on at least the image features. ### Claim 12 (depends on 1) The CDS system of claim 1, wherein the inference engine that determines a distance or similarity measure associated with each clinically related group based on values of the probative features determined from the received current patient data and identifies the select clinically related group based on the distance or similarity measure. ### Claim 13 A clinical decision support (CDS) method implemented by a CDS system, the CDS method comprising: providing, by a processor of the CDS system, a graphical user interface to a user, the graphical user interface graphically representing patient cases on a display device by patient case icons comprising patient images or thumbnail images generated from patient images; receiving, from the user via the graphical user interface, a selection of patient cases that form each of a plurality of user-defined clinically related groups; determining, by the processor, probative features of the patient cases having values that distinguish among the user-defined clinically related groups using a machine learning component that implements an algorithm or combination of algorithms selected from a group consisting of: a genetic algorithm, a k-nearest neighbor algorithm; a k-mean clustering algorithm, a simulated annealing algorithm, an exhaustive search algorithm, a feature reduction algorithm, and a floating search algorithm; receiving current patient data related to a current patient; identifying, by the processor, a select clinically related group of the plurality of clinically related groups based on values of the probative features determined from the received current patient data; determining clinical decision support information for the current patient based on one or more treatments associated with the select clinically related group; and providing the clinical decision support information to the user for prescribing to the current patient; wherein the clinical decision support information includes a medical regime to be applied corresponding to a prior medical regime of one or more prior patient cases having values of probative features most closely related to the values of the probative features determined from the current patient data. ### Claim 14 (depends on 13) The CDS method of claim 13, wherein the graphical user interface provides drag-and-drop manipulation of the patient case icons by which a user groups patient cases into each clinically related group. ### Claim 15 (depends on 13) The CDS method of claim 13, wherein the graphical user interface displays non-image patient case information responsive to user selection of a patient case icon via the graphical user interface. ### Claim 16 (depends on 1) The CDS method of claim 1, wherein the determining of the probative features includes extracting image features from patient images, the image features being processed by the machine learning component. ### Claim 17 (depends on 16) The CDS method of claim 16, wherein the machine learning component further processes at least some non-image features. ### Claim 18 A non-transitory computer readable medium that includes instructions executable by a digital processor to: provide a graphical user interface by which a user groups patient cases into a plurality of user-defined clinically related groups, the graphical user interface graphically representing patient cases by patient case icons comprising patient images or thumbnail images generated from patient images; determine probative features of the patient cases having values that distinguish among the user-defined clinically related groups using a machine learning component that implements an algorithm or combination of algorithms selected from a group consisting of: a genetic algorithm, a k-nearest neighbor algorithm, a k-mean clustering algorithm, a simulated annealing algorithm, an exhaustive search algorithm, a feature reduction algorithm, and a floating search algorithm; receive current patient data related to a current patient; identify a select clinically related group of the plurality of clinically related groups based on values of the probative features determined from the received current patient data, determine clinical decision support information for the current patient based on one or more treatments associated with the select clinically related group; and provide the clinical decision support information to the user for prescribing to the current patient; wherein the clinical decision support information includes a medical regime to be applied corresponding to a prior medical regime of one or more prior patient cases having values of probative features most closely related to the values of the probative features determined from the current patient data. ### Claim 19 (depends on 17) The medium of claim 17, wherein the machine learning component determines a distance or similarity measure associated with each clinically related group based on values of the probative features determined from the received current patient data and identifies the select clinically related group based on the distance or similarity measure. ### Claim 20 (depends on 17) The medium of claim 17, wherein the graphical user interface provides drag-and-drop manipulation of the patient case icons by which a user groups patient cases into each clinically related group.