Automated generation of medical training data for training AI-algorithms for supporting clinical reporting and documentation

A computer system for medical imaging, the computer system comprising:

• memory comprising instructions;
• and
• processing circuitry configured to execute the instructions to:
• a. receive medical image data of a patient comprising medical image data elements;
• b. at least one of generate, edit, or receive a synoptic structured medical findings report comprising a hierarchical tree-like data structure which comprises machine-readable findings-report elements, wherein;
• the contents of the findings-report elements each comprise at least one semantic feature providing information about at least one of the patient or the received medical image data;
• the tree-like data structure reflects dependencies of the provided information;
• and
• each findings-report element and the contents of each findings-report element is assigned to a unique identifier of a database of identifiers each uniquely representing a medical semantic content of an individual medical finding;
• c. annotate the received medical image data by linking one of the medical image data elements to the unique identifier to which the findings-report element is assigned, the contents of which provides information about the medical image data element;
• d. store the annotated received medical image data as training data for an artificial intelligence (AI) algorithm;
• and
• e. cause the AI algorithm to be trained using the training data such that the AI algorithm can be used to generate a medical image of a second patient.

A computer system for medical imaging, the computer system comprising:

• memory comprising instructions;
• and
• processing circuitry configured to execute the instructions to:
• a. receive medical image data of a patient comprising medical image data elements;
• b. at least one of generate, edit, or receive a synoptic structured medical findings report comprising a hierarchical tree-like data structure which comprises machine-readable findings-report elements, wherein:
• the contents of the findings-report elements each comprise at least one semantic feature providing information about at least one of the patient or the received medical image data;
• the tree-like data structure reflects dependencies of the provided information;
• and
• each findings-report element and the contents of each findings-report element is assigned to a unique identifier of a database of identifiers each uniquely representing a medical semantic content of an individual medical finding;
• c. annotate the received medical image data by linking one of the medical image data elements with the unique identifier to which the findings-report element is assigned, the contents of which provides information about the medical image data element;
• d. store the annotated received medical image data as first training data for an artificial intelligence (AI) algorithm;
• e. at least one of generate, edit, or receive at least one of classification or segmentation data comprising segmentation or classification information associated with at least one of the received medical image data or the patient;
• f. further annotate the received medical image data by linking the medical image data element with one of the classification or segmentation data;
• g. link the classification or segmentation data with the findings-report element and the contents of the findings-report element;
• h. store, as second training data for the AI algorithm, at least one of the further annotated received medical image data, or the linked classification or segmentation data and findings-report element and the contents of the findings-report element;
• and
• i. cause the AI algorithm to be trained using the first training data and the second training data such that the AI algorithm can be used to generate a medical image of a second patient.

A non-transitory computer-readable storage medium having instructions stored thereon that, when executed by processing circuitry, cause the processing circuitry to:

• a) receive medical image data of a patient comprising medical image data elements;
• b) at least one of generate, edit, or receive a synoptic structured medical findings report that summarizes individual medical findings and comprises a hierarchical tree-like data structure which comprises machine-readable findings-report elements, wherein:
• the contents of the findings-report elements each comprise at least one semantic feature providing information about at least one of the patient or the received medical image data;
• the tree-like data structure reflects dependencies of the provided information;
• and
• each findings-report element and the contents of each findings-report element is assigned to a unique identifier of a database of identifiers each uniquely representing a medical semantic content of an individual medical finding;
• c) annotate the medical image data by linking one of the medical image data elements to the unique identifier to which the findings-report element is assigned, the contents of which provides information about the medical image data element;
• d) store the annotated received medical image data as training data for an artificial intelligence (AI) algorithm;
• and
• e) cause the AI algorithm to be trained using the training data such that the AI algorithm can be used to generate a medical image of a second patient.