Imaging related clinical context apparatus and associated methods

A computer-readable storage medium including instructions which, when executed, cause at least one processor to at least:

• execute a scheduler to perform a prefetch operation for clinical data located in a plurality of documents residing in one or more connected systems;
• and
• analyze a plurality of documents to identify a subset of relevant documents in the plurality of documents by:
• applying natural language processing to identify terms in the plurality of documents, a subset of the identified terms forming tagged concepts;
• constructing a data structure of the tagged concepts;
• processing the identified terms using a machine learning model with respect to a relevancy criterion for an examination to select the subset of relevant documents, the machine learning model trained using at least radiologist usage patterns and keywords and deployed with the at least one processor to leverage the data structure of the tagged concepts to process the identified terms to select the subset of relevant documents;
• executing a gap analysis in real-time to select additional relevant documents for automated review, the gap analysis to identify, retrieve, and process, using the machine learning model, additional clinical data entered after the prefetch operation, wherein the additional clinical data indicates whether to administer a contrast agent to a patient, an environmental context for the examination, the environmental context comprising an inpatient, an outpatient, or an emergency department (ED) context, a disease category representing a neoplasm category, and a phase of treatment comprising a screening, a follow-up, a pre-surgery treatment, or a post-surgery treatment;
• modifying the subset of relevant documents by adding an emphasis to the tagged concepts found in the subset of relevant documents and the additional clinical data;
• and
• output the subset of relevant documents including emphasized tagged concepts.

A computer-implemented method comprising:

• executing a scheduler to perform a prefetch operation for clinical data located in a plurality of documents residing in one or more connected systems;
• and
• analyzing, by executing an instruction with a processor, a plurality of documents to identify a subset of relevant documents in the plurality of documents by:
• applying natural language processing to identify terms in the plurality of documents, a subset of the identified terms forming tagged concepts;
• constructing a data structure of the tagged concepts;
• processing the identified terms using a machine learning model with respect to a relevancy criterion for an examination to select the subset of relevant documents, the machine learning model trained using at least radiologist usage patterns and keywords and deployed with the at least one processor to leverage the data structure of the tagged concepts to process the identified terms to select the subset of relevant documents;
• executing a gap analysis in real-time to select additional relevant documents for automated review, the gap analysis to identify, retrieve, and process, using the machine learning model, additional clinical data entered after the prefetch operation, wherein the additional clinical data indicates whether to administer a contrast agent to a patient, an environmental context for the examination, the environmental context comprising an inpatient, an outpatient, or an emergency department (ED) context, a disease category representing a neoplasm category, and a phase of treatment comprising a screening, a follow-up, a pre-surgery treatment, or a post-surgery treatment;
• modifying the subset of relevant documents by adding an emphasis to the tagged concepts found in the subset of relevant documents and the additional clinical data;
• and
• outputting, by executing an instruction with the processor, the subset of relevant documents including emphasized tagged concepts.