Medical information processing system and method

1 . A medical information processing system comprising:

a data store storing a medical ontology, knowledge graph or other knowledge base; and

processing circuitry configured to:

receive medical image data;

receive an input regarding a position on the medical image data;

use semantic functionality to map the input to at least one anatomical region;

map the at least one anatomical region to a concept in the medical ontology, knowledge graph or other knowledge base which includes or is related to the position;

determine at least one surface form for the concept;

perform a semantic search, using the concept, in one or more pre-existing clinical notes or pre-existing health records for a patient that existed before the medical image data is received, the semantic search comprising:

searching for the at least one surface form in text data of the one or more pre-existing clinical notes or pre-existing health records;

identifying a portion of the one or more pre-existing clinical notes or pre-existing health records including the at least one surface form; and

displaying the portion with the at least one surface form highlighted; and

use a concept distance measure in a concept space of the medical ontology, knowledge graph or other knowledge base to rank and/or filter concepts and/or terms, wherein

the concept space has nodes, each node representative of a concept,

the nodes are connected by edges, each edge of the edges representative of a relationship between two nodes that the edge connects, and

a result of the concept distance measure is expressed as a number of the edges.

2 . A system according to claim 1 , wherein the specified text data further comprises one or more synonyms and/or hyponyms for the specified concept.

3 . A system according to claim 1 , wherein the specifying of the concept comprises:

determining based on the position and the medical image data at least one anatomical region; and

mapping the at least one anatomical region to the medical ontology, knowledge graph or other knowledge base to obtain the specified concept.

4 . A system according to claim 3 , wherein the at least one anatomical region comprises at least one anatomical landmark.

5 . A system according to claim 3 , wherein the determining of the at least one anatomical region is dependent on a segmentation.

6 . A system according to claim 3 , wherein the specifying of the concept is based on a distance in image space from the position to the at least one anatomical region.

7 . A system according to claim 1 , wherein the receiving of the input regarding position comprises receiving user input that is representative of a selection of a point or region on the medical image data by a user.

8 . A medical information processing method comprising:

receiving medical image data;

receiving an input regarding a position on the medical image data;

using semantic functionality to map the input to at least one anatomical region;

mapping the at least one anatomical region to a concept in a medical ontology, knowledge graph or other knowledge base which includes or is related to the position;

determining at least one surface form for the concept;

performing a semantic search, using the concept, in one or more pre-existing clinical notes or pre-existing health records for a patient that existed before the medical image data is received, the semantic search comprising:

searching for the at least one surface form in text data of the one or more pre-existing clinical notes or pre-existing health records;

identifying a portion of the one or more pre-existing clinical notes or pre-existing health records including the at least one surface form; and

displaying the portion with the at least one surface form highlighted; and

using a concept distance measure in a concept space of medical ontology, knowledge graph or other knowledge base to rank and/or filter concepts and/or terms, wherein

the concept space has nodes, each node representative of a concept,

the nodes are connected by edges, each edge of the edges representative of a relationship between two nodes that the edge connects, and

a result of the concept distance measure is expressed as a number of the edges.

9 . A medical information processing system comprising:

a data store storing a medical ontology, knowledge graph or other knowledge base; and

processing circuitry configured to:

receive a text query and medical image data;

specify a node of the medical ontology, knowledge graph or other knowledge base based on the text query;

use semantic functionality to map a concept corresponding to the specified node and/or a related concept to at least one anatomical region;

specify a point or region on the medical image data corresponding to the at least one anatomical region;

determine at least one surface form for the concept;

perform a semantic search, using the concept, in one or more pre-existing clinical notes or pre-existing health records for a patient that existed before the medical image data is received, the semantic search comprising:

searching for the at least one surface form in text data of the one or more pre-existing clinical notes or pre-existing health records;

identifying a portion of the one or more pre-existing clinical notes or pre-existing health records including the at least one surface form; and

displaying the portion with the at least one surface form highlighted; and

use a concept distance measure in a concept space of the medical ontology, knowledge graph or other knowledge base to rank and/or filter concepts and/or terms, wherein

the concept space has nodes, each node representative of a concept,

the nodes are connected by edges, each edge of the edges representative of a relationship between two nodes that the edge connects, and

a result of the concept distance measure is expressed as a number of the edges.

10 . A system according to claim 9 , wherein the processing circuitry is further configured to display an image based on the medical image data and to highlight the specified point or region on the displayed image.

11 . A system according to claim 9 , wherein the anatomical information comprises at least one anatomical landmark or anatomical region, and the specifying of the point or region on the medical image data comprises locating the at least one anatomical landmark or anatomical region in the medical image data.

12 . A system according to claim 9 , wherein the specifying of the anatomical information and/or the specifying of the point or region is dependent on a segmentation.

13 . A system according to claim 9 , wherein the specifying of anatomical information based on the concept corresponding to the specified node comprises obtaining at least one further concept or further term related to the specified node.

14 . A system according to claim 9 , wherein the specifying of the anatomical information comprises specifying anatomical information at different levels of generality.

15 . A system according to claim 9 , wherein the specifying of the anatomical information and/or specifying of the point or region is based on text information associated with the medical image data.

16 . A medical information processing method comprising:

receiving a text query and medical image data;

specifying a node of a medical ontology, knowledge graph or other knowledge base based on the text query;

using semantic functionality to map a concept corresponding to the specified node and/or a related concept to at least one anatomical region;

specifying a point or region on the medical image data corresponding to the at least one anatomical region;

determining at least one surface form for the concept;

performing a semantic search, using the concept, in one or more pre-existing clinical notes or pre-existing health records for a patient that existed before the medical image data is received, the semantic search comprising:

searching for the at least one surface form in text data of the one or more pre-existing clinical notes or pre-existing health records;

identifying a portion of the one or more pre-existing clinical notes or pre-existing health records including the at least one surface form; and

displaying the portion with the at least one surface form highlighted; and

using a concept distance measure in a concept space of the medical ontology, knowledge graph or other knowledge base to rank and/or filter concepts and/or terms, wherein

the concept space has nodes, each node representative of a concept,

the nodes are connected by edges, each edge of the edges representative of a relationship between two nodes that the edge connects, and

a result of the concept distance measure is expressed as a number of the edges.