Method and system for identifying anomalies in medical images especially those including one of a pair of symmetric body parts

• US 9,779,504 B1
• Filed: 04/28/2016
• Issued: 10/03/2017
• Est. Priority Date: 12/14/2011
• Status: Active Grant

• Current Assignee: Atti International Services Company, Inc.
• Original Assignee: Atti International Services Company, Inc.
• Inventors: Biagiotti, William, Levi, Eli, Levi, David
• Primary Examiner(s): Li, Ruiping
• Application Number:
  US15/140,746
• Time in Patent Office: 523 Days
• Field of Search: 382128
• CPC Class Codes: G06F 16/5838 using colour
• CPC Class Codes: G06F 18/22 Matching criteria, e.g. proximity measures
• CPC Class Codes: G06F 18/2433 Single-class perspective, e.g. one-against-all classification; Novelty detection; Outlier detection
• CPC Class Codes: G06F 3/04842 Selection of displayed objects or displayed text elements G06F3/0482 takes precedence
• CPC Class Codes: G06T 11/60 Editing figures and text; Combining figures or text
• CPC Class Codes: G06T 2207/10072 Tomographic images
• CPC Class Codes: G06T 2207/20104 Interactive definition of region of interest [ROI]
• CPC Class Codes: G06T 2207/30016 Brain
• CPC Class Codes: G06T 7/0014 using an image reference approach
• CPC Class Codes: G06T 7/13 Edge detection
• CPC Class Codes: G06T 7/68 of symmetry
• CPC Class Codes: G06T 7/74 involving reference images or patches
• CPC Class Codes: G06T 7/90 Determination of colour characteristics
• CPC Class Codes: G06V 10/761 Proximity, similarity or dissimilarity measures
• CPC Class Codes: G06V 10/764 using classification, e.g. of video objects
• CPC Class Codes: G06V 2201/03 Recognition of patterns in medical or anatomical images
• CPC Class Codes: G16B 50/00 ICT programming tools or database systems specially adapted for bioinformatics
• CPC Class Codes: G16H 30/40 for processing medical images, e.g. editing
• CPC Class Codes: G16H 50/20 for computer-aided diagnosis, e.g. based on medical expert systems
• CPC Class Codes: G16H 70/20 relating to practices or guidelines

Abstract

Method for locating anomalous features in medical images in which medical images are supplied by an external source. A sequence of specific measurements is executed on the supplied data to obtain image-related metrics. The metrics are compared to corresponding values in a database resulting in an anomalous/not anomalous determination. Anomalous determinations are presented to the test operator for analysis along with supplemental historical data. In application to all types of medical imagery, potential anomalies are located resulting in an efficient and accurate diagnosis. For paired body parts having symmetrical properties, as a pre-screening stage, a line of symmetry between the body parts is determined and symmetric regions of the two body parts on opposite sides of the determined line of symmetry are analyzed for differences. The metric comparison using the database is performed between both symmetric regions only when differences between the symmetric regions of the subject medical image are present.

Claims

1. A method for analyzing a subject medical image containing at least a first body part having symmetrical properties with a second body part, comprising:

determining, using a processor, a line of symmetry between the first body part in the subject medical image and the second body part;

analyzing, using the processor, regions of the first and second body parts on opposite sides of the determined line of symmetry for differences; and

only when one or more differences between one of the regions of the first body part in the subject medical image relative to a symmetric one of the regions of the second body part are found,

retrieving, from a database, at least one set of a plurality of specific metrics from a plurality of sets of metrics stored in the database, each of the sets of metrics stored in the database being derived from an anomaly or a non-anomalous region in an associated one of a plurality of historical medical images such that at least one set is derived from an anomaly of an associated historical medical image and at least one set is derived from a non-anomalous region of an associated historical medical image, each of the metrics in each set of metrics being a different measurement of the associated historical medical image;

determining, using the processor, a plurality of specific metrics from each of the regions of the first body part in the subject medical image, each of the metrics being a different measurement of the region of the first body part in the subject medical image;

comparing, using the processor, each of the determined metrics from each of the regions of the first body part in the subject medical image to a respective one of the metrics of the retrieved at least one set of the plurality of metrics from the database to determine a degree of similarity therebetween;

creating, using the processor, a contour map based on the determined degree of similarity between the plurality of determined metrics from each of the regions of the first body part in the subject medical image and the retrieved at least one set of the plurality of metrics from the database; and

displaying the created contour map on the subject medical image on a display.

2. The method of claim 1, wherein the subject medial image includes the first and second body parts and wherein the step of analyzing regions of the first and second body parts on opposite sides of the determined line of symmetry for differences comprises:

determining, using the processor, a plurality of specific metrics from each of the regions of the first and second body parts in the subject medical image, each of the metrics being a different measurement of the region of the first or second body part in the subject medical image; and

comparing, using the processor, each of the determined metrics from one of the regions of the first body part in the subject medical image to a respective one of the determined metrics from the region of the second body of the subject medical image to determine a degree of similarity therebetween.

3. The method of claim 1, further comprising storing the specific metrics derived from anomalies and non-anomalous regions in historical medical images in the database arranged by body categories, components of the category, and secondary components of the category.

4. The method of claim 1, further comprising:

storing additional information related to the historical medical images in the database that include anomalies, including for each such medical image, patient age, patient sex, area of anomaly, shape of anomaly, diagnostic tests performed and their reports, diagnosis/prognosis, treatment, outcome, time to resolution, and actual medical image;

forming the database from the historical medical images; and

when forming the database, reducing an overall size of the historical medical images using the processor by delineating an outer boundary of any anomaly in each historical medical image, then enlarging the delineated outer boundary to provide an enlarged contour of the diagnosed anomaly, and using the enlarged contour of the diagnosed anomaly in metric calculations to be saved in the database.

5. The method of claim 1, further comprising organizing the database to include a cross reference indication when one anomaly in a historical medical image in the database occurs simultaneously with one or more other anomalies in one or more other historical medical images in the database, the cross reference indication inferring a relationship between the anomalies in the two or more historical medical images when any of the linked anomalies is retrieved.

6. The method of claim 1, wherein the specific metrics consists of one or more of the following:

an intensity range of a region of the anomaly,

an intensity range of the non-anomalous region,

a ratio between the intensity range of the anomaly and the intensity range of the non-anomalous region,

a variance of the anomaly region,

a variance of the non-anomalous region, and

a ratio between the between the variance of the anomaly region and the variance of the non-anomalous region.

7. The method of claim 1, wherein the subject medical image does not have manual delineations, further comprising:

retrieving, using the processor, at least one analogous entry from the historical medical image database related to the subject medical image; and

configuring, using the processor, a subset of the subject medical image from which the plurality of specific metrics is determined as a virtual container of size and shape derived from the retrieved at least one analogous entry from the historical medical image database such that the specific metrics are determined based only on that portion of the subject medical image in the virtual container.

8. The method of claim 7, further comprising moving, using the processor, the virtual container incrementally across the subject medical image to different positions, the plurality of specific metrics being determined at each of a plurality of positions to which the virtual container is moved.

9. The method of claim 8, further comprising:

moving, using the processor, the virtual container incrementally across a reduced resolution of the original subject medical image, when resolution of the original image exceeds a programmable threshold; and

comparing, using the processor, the metrics from the subject medical image to the metrics stored for each of a plurality of analogous entries retrieved from the historical medical image database until a successful condition is determined by using a pass/fail threshold metric stored in the database as the criteria, the successful condition being present when one of the analogous entries retrieved from the historical medical image database passes the threshold metric.

10. The method of claim 8, further comprising:

comparing, using the processor, the metrics from each of the regions of the first body part in the subject medical image to the metrics stored for each of a plurality of analogous entries retrieved from the historical medical image database until a successful condition is determined by using a pass/fail threshold metric stored in the database as the criteria, the successful condition being present when one of the analogous entries retrieved from the historical medical image database passes the threshold metric;

comparing, using the processor, all analogous entries in the historical medical image database to the first body part of the subject medical image; and

when all comparisons have not passed the criteria of the pass/fail threshold metrics stored in the historical medical image database, visually presenting, using the processor, the subject medical image on a user interface with status information showing that no comparative match was found in the historical medical image database.

11. The method of claim 1, further comprising:

enabling, using the processor, a test operator to confirm an anomalous object in the first body part of the subject medical image; and

after such confirmation, modifying, using the processor, the database to include the plurality of determined metrics from the subject medical image as a new set of metrics associated with the subject medical image and confirmation of the anomalous object as an anomaly.

12. A system for analyzing a subject medical image containing at least a first body part having symmetrical properties with a second body part, comprising:

a database containing sets of a plurality of specific metrics, each of the sets of metrics being derived from an anomaly or a non-anomalous region in an associated one of a plurality of historical medical images such that at least one set is derived from an anomaly of an associated historical medical image and at least one set is derived from a non-anomalous region of an associated historical medical image, each of the metrics in each set of metrics being a different measurement of the associated historical medical image;

at least one processor coupled to said database and that determines a line of symmetry between the first body part in the subject medical image and the second body part, analyzes regions of the first and second body parts on opposite sides of the determined line of symmetry for differences, and only when one or more differences between one of the regions of the first body part of the subject medical image relative to a symmetric one of the regions of the second body part are found, determines a plurality of specific metrics from each of the regions of the first body part of the subject medical image, each of the determined metrics being a different measurement of the subject medical image,

said at least one processor being configured to compare each of the determined metrics from each of the regions of the first body part of the subject medical image to a respective one of the metrics of the retrieved at least one set of the plurality of metrics from said database to determine a degree of similarity therebetween,

said at least one processor being further configured to create a contour map based on the determined degree of similarity between the plurality of determined metrics from each of the regions of the first body part of the subject medical image and the retrieved at least one set of the plurality of metrics from said database; and

a display device coupled to said at least one processor, said display device being directed by said at least one processor to display the created contour map on the subject medical image.

13. The system of claim 12, further comprising at least one user interface coupled to said at least one processor to enable a user to delineate regions of the first body part of the subject medical image, said at least one processor being further configured to determine the plurality of specific metrics for each delineated region of the subject medical image and then compare the plurality of metrics from each delineated region of the subject medical image to the sets of the plurality of metrics from said database to determine a degree of similarity therebetween.

14. A system for analyzing a subject medical image containing at least a first body part having symmetrical properties with a second body part, comprising:

a database containing sets of a plurality of specific metrics derived from both anomalies and non-anomalous regions in a plurality of historical medical images, wherein the anomalies and non-anomalous regions are considered to be “

good”

or normal;

a second database containing sets of a plurality of specific metrics derived from both anomalies and non-anomalous regions in a plurality of historical medical images, wherein the anomalies and non-anomalous regions are considered to be “

bad”

or abnormal;

a processor coupled to said first and second databases and configured to;

determine a line of symmetry between the first body part in the subject medical image and the second body part,

analyze regions of the first and second body parts on opposite sides of the determined line of symmetry for differences, and

only when one or more differences between one of the regions of the first body part of the subject medical image relative to a symmetric one of the regions of the second body part are found,

determine the plurality of specific metrics from each of the regions of the first body part of the subject medical image obtained by said medical imaging device, each of the determined metrics being a different measurement of the subject medical image,

compare the plurality of metrics from each of the regions of the first body part of the subject medical image to the sets of the plurality of metrics from said first database to determine a degree of similarity between the plurality of determined metrics from each of the regions of the first body part of the subject medical image and each set of the plurality of metrics from said first database, and

thereafter and only when the degree of similarity between the plurality of determined metrics from at least one of the regions of the first body part of the subject medical image and all of the sets of the plurality of metrics from said first database is below a threshold, compare the plurality of metrics from the at least one of the regions of the first body part of the subject medical image to the sets of the plurality of metrics from said second database to determine a degree of similarity between the plurality of determined metrics from the at least one of the regions of the first body part of the subject medical image and each set of the plurality of metrics from said second database, and

create a contour map based on the determined degree of similarity between the plurality of metrics from the at least one of the regions of the first body part of the subject medical image and the sets of the plurality of metrics from said second database; and

a display device coupled to said processor, said display device being directed by said processor to display the created contour map on the subject medical image.