Ultrasound clinical feature detection and associated devices, systems, and methods

Patent metadata

Field	Value
Publication number	US-2020043602-A1
Application number	US-201816498667-A
Country	US
Kind code	A1
Filing date	Mar 28, 2018
Priority date	Mar 28, 2017
Publication date	Feb 6, 2020
Grant date	—

Abstract

Ultrasound image devices, systems, and methods are provided. A clinical condition detection system, comprising a communication device in communication with an ultrasound imaging device and configured to receive a sequence of ultrasound image frames representative of a subject body across a time period; and a processor in communication with the communication device and configured to classify the sequence of ultrasound image frames into a first set of clinical characteristics by applying a first predictive network to the sequence of ultrasound image frames to produce a set of classification vectors representing the first set of clinical characteristics; and identify a clinical condition of the subject body by applying a second predictive network to the set of classification vectors.

First claim

1 . A clinical condition detection system, comprising:
a communication device in communication with an ultrasound imaging device and configured to receive a sequence of ultrasound image frames representative of a subject body across a time period; and a processor in communication with the communication device and configured to:
classify spatial domain information in the sequence of ultrasound image frames by applying a first deep learning network to each ultrasound image frame of the sequence of ultrasound image frames to generate a classification vector including a probability for each clinical characteristic of a first set of clinical characteristics;
classify time domain information in the classification vectors corresponding to the sequence of ultrasound image frames across the time period by applying a second deep learning network to generate a probability for each clinical characteristic of a second set of clinical characteristics; and
identify a clinical condition of the subject body based on the probabilities for the second set of clinical characteristics.

2 . (canceled)

3 . The system of claim 1 , wherein the first deep learning network is a convolutional neural network.

4 . The system of claim 1 , wherein the processor is further configured to identify the clinical condition by:
selecting a highest probability from the probabilities for the second set of clinical characteristics.

5 . The system of claim 1 , wherein the first set of clinical characteristics is identical to the second set of clinical characteristics.

6 . The system of claim 1 , wherein the first set of clinical characteristics is associated with different types of clinical features, and wherein the second set of clinical characteristics is associated with different degrees of severity for at least one of the different types of clinical features.

7 . The system of claim 1 , wherein the first set of clinical characteristics is associated with a first categorization of at least one of different types of clinical features or different degrees of severity, and wherein the second set of clinical characteristics is associated with a second categorization of the at least one of the different types of clinical features or the different degrees of severity.

8 . The system of claim 1 , wherein the subject body includes at least a portion of a lung.

9 . The system of claim 8 , wherein the clinical condition includes features associated with at least one of a normal lung, a B-line artifact, consolidation, a bronchogram, or a pleural effusion.

10 . The system of claim 8 , wherein the clinical condition includes features associated with a degree of severity of at least one of a normal lung, a B-line artifact, consolidation, a bronchogram, or a pleural effusion.

11 . The system of claim 1 , wherein at least the first predictive network or the second predictive network is trained by:
providing a plurality of scan-formatted ultrasound images representative of test subject bodies including the clinical condition captured by different ultrasound imaging devices; converting the plurality of scan-formatted ultrasound images into a plurality of pre-scan-formatted ultrasound images based on at least one of a common dimension or a common format independent of the different ultrasound imaging devices; and assigning a score to each ultrasound image of the plurality of pre-scan-formatted ultrasound images with respect to the clinical condition.

12 . The system of claim 11 , wherein the different ultrasound imaging devices include at least one of a linear ultrasound transducer device, curvilinear ultrasound transducer device, or a phased-array ultrasound transducer device.

13 . The system of claim 1 , further comprising:
a display in communication with the processor and configured to display an indication of the clinical condition.

14 . A method for clinical condition detection, comprising:
receiving, from an ultrasound imaging device, a sequence of ultrasound image frames representative of a subject body across a time period; classifying spatial domain information in the sequence of ultrasound image frames by applying a first deep learning network to each ultrasound image frame of the sequence of ultrasound image frames to generate a classification vector including a probability for each clinical characteristic of a first set of clinical characteristics; classifying time domain information in the classification vectors corresponding to the sequence of ultrasound image frames across the time period by applying a second deep learning network to generate a probability for each clinical characteristic of a second set of clinical characteristics; and identifying a clinical condition of the subject body based on the probabilities for the second set of clinical characteristics.

15 . (canceled)

16 . The method of claim 13 , wherein the first predictive network is a convolutional neural network.

17 . The method of claim 14 , wherein the identifying includes:
selecting a highest probability from the probabilities for the second set of clinical characteristics.

18 . The method of claim 14 , wherein the subject body includes at least a portion of a lung, and wherein the clinical condition includes features associated with at least one of a normal lung, a B-line artifact, consolidation, a bronchogram, or a pleural effusion.

19 . The method of claim 14 , wherein the subject body includes at least a portion of a lung, and wherein the clinical condition includes features associated with a degree of severity of at least one of a normal lung, a B-line artifact, consolidation, a bronchogram, or a pleural effusion.

20 . The method of claim 14 , further comprising:
displaying, by a display, an indication of the clinical condition.

Assignees

• Koninklijke Philips Nv

Inventors

• Kim Seungsoo
• Raju Balasundar Iyyavu

Classifications

• G16H50/20

  for computer-aided diagnosis, e.g. based on medical expert systems · CPC title

• G16H30/40 Primary

  for processing medical images, e.g. editing · CPC title

• G16H50/70

  for mining of medical data, e.g. analysing previous cases of other patients · CPC title

• G06N3/02

  Neural networks · CPC title

• A61B8/483

  involving the acquisition of a 3D volume of data · CPC title