Search

EP-4312783-B1 - PROGRESSIVE SCANS WITH MULTIPLE PULSED X-RAY SOURCE-IN-MOTION TOMOSYNTHESIS IMAGING SYSTEM

EP4312783B1EP 4312783 B1EP4312783 B1EP 4312783B1EP-4312783-B1

Inventors

  • MAOLINBAY, MANAT
  • KU, CHWEN-YUAN
  • YANG, LINBO
  • LIU, JIANQIANG

Dates

Publication Date
20260506
Application Date
20220302

Claims (15)

  1. A method of progressive scans with multiple pulsed X-ray source-in-motion tomosynthesis imaging system, the method comprising: placing an object in a predetermined position; controlling a multiple pulsed X-ray source-in-motion tomosynthesis imaging system (1); taking a first set of data from different X-ray source (2) at different angles using said tomosynthesis imaging system (1) and performing an image reconstruction; applying artificial intelligence with machine learning to perform diagnostics and repeating one or more scans to reach a predetermined image reconstruction quality; and constructing a 3D tomosynthesis volume therefrom.
  2. The method of claim 1, comprising selecting an imaging task and loading a predetermined protocol to the imaging system (1).
  3. The method of claim 1, comprising acquiring one or more projection images from one or more selected sources.
  4. The method of claim 1, comprising accumulating the one or more projection images and reconstructing the 3D tomosynthesis volume therefrom.
  5. The method of claim 1, comprising applying machine learning to the reconstructed 3D tomosynthesis volume to determine the image reconstruction quality.
  6. The method of claim 1, comprising applying machine learning to search for lung nodules in the 3D tomosynthesis volume.
  7. The method of claim 1, comprising: applying artificial intelligence with machine learning to search for lung nodules in the 3D tomosynthesis volume; and identifying nodule characteristics including size, shape, numbers, and locations for the nodules and generating a report.
  8. The method of claim 5, comprising acquiring additional projection images and applying machine learning to the reconstructed 3D tomosynthesis volume until the predetermined image reconstruction quality threshold is met.
  9. The method of claim 6, comprising identifying nodule characteristics including size, shape, numbers, and locations for the nodules and generating a report.
  10. The method of claim 1, further comprising: selecting an imaging task and loading a predetermined protocol to the imaging system (1); the performing an image reconstruction being done; by acquiring one or more set of projection images from one or more selected sources of the imaging system; accumulating the one or more sets of projection images; acquiring additional set of projection images and applying artificial intelligence with machine learning to the reconstructed 3D tomosynthesis volume until the predetermined image reconstruction quality threshold is met; and applying artificial intelligence with machine learning to search for lung nodules in the 3D tomosynthesis volume.
  11. An X-ray imaging system, comprising: a multiple pulsed X-ray source-in-motion tomosynthesis imaging system (1) to receive an object in a predetermined position; a processor coupled to the imaging system to run computer code to: control said multiple pulsed X-ray source-in-motion tomosynthesis imaging system (1); obtain a first set of data from different X-ray source (2) at different angles and performing an image reconstruction; apply artificial intelligence with machine learning to perform diagnostics and repeat one or more scans to reach a predetermined image reconstruction quality; and constructing a 3D tomosynthesis volume therefrom.
  12. The system of claim 11, wherein the processor selects an imaging task and loading a predetermined protocol to the imaging system.
  13. The system of claim 11, wherein the processor controls the imaging system to acquire one or more projection images from one or more selected sources.
  14. The system of claim 11, wherein the processor accumulates the one or more projection images to reconstruct the 3D tomosynthesis volume.
  15. The system of claim 11, wherein the processor runs machine learning code to the reconstructed 3D tomosynthesis volume to determine the image reconstruction quality.

Description

Field of the Invention The present invention relates generally to X-ray diagnostic imaging and, more particularly, to a method and apparatus of multiple pulsed X-ray source-in-motion tomosynthesis imaging using a series of partial-scans to sample an imaging volume and using Al to determine if information is enough from image reconstruction. Background Tomosynthesis, also digital tomosynthesis (DTS), is a method for performing high resolution limited-angle tomography at radiation dose levels comparable with projection radiography. It has been studied for a variety of clinical applications, including vascular imaging, dental imaging, orthopedic imaging, mammographic imaging, musculoskeletal imaging, and lung imaging. DTS dose level is far less than that of a CT, DTS is also much faster than that of CT, and DTS itself costs far less. In order to further reduce X-ray dose on a patient and even faster perform X-ray scans, the current invention is to a method and system of imaging acquisition from distributed wide-angle sparse partial scans. There are some prior arts to perform progressive image reconstructions using different resolutions. However, there are several disadvantages associated with the prior arts. The first disadvantage is that there are only resolution changes at the same location; the second disadvantage is that there is no view angle location information. Second disadvantage; the third disadvantage is that there is no artificial intelligence (Al) involved. There are also other prior arts regarding progressive scan for CT. The first disadvantage is that the progressive scan is incremental position and angle coverage is quite small because it is for CT only. It includes an imaging technique whereby a subject is incrementally translated through a number of discrete scan positions to acquire CT data from a region of the subject. In this regard, the subject is not translated to the next scan position until valid or acceptable data is acquired for a current scan position. The second disadvantage is that slower with one X-ray source. Most CT apparatus only has an X-ray source. It has to relatively rotate quite some angle to get larger coverage. Sampling imaging volume starts from a small angle and then increases incrementally. So, it would be slow. The third disadvantage in the prior art is that there is usually no Al involved in the prior art regarding incremental progressive scan. US 2020/0211240 A1 discloses methods, apparatus and systems for deep learning-based image reconstruction. At least one computer-readable storage medium includes instructions that, when executed, cause at least one processor to at least: obtain a plurality of two-dimensional tomosynthesis projection images of an organ by rotating an x-ray emitter to a plurality of orientations relative to the organ and emitting a first level of x-ray energization from the emitter for each projection image of the plurality of 2D tomosynthesis projection images; reconstruct a 3D volume of the organ from the plurality of 2D tomosynthesis projection images; obtain an x-ray image of the organ with a second level of x-ray energization; generate a synthetic 2D image generation algorithm from the reconstructed 3D volume based on a similarity metric between the synthetic 2D image and the x-ray image; and deploy a model instantiating the synthetic 2D image generation algorithm. US 2020/0118308 A1 discloses a medical image processing apparatus including processing circuitry which acquires information on missing data based on first projection data obtained by scanning a subject. The processing circuitry generates second projection data by interpolating missing data in the first projection data based on the information on missing data. The processing circuitry generates a first reconstructed image by reconstructing the second projection data. The processing circuitry generates third projection data by performing forward projection on the first reconstructed image. The processing circuitry generates fourth projection data by updating the second projection data based on the third projection data. The processing circuitry generates a second reconstructed image based on the fourth projection data. US 2015/0238159 A1 discloses a method including, in a bi-plane interventional imaging system, moving a first C-arm supporting a first X-ray source and a first X-ray detector about first and second axes while obtaining a plurality of first X-ray attenuation data sets relating to a subject of interest; moving a second C-arm, positioned crosswise with respect to the first C-arm and supporting a second X-ray source and a second X-ray detector, about the first axis while obtaining a plurality of second X-ray attenuation data sets relating to the subject of interest; and synchronizing the movement of the first and second C-arms to avoid collision therebetween. Summary In one aspect, a method of progressive scans with multiple pulsed X-ray source-in-motion tomosynthesis im