Apparatus and corresponding method for measuring lumbar spine

Patent metadata

Field	Value
Publication number	US-12582372-B2
Application number	US-202318296941-A
Country	US
Kind code	B2
Filing date	Apr 6, 2023
Priority date	Apr 7, 2022
Publication date	Mar 24, 2026
Grant date	Mar 24, 2026

Abstract

Described herein is a technique (e.g., apparatus and method) to improve consistency of lumbar lordosis (LL) and disc height index (DHI) measurements. In at least one embodiment, one or more processors are to measure lumber spine based, at least in part, on cross-sectional area and optical center of mass of segmented intervertebral discs (IVDs).

First claim

We claim:

1 . A method comprising:
receiving images of segmented intervertebral discs (IVDs) from image segmentation circuitry comprising neural network architecture configured to apply filters for feature extraction and classification to automatically segment intervertebral disc regions from medical imaging data; identifying a first disc and a second disc from the images of the segmented IVDs, wherein the second disc is a caudal neighbor of the first disc; computing a centroid and orientation for the first disc and the second disc from the images of segmented IVDs; computing a first pixel width of the first disc; computing a first disc height of the first disc based on pixel area and the first width by calculating the first disc height as pixel area divided by the first width; computing a second width of the second disc; computing a second disc height of the second disc based on the pixel area and the second width by calculating the second disc height as the pixel area divided by the second width, wherein computing the first disc height and the second disc height comprises rotating the first disc and the second disc such that the first disc and the second disc are normal to a reading frame, and creating a first bounding box around the first disc; and a second bounding box around the second disc in response to rotating the first disc and the second disc, and wherein the first width is a width of the first bounding box and the second width is a width of the second bounding box; and generating, by report generation circuitry, a clinical diagnostic report containing the first disc height, second disc height, and comparison to normative data for display to healthcare providers.

2 . The method of claim 1 , wherein computing the first disc height and the second disc height comprises:
determining a first top edge of the first disc prior to rotating the first disc such that the first disc is an original orientation, wherein determining the first top edge comprises applying a half of the first disc height along a first short axis of the original orientation of the first disc; and determining a second bottom edge of the second disc prior to rotating the second disc such that the second disc in its original orientation, wherein determining the second bottom edge comprises applying a half of the second disc height along a second short axis of the original orientation of the second disc.

3 . The method of claim 2 , comprising computing a first vertebral height, wherein computing the first vertebral height comprises determining an approximate vertebral height by calculating a distance between a first point located at the first top edge where it intersects with the first short axis of the first disc and a second point extending from the second bottom edge where it intersects with the second short axis of the second disc.

4 . The method of claim 3 , comprises computing a first disc height index by dividing the first disc height by the approximate vertebral height to determine a disc height index.

5 . The method of claim 1 , wherein computing the centroid and the orientation for the first disc and the second disc comprises:
determining a first centroid and a first orientation of the first disc, wherein the first disc is T12/L1 or L1/L2; determining a second centroid and a second orientation of the second disc, wherein the second disc is T5/S1; and computing an angle of lumbar lordosis based on the first centroid and the first orientation, and the second centroid and the second orientation.

6 . An apparatus comprising:
image segmentation circuitry comprising neural network architecture configured to apply filters for feature extraction and classification to generate segmented intervertebral discs (IVDs); one or more processors to measure lumbar spine based, at least in part, on cross-sectional area and center of mass of the segmented IVDs, wherein the image segmentation circuitry provides the segmented IVDs to the one or more processors, and wherein the one or more processors are to: compute a first width of a first disc; compute a first disc height of the first disc based on pixel area and the first width by calculating the first disc height as pixel area divided by the first width; compute a second width of a second disc; compute a second disc height of the second disc based on pixel area and the second width by calculating the second disc height as pixel area divided by the second width, wherein the one or more processors are to compute the first disc height and the second disc height by rotation of the first disc and the second disc such that the first disc and the second disc are normal to a reading frame, and by creation of a first bounding box around the first disc and a second bounding box around the second disc in response to rotating the first disc and the second disc; report generation circuitry configured to receive the computed measurements from the one or more processors and generate clinical diagnostic reports containing disc measurements and comparisons to normative data; and display circuitry configured to present the clinical diagnostic reports to healthcare providers.

7 . The apparatus of claim 6 , wherein the one or more processors applies an image center of the mass and an axis orientation of segmented IVDs to calculate Lumbar Lordosis.

8 . The apparatus of claim 6 , wherein the one or more processors are to:
receive images of the segmented IVDs; identify a first disc and a second disc from the images of the segmented IVDs, wherein the second disc is a caudal neighbor of the first disc; and compute a centroid and orientation for the first disc and the second disc from the images of the segmented IVDs.

9 . A system comprising:
image segmentation circuitry comprising neural network architecture configured to generate segmented intervertebral discs (IVDs) by applying filters for feature extraction and classification to medical imaging data; one or more memories; and one or more processors coupled to the one or more memories and the image segmentation circuitry, wherein the image segmentation circuitry receives the medical imaging data and generates the segmented IVDs using the neural network architecture, and wherein the one or more processors are to:
receive images of the segmented IVDs from the image segmentation circuitry;
identify a first disc and a second disc from the images of the segmented IVDs, wherein the second disc is a caudal neighbor of the first disc;
compute a centroid and an orientation for the first disc and the second disc from the images of segmented IVDs;
compute a first width of the first disc; compute a first disc height of the first disc based on pixel area and the first width by calculating the first disc height as pixel area divided by the first width;
compute a second width of the second disc;
compute a second disc height of the second disc based on pixel area and the second width by calculating the second disc height as pixel area divided by the second width, wherein the one or more processors are to compute the first disc height and the second disc height by rotation of the first disc and the second disc such that the first disc and the second disc are normal to a reading frame, and by creation of a first bounding box around the first disc and a second bounding box around the second disc in response to rotating the first disc and the second disc, and wherein the first width is a width of the first bounding box and the second width is a width of the second bounding box;
report generation circuitry configured to receive computed measurements from the one or more processors and generate clinical diagnostic reports with disc measurements and normative comparisons; a

Assignees

• Univ Oregon State

Inventors

• Giers Morgan B
• Ahrens Sonia

Classifications

• G06T7/0012

  Biomedical image inspection · CPC title

• G06T2207/10088

  Magnetic resonance imaging [MRI] · CPC title

• G06T2207/10116

  X-ray image · CPC title

• G06T2207/30012

  Spine; Backbone · CPC title

• G06T7/11

  Region-based segmentation · CPC title