Method for calculating local narrowing in a pipe

Abstract

The method for calculating local narrowing from a three-dimensional image of a pipe comprises the following steps: - determining (step E3) a minimum diameter profile of the pipe section along the length of the pipe; - defining (step E4) at least one sliding window for inspecting the pipe, for each position of said window on the pipe, the minimum diameter of the pipe at the longitudinal ends of said window being known; and - for each position of the at least one window, calculating (step E5) the maximum loss of interpolation of the minimum diameter of the pipe within the window relative to the diameters at the longitudinal ends of the window.

Description

Technical Field

[0001] This invention relates to the imaging and acquisition of visual information in and around channels such as blood vessels.

[0002] Specifically, this invention relates to the identification of narrowing of blood vessel diameter.

[0003] Generally, the present invention is applicable to any mechanical or physiological conduit for which diameter distribution checks can be performed, for example for medical or maintenance purposes. Existing technology

[0004] In the field of cardiovascular imaging, current tools enable the visualization of blood vessels, measurement of their diameter, and manual placement of markers at different reference points on images of these vessels to infer specific characteristics, such as the presence of a particular narrowing in the vessel diameter, sometimes a symptom of stenosis (e.g., narrowing of coronary arteries due to the presence of calcification on the artery). Helping to identify stenosis is important because doing so can ultimately reduce a patient's risk of heart attack.

[0005] However, these manual methods are unreliable, time-consuming, and non-reproducible, and lead to uncertainty about the measurements obtained.

[0006] Alternatively, tools developed using neural networks can be used to identify narrowing in blood vessels. However, this identification does not include any actual, quantifiable data that would allow for consensus on the severity of the identified narrowing. Summary of the Invention

[0007] Therefore, the object of the present invention is to overcome the above-mentioned disadvantages and to provide a reliable and reproducible method for identifying and quantifying narrowing in channels such as blood vessels.

[0008] This invention relates to a method for calculating local narrowing from a three-dimensional image of a pipe, the method comprising the following steps:

[0009] - Determine the minimum diameter distribution of pipe sections along the length of the pipe;

[0010] - Define at least one sliding window for inspecting the pipe, wherein for each position of the window on the pipe, the minimum diameter of the pipe at the longitudinal end of the window is known; and

[0011] - For each location of at least one window, calculate the maximum interpolation loss of the minimum diameter of the pipe within the window relative to the diameter at the longitudinal end of the window.

[0012] In this way, the solution enables the automatic detection and quantification of narrowing of the channel, such as vascular stenosis, via a method simple for the operator, without the need for manual reference selection. Clinical confidence is also enhanced.

[0013] Advantageously, the method further includes the step of recording the position of the window when the maximum loss calculated for the minimum diameter of the window is greater than a predetermined threshold, which is preferably 50% loss, or even more preferably 70% loss.

[0014] In a particular implementation, the method further includes the step of: for each window where a position is recorded and which at least partially overlaps with another window where a position is recorded, evaluating the average loss of interpolation of the minimum diameter of the pipe within the window relative to the diameter at the longitudinal end of the window.

[0015] Advantageously, the method includes the following steps: for each set of windows that have recorded positions and at least partially overlap, classifying the windows in the set of windows with the highest average loss based on the smallest evaluated diameter.

[0016] Optionally, the method includes a prior step of acquiring a three-dimensional image of the pipeline tree and a step of processing the three-dimensional image by segmenting the pipeline tree into at least one three-dimensional image of the pipeline.

[0017] Advantageously, the minimum diameter distribution is determined with a measurement step size between 0.1 mm and 1 mm, preferably between 0.2 mm and 0.5 mm, or even more preferably 0.3 mm, and the sliding window for inspection slides along the pipe in increments equal to the measurement step size of the minimum diameter distribution.

[0018] Advantageously, the longitudinal ends of the windows are spaced between 1.8 mm and 14.1 mm apart.

[0019] Advantageously, multiple windows are defined, each including a distance different from other windows between its longitudinal ends.

[0020] In a particular implementation, the recording step is performed only when the maximum loss of the minimum diameter of the pipe within the window relative to the minimum diameter of the longitudinal end of the window with the minimum minimum diameter is greater than a predetermined threshold (preferably 40% loss).

[0021] Advantageously, the method is implemented only for the portion of the pipe whose minimum diameter is greater than a value between 1 mm and 3 mm, preferably between 1.5 mm and 2 mm, or even more preferably 1.6 mm.

[0022] In a particular implementation, the conduit is a physiological conduit, such as a blood vessel, preferably a coronary artery.

[0023] The present invention also relates to a computer program comprising instructions that, when executed by a computer, cause the computer to perform the steps of a previously defined method.

[0024] The present invention also relates to a system comprising means for implementing the steps of the method as defined above.

[0025] The system includes, for example, a computer and a computer-readable data medium, on which computer programs as defined above are stored. Attached Figure Description

[0026] Other objects, features, and advantages of the invention will become apparent from the following description, which is given only by way of non-limiting example and with reference to the accompanying drawings, wherein:

[0027] [ Figure 1 ]

[0028] This is a schematic diagram illustrating the various steps implemented in the method according to the present invention;

[0029] [ Figure 2 ]

[0030] It is a three-dimensional pipeline tree; and

[0031] [ Figure 3 ]

[0032] It is a schematic two-dimensional view including the narrowing of the pipe. Detailed description of at least one implementation scheme

[0033] Figure 1 The illustration shows the method used according to Figure 2 and Figure 3 The steps in the method for calculating local narrowing of the three-dimensional image of pipe 1 shown.

[0034] For example, conduit 1 is a physiological conduit 1 of the human body, such as a blood vessel. Conduit 1 is, for example, a coronary artery, and may consist of branches of the aorta at the tip P of the coronary artery, excluding the bifurcation B.

[0035] The method can begin with step E1 to obtain a three-dimensional image of pipe 1 tree 3.

[0036] The tree 3 is in Figure 2 It is illustrated schematically. In the same... Figure 2 In the accompanying drawing, reference numeral 1 illustrates the path of the pipes referred to as tree 3 in this application. Clearly, tree 3 comprises multiple pipes 1, which... Figure 2 The base on the left is public.

[0037] Then step E2 can be performed to process the three-dimensional image by segmenting the pipe 1 tree 3 into at least one three-dimensional image of the pipe 1.

[0038] Figure 3 The diagram shows a schematic close-up view of the pipe section 1, including the narrowing section 5.

[0039] To implement this method, step E3 is performed to determine the minimum diameter distribution D of segments of conduit 1 along its length. Conduit 1 can be pre-selected by the operator from a conduit tree 3. Since conduit 1 is three-dimensional, it is practically necessary to determine the minimum diameter of each segment of conduit 1, for example, through image processing. The image processing performed includes, for example, tracing a centerline L at the center of conduit 1 and tracing the radius virtually extended from that centerline L to the wall of conduit 1 (actually the lumen if conduit 1 is a coronary artery).

[0040] The position of the diameter along the length of pipe 1 is represented by x, that is, the position of the section perpendicular to the longitudinal axis of pipe 1, defined by the centerline L.

[0041] The minimum diameter distribution D can optionally be determined with a measurement step size between 0.1 mm and 1 mm, preferably between 0.2 mm and 0.5 mm, or even more preferably 0.3 mm. Therefore, in the latter case, the minimum diameter of pipe 1 is known every 0.3 mm along the longitudinal axis of the pipe.

[0042] For the remainder of the method, condition C1 may optionally be applied such that the following steps are performed only on the portion of the minimum diameter D of the conduit 1 that is greater than a value between 1 mm and 3 mm, preferably between 1.5 mm and 2 mm, or even more preferably 1.6 mm. In fact, if the conduit 1 is a blood vessel, which does not have a constant minimum diameter D, and the tip P of the conduit 1 is very fine, for example, with a minimum diameter less than 1 mm, then the implementation of this method would be irrelevant for this type of minimum diameter D.

[0043] Subsequently, step E4 is performed to define at least one sliding window for inspecting pipe 1, wherein for each position of the window on pipe 1, the minimum diameter D of pipe 1 at the longitudinal end of the window is known.

[0044] In practice, x1 and x2 are the positions of the longitudinal ends of each window, with x2 typically corresponding to a position where the minimum diameter D is smaller than position x1, and is usually positioned closer to the tip P.

[0045] For example, the longitudinal ends of the windows are spaced between 1.8 mm and 14.1 mm apart (referred to as the window width) to allow for proper inspection of foreign objects on pipes 1 of similar size.

[0046] Each inspection window is designed to slide along pipe 1 in order to inspect for local narrowing 5 of the minimum diameter D of pipe 1, which is sometimes a symptom of narrowing.

[0047] For example, the window slides along pipe 1 with incremental displacements equivalent to the measurement of the minimum diameter distribution D.

[0048] Optionally, multiple windows are defined, each including a distance different from other windows between its longitudinal ends. In a particular embodiment, a large number of windows are defined to cover a wide range of window widths, for example from 1.8 mm to 14.1 mm.

[0049] For example, the window is defined as having a window with each possible width between 1.8mm and 14.1mm, where the increment between each different width is 0.3mm.

[0050] Then perform step E5: For each location of at least one window, calculate the maximum interpolation loss of the minimum diameter D of the pipe 1 within the window relative to the diameter at the longitudinal end of the window.

[0051] Therefore, for each x between x1 and x2, the minimum diameter loss D of the interpolated diameter of pipe 1 relative to the longitudinal end of the window is calculated. Then, the maximum value of these losses is calculated according to the global formula:

[0052]

[0053] in It is the interpolation of the minimum diameter D at the longitudinal end of the window at position x, and w(x) is a linear function with a value of 1 at x1 and a value of 0 at x2.

[0054] Therefore, this calculation provides particularly relevant data regarding the presence of narrowing 5 in pipe 1, where the higher the maximum loss, the more pipe 1 narrows in its cross-sectional direction, such as the symptoms of severe stenosis in the case of pipe 1 being a coronary artery.

[0055] For example, according to condition C2, when the maximum loss calculated for the minimum diameter of the window is greater than a predetermined threshold (preferably 50% loss, or even more preferably 70% loss), step E6 can be performed to record the position of the window. These percentages correspond to a narrowing 5 that may potentially indicate the presence of severe narrowing, or in any case, a narrowing 5 whose cause needs to be investigated.

[0056] The recording step E6 includes, for example, writing the window's position, its width, and the maximum loss of the minimum diameter D calculated for the window into memory.

[0057] Optionally, according to condition C3, the recording step E6 is performed only if the maximum loss of the minimum diameter D of the pipe 1 within the window relative to the minimum diameter D of the longitudinal end of the window with the smallest minimum diameter D is greater than a predetermined threshold (preferably 40% loss). This dual condition C2 and C3 avoids confusion between the undesirable narrowing 5 and the presence of the bifurcation B from the pipe 1.

[0058] Then, for each window whose position was recorded in the preceding step E6 and which at least partially overlaps with another window whose position was recorded, step E7 can be performed to evaluate the average loss of the interpolation of the minimum diameter D of the pipe 1 within the window relative to the minimum diameter D at the longitudinal end of the window. The average of these losses is evaluated according to a global formula:

[0059]

[0060] Finally, step E8 can be performed to classify the windows with the highest evaluated average minimum diameter loss D in each set of windows that have recorded positions and at least partially overlap.

[0061] These last two steps, E7 and E8, allow the same narrowing 5 to be considered multiple times in different window scenarios, and allow filtering to be performed to evaluate the window that constitutes the best reference for identifying the narrowing 5 (e.g., the narrowing). The best checking reference here is equivalent to the manual reference in the prior art, but is now obtained automatically and in a reproducible, optimized, and reliable manner.

[0062] Therefore, the window is categorized, for example, stored in memory, where the window's position, width, and the maximum and average losses of the minimum diameter D are calculated and evaluated separately.

[0063] The present invention also relates to a system comprising means for implementing the steps of the illustrated method, the means including, for example, a computer.

Claims

1. A method for calculating local narrowing from a three-dimensional image of a pipe (1), characterized in that, The method includes the following steps: - Determine (step E3) the minimum diameter distribution (D) of the section of the pipe (1) along the length of the pipe (1); - A limitation (step E4) is made for inspecting at least one sliding window of the pipe (1), wherein for each position of the window on the pipe (1), the minimum diameter (D) of the pipe (1) at the longitudinal end of the window is known; and - For each location of the at least one window, calculate (step E5) the maximum interpolation loss of the minimum diameter of the pipe (1) within the window relative to the diameter (D) at the longitudinal end of the window.

2. The method according to claim 1, further comprising step (E6): when the maximum loss of the minimum diameter (D) calculated for the window is greater than a predetermined threshold, recording the position of the window, wherein the predetermined threshold is preferably 50% loss, or even more preferably 70% loss.

3. The method according to claim 2, further comprising the step (E7): for each window in which the position is recorded and which at least partially overlaps with another window in which the position is recorded, evaluating the average loss of the interpolation of the minimum diameter (D) of the pipe (1) within the window relative to the diameter (D) at the longitudinal end of the window.

4. The method of claim 3, the method comprising the step (E8): for each set of windows that have recorded the positions and at least partially overlap, classifying the windows in the set of windows with the highest average loss based on the smallest evaluated diameter (D).

5. The method according to any one of claims 1 to 4, the method comprising a prior step (E1) of acquiring a three-dimensional image of a pipeline (1) tree (3) and a step (E2) of processing the three-dimensional image by segmenting the pipeline (1) tree (3) into at least one three-dimensional image of a pipeline (1).

6. The method according to any one of claims 1 to 5, wherein the minimum diameter distribution (D) is determined with a measurement step size between 0.1 mm and 1 mm, preferably between 0.2 mm and 0.5 mm, or even more preferably 0.3 mm, and the sliding window being inspected slides along the pipe (1) in increments equal to the measurement step size of the minimum diameter distribution (D).

7. The method according to any one of claims 1 to 6, wherein the longitudinal ends of the window are spaced apart by a distance between 1.8 mm and 14.1 mm.

8. The method of claim 7, wherein a plurality of windows are defined, each window comprising a distance different from the other windows between its longitudinal ends.

9. The method according to any one of claims 1 to 8, wherein the recording step (E6) is performed only when the maximum loss of the minimum diameter (D) of the pipe (1) within the window relative to the minimum diameter of the longitudinal end of the window having the minimum minimum diameter is greater than a predetermined threshold, the predetermined threshold being 40% loss.

10. The method according to any one of claims 1 to 9 is implemented only for the portion of the minimum diameter (D) of the pipe (1) that is greater than a value between 1 mm and 3 mm, preferably between 1.5 mm and 2 mm, or even more preferably 1.6 mm.

11. The method according to any one of claims 1 to 10, wherein the conduit (1) is a physiological conduit (1), such as a blood vessel, preferably a coronary artery.

12. A system comprising means for implementing the steps of the method according to any one of claims 1 to 11.

Images