Search

EP-4736773-A1 - SYSTEM AND METHOD FOR DETERMINING INNER AND OUTER VESSEL CURVE

EP4736773A1EP 4736773 A1EP4736773 A1EP 4736773A1EP-4736773-A1

Abstract

A method and system for preprocedural planning is disclosed. The method includes receiving a 3D medical image (205) comprising 3D segmentation (210) of a vessel (220). The 3D segmentation (210) comprises a centerline (214A) of the vessel (220) and contours (218 A-N) of an inner wall of the vessel (220). The centerline (214A) and contours (218 A-N) of the inner wall of the vessel (220) are represented as centerline points (222 A-N) and contour points (224 A-N) in a 3D space. The method includes identifying an inner (212) and an outer curve (216) of the inner wall of the vessel (220) and outputting the inner (212) and the outer curve (216) of inner wall of vessel (220) on a display. Additionally, when an aneurysm is detected, the method (100) includes recomputing the centerline (214B), modified contours in the region of aneurysm and inner (212) and outer (216) curves.

Inventors

  • THAYYULLATHIL,, Hemanth
  • RABOTNIKOV, Mark
  • SETTY, Vasavi

Assignees

  • Koninklijke Philips N.V.

Dates

Publication Date
20260506
Application Date
20241029

Claims (15)

  1. A computer-implemented method (100) in particular for preprocedural planning and/or computer assisted diagnosis, the method (100) comprising the steps of: - receiving (102), a 3D medical image (205) comprising a 3D segmentation (210) of a vessel (220), wherein the 3D segmentation (210) of the vessel (220) comprises a centerline (214A) of the vessel (220) and a plurality of contours (218 A-N) of an inner wall of the vessel (220), wherein the centerline (214A) of the vessel (220) and the plurality of contours (218 A-N) of the inner wall of the vessel (220) are represented as centerline points (222 A-N) and contour points (224 A-N) in a 3D space; - identifying (106), an inner curve (212) and an outer curve (216) of the inner wall of the vessel (220) in the 3D segmentation (210) of the vessel (220) using a plurality of direction vectors (228 A-N) orthogonal from the plurality of centerline points (222 A-N) to the plurality of corresponding contour points (224 A-N); and - outputting (108), the inner curve (212) and the outer curve (216) of the inner wall of the vessel (220) on a display together with the 3D medical image (205) and the 3D segmentation (210) of the vessel (220).
  2. The computer-implemented method (100) according to claim 1, wherein using the plurality of direction vectors (228 A-N) orthogonal from the plurality of centerline points (222 A-N) to the plurality of corresponding contour points (224 A-N) comprises the step of mapping the plurality of centerline points (222 A-N) and the plurality of contour points (224 A-N) into virtual latitudes (232) and longitudes (234).
  3. The computer-implemented method (100) according to claim 1, further comprising the steps of: - identifying, a region (236) of the vessel (220) in the 3D medical image (205) that comprises an aneurysm (236); - recomputing (104), the centerline (214A) in the identified region (236) for obtaining a modified centerline (214B); and - generating the modified centerline (214B) that replaces the centerline (214A).
  4. The computer-implemented method (100) according to claim 3, wherein the steps for recomputing the centerline (214A) in the identified region (236) comprise (i) interpolating with a polynomial using two points each on either side of the identified region (236) and (ii) resampling the interpolated points to keep them equidistant, such as to identify the modified centerline (214B) in the identified region (236).
  5. The computer-implemented method (100) according to claim 4, further comprising the steps of: - determining the plurality of contours (218 A-N) between the contour points (224 A-N) at the start (238A) and at the end (238B) of the identified region (236) in the 3D medical image (205); - computing a modified radii of the plurality of contours (218 A-N) in the identified region (236) of the 3D medical image (205) by interpolating a radii of the contour points (224 A-N) of the plurality of contours (218 A-N) along the entire length between the start (238A) and the end (238B) of the identified region (236) in the 3D medical image (205); and - generating modified contours from the modified radii and the modified centerline (214B).
  6. The computer-implemented method (100) according to claim 5, wherein the inner curve (212) and the outer curve (216) are computed from the modified centerline (214B) and the modified contour.
  7. The computer-implemented method (100) according to claim 6, using the computed inner curve (212) and the outer curve (216) for estimating a stent length for preprocedural planning and/or computer assisted diagnosis.
  8. A system (600) in particular for preprocedural planning and/or computer assisted diagnosis, the system (600) comprising: a display (648); and one or more processors (642) configured to: - receive a 3D medical image (205) comprising a 3D segmentation (210) of a vessel (220), wherein the 3D segmentation (210) of the vessel (220) comprises a centerline (214A) of the vessel (220) and a plurality of contours (218 A-N) of an inner wall of the vessel (220), wherein the centerline (214A) of the vessel (220) and the plurality of contours (218 A-N) of the inner wall of the vessel (220) are represented as centerline points (222 A-N) and contour points (224 A-N) in a 3D space; - identify an inner curve (212) and an outer curve (216) of the inner wall of the vessel (220) in the 3D segmentation (210) of the vessel (220) using a plurality of direction vectors (228 A-N) orthogonal from the plurality of centerline points (222 A-N) to the plurality of corresponding contour points (224 A-N); and - output the inner curve (212) and the outer curve (216) of the inner wall of the vessel (220) on a display together with the 3D medical image (205) and the 3D segmentation (210) of the vessel (220).
  9. The system (600) according to claim 8, wherein using the plurality of direction vectors (228 A-N) orthogonal from the plurality of centerline points (222 A-N) to the plurality of corresponding contour points (224 A-N) comprises the step of mapping the plurality of centerline points (222 A-N) and the plurality of contour points (224 A-N) into virtual latitudes (232) and longitudes (234).
  10. The system (600) according to claim 8, wherein the one or more processors (642) are further configured to: - identify a region of the vessel (220) in the 3D medical image (205) that comprises an aneurysm; - recompute the centerline (214A) in the identified region (236) to obtain a modified centerline (214B); and - generate the modified centerline (214B) that replaces the centerline (214A).
  11. The system (600) according to claim 10, wherein to recompute the centerline in the identified region (236), the one or more processors (642) are further configured to (i) interpolate with a polynomial using two points each on either side of the identified region (236) and (ii) resample the interpolated points to keep them equidistant, such as to identify the modified centerline (214B) in the identified region (236).
  12. The system (600) according to claim 11, wherein the one or more processors (642) are further configured to: - determine the plurality of contours (218 A-N) between the contour points (224 A-N) at the start (238A) and at the end (238B) of the identified region (236) in the 3D medical image (205); - compute a modified radii of the plurality of contours (218 A-N) in the identified region (236) of the 3D medical image (205) by interpolating a radii of the contour points (224 A-N) of the plurality of contours (218 A-N) along the entire length between the start (238A) and the end (238B) of the identified region (236) in the 3D medical image (205); and - generate modified contours from the modified radii and the modified centerline (214B).
  13. The system (600) according to claim 12, wherein the inner curve (212) and the outer curve (216) are computed from the modified centerline (214B) and the modified contour.
  14. The system (600) according to claim 13, wherein the one or more processors (642) are further configured to estimate a stent length for preprocedural planning and/or computer assisted diagnosis by using the computed inner curve (212) and the outer curve (216).
  15. A computer program comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to any one of claim 1 to 7.

Description

FIELD OF THE INVENTION The invention relates to image processing and more particularly relates to techniques for representing the segmentation of features derived from 3D contours in a medical image. BACKGROUND OF THE INVENTION CT angiography (CTA) is commonly used as a preprocedural imaging tool in various medical contexts, especially for vascular and cardiac procedures. This feature aids the clinicians in preprocedural planning for Endovascular stent-graft placement. The primary purpose of this feature is to aid with an estimation of the stent length to be placed in a vessel. Additionally, the length measurement of the stent could be used to confirm suspected abnormality. Endovascular stents or stent grafts have become a viable treatment for thoracic and abdominal aortic aneurysms in both elective and emergency situations. The ability of stent deployment in a vessel is largely dependent on the compatibility between vessel geometry and the stent hardware. The deployment of the stent in a blood vessel is complex and involves several critical challenges. Such challenges for deployment of the stent in the blood vessel include navigating the stent through tortuous anatomy, managing varying plaque characteristics of the vessels, minimizing the risk of dissection or perforation, and ensuring precise positioning of the stent to cover the entire lesion while avoiding placement over critical branch vessels. One of such critical challenges of stent deployment involves manoeuvring the stent through winding or narrow vessels, especially in patients with complex vascular structures. Another challenge involves improper positioning of the stent in the vessel that can damage the vessel wall during deployment, leading to complications such as dissection or perforation. The complex geometry of vessels, along with patient-specific factors such as vessel size, blood flow dynamics, and existing comorbidities, can influence the complexity and outcome of the procedure. These factors also increase the risk of the stent recoiling after deployment or migrating from its intended location, which could result in inadequate coverage or re-narrowing of the vessel. Wrong stent deployment may lead to complications, such as stent graft migration, endo leaks, blood clot formations, vessel injury as well various graft complications. SUMMARY OF THE INVENTION It is an object of the invention to provide improved preprocedural planning and/or computer aided diagnosis methods, systems and computer program products by aiding clinicians in preprocedural planning for endovascular stent-graft placement and/or computer aided diagnosis. Additionally, the accurate length measurement of the stent may be used to confirm a precise suspected abnormality. The ability of placement of stent-graft is largely dependent on the compatibility between aneurysm geometry and the stent hardware. The inner and outer curves of a vessel play a primary role in establishing this compatibility. The placement of a stent-graft is efficient only when the measurements of the vessel, such as its length, luminal diameter, and collapsed diameter are estimated accurately. These measurements are most precise when they are based on both the inner and outer curves of the vessel. This is because, during stent deployment, the guidewire in a long aneurysm neck is forced to follow the path of the aneurysm neck, navigating along the inner and outer curves. Starting from preprocedural planning through to stent deployment, the inner and outer curves of the vessel play a crucial role. The inner and outer curve lengths of the vessel more accurately reflect the length of the deployed endograft and are more precise than centerlines of the vessel in planning thoracic endografts. The greater the curvature and the larger the vessel, the more significant the underestimation when centerlines are used for length calculation. The concept of using outer and inner curves, especially for the aorta, shows promising results, even accounting for longitudinal strain and axial twists. While centerlines are useful tools in vessel analysis, incorporating the inner and outer curves adds significant value. When both features and measurements are used together in vessel analysis, diagnosis, and preprocedural planning, they can certainly lead to new and improved applications in radiology. To this end the invention provides a system and a computer-implemented method for preprocedural planning and/or computer aided diagnosis. The method comprises the steps of receiving a 3D medical image comprising a 3D segmentation of a vessel. The 3D segmentation comprises a centerline of the vessel and a plurality of contours of an inner wall of the vessel. The centerline of the vessel and the plurality of contours of the inner wall of the vessel are represented as centerline points and contour points in a 3D space. The method comprises the steps of identifying an inner curve and an outer curve of the inner wall of the vessel in the