Search

EP-4736107-A1 - ANATOMICAL CONTOURING

EP4736107A1EP 4736107 A1EP4736107 A1EP 4736107A1EP-4736107-A1

Abstract

Proposed concepts thus aim to provide schemes, solutions, concept, designs, methods and systems pertaining to prioritizing medical images for anatomical contouring. In particular, embodiments aim to provide a method for prioritizing medical images for anatomical contouring by processing a plurality of medical images in order to generate at least one visual characteristic value for each respective image. The visual characteristic value for each image is based, in part, on obtained boundary data describing a region of at least one of the images, such as an unfinished contour line. The region describes a boundary and the visual characteristic value comprises a contrast value between pixels on either side of the boundary. A priority image can then be determined based on the plurality of visual characteristic values.

Inventors

  • PAPACONSTADOPOULOS, Pavlos
  • KLAASSEN, Remy
  • HEIJMAN, EDWIN

Assignees

  • Koninklijke Philips N.V.

Dates

Publication Date
20260506
Application Date
20240627

Claims (15)

  1. 1. A computer-implemented method (100) for prioritizing medical images for anatomical contouring, the method comprising: obtaining a plurality of medical images (110) including the same anatomical region, each medical image being captured using a different imaging modality or imaging sequence; obtaining boundary data (120) describing a region of at least one of the plurality of medical images, wherein the region describes a boundary; processing each of the plurality of medical images (130) based on the boundary data to generate at least one visual characteristic value for each respective medical image, wherein the visual characteristic value comprises a contrast value between pixels on either side of the boundary; and determining a priority image (140) of the plurality of medical image based on the generated plurality of visual characteristic values.
  2. 2. The method of claim 1, wherein the visual characteristic value comprises a combination of the contrast value and at least one of a visual characteristic value for: sharpness; tone; gamma; texture; dynamic range; colour; saturation; and white balance.
  3. 3. The method of any of claims 1 or 2, wherein an imaging modality or imaging sequence comprises at least one of: conventional CT imaging; spectral CT imaging; monoenergetic CT imaging; iodine density imaging; Z-Effective imaging; x-ray imaging; MRI imaging; multi -parametric MRI imaging; and PSMA PET/CT imaging.
  4. 4. The method of any prior claim, wherein obtaining boundary data comprises: analyzing at least one of the plurality of medical images (320) to generate boundary data describing a region for the at least one image respectively.
  5. 5. The method of any prior claim, wherein in the step of processing each of the plurality of medical images (130) only a subset of each of the plurality of medical images is processed.
  6. 6. The method of any of the preceding claims, wherein the boundary comprises a boundary generated by a user.
  7. 7. The method of any of the preceding claims, wherein the boundary comprises an open boundary.
  8. 8. The method of any prior claim, wherein processing each of the plurality of medical images comprises: generating corresponding boundary data (325) for each of a remaining plurality of medical images describing corresponding regions in each of the remaining plurality of medical images respectively; and processing each of the plurality of medical images (330) based on their corresponding boundary data respectively to generate at least one visual characteristic value for each respective medical image.
  9. 9. The method of any prior claim, wherein the method further comprises generating a signal (350) for informing a user that the priority image is a different image to a medical image the user is currently viewing; and, preferably, outputting the generated signal (355).
  10. 10. The method of claim 9, wherein the generated signal comprises a signal configured to make an input device vibrate.
  11. 11. The method of any prior claim, wherein the method further comprises: in response to receiving an acceptance signal from a user, outputting the priority image (360) on at least one display device.
  12. 12. The method of any of the preceding claims 1-9, wherein the method further comprises outputting the priority image (360) on at least one display device.
  13. 13. A computer program comprising code means for implementing the method of any preceding claim when said program is run on a processing system.
  14. 14. Computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the steps of the method according to any of claims 1-12.
  15. 15. A system (400) for prioritizing medical images for anatomical contouring, the system comprising: an input interface (410) configured to: obtain a plurality of medical images (110) including the same anatomical region, each medical image being captured using a different imaging modality or imaging sequence; and obtain boundary data (120) describing a region of at least one of the plurality of medical images, wherein the region describes a boundary; and a processor arrangement (420) configured to: process each of the plurality of medical images (130) based on the boundary data to generate at least one visual characteristic value for each respective medical image, wherein the visual characteristic value comprises a contrast value between pixels on either side of the boundary; and determine a priority image (140) of the plurality of medical images based on the generated plurality of visual characteristic values.

Description

ANATOMICAL CONTOURING FIELD OF THE INVENTION This invention relates to the field of medical images for anatomical contouring, and in particular, to the field of prioritizing medical images for anatomical contouring. BACKGROUND OF THE INVENTION During clinical pre-treatment phases in Oncology, a clinical expert (e.g., a radiologist, interventional radiologist, medical or radiation oncologist) typically utilizes patient medical image scans to localize and delineate the tumour(s). Such activity traditionally occurs with the aid of an electronic contouring tool and is often described as contouring (or delineation). This process is of utmost critical nature as it is the starting point of identifying which area of the body needs to be targeted for biopsy or removed, ablated, or irradiated depending on the choice of treatment. The contouring tool is often an electronic paint brush or pen which allows the clinician to directly draw the tumour in 3D image space, slice-by-slice. The most common data format for storing the contours are DICOM files, similarly to medical images. The contours are saved as a series of (x, y, z) coordinates in the DICOM file, following the coordinate system of the reference image. During the activity of contouring the clinician essentially defines the boundary that separates the tumor to the healthy tissue. Such boundaries can typically be depicted visually as a contrast differentiation level between the different anatomical regions. An ideal image modality would offer 100 % contrast differentiation between healthy tissue and cancer, making contouring much easier. However, no image modality is perfect. Different imaging modalities may offer optimum contrast differentiation in some clinical settings or tissue characteristics but not in others. Historically, a single image modality, i.e. Computed Tomography (CT), is chosen for scanning the patient and performing the delineation. However, in recent years the role and importance of multi-modality imaging has been elevated. The principal reasons behind this shift is the greater accessibility of various image modalities within a hospital as well as the appreciation that different image modalities offer complementary information to each other and should be used for tumour delineation. A good example of such a situation is prostate cancer: multi-parametric MRI (mpMRI) is considered broadly now as the standard of care for prostate cancer diagnosis. International guidelines (e.g. PI-RADS) require that various mpMRI image sequences (e.g. T2w, DWI, DCE) are reviewed in order to identify the intra-prostatic tumour lesions. In addition, PSMA PET/CT is considered superior for Lymph Node (LN) involvement assessment for prostate cancer. As Spectral CT is also evolving, it will offer complementary information in both the primary tumour and LN assessment, via even more multi-dimensional information (e.g. Z-effective maps, Iodine maps or various virtual monoenergetic beams). In current practice, but also in future, it can only be expected that the utility of multimodality imaging will be even further enhanced. The situation described above demonstrates that there is already a vast and diverse range of image modalities that the physician needs to consult during contouring. However, the tools via which the physician delineates/contours have not changed. Typically, an electronic contour brush/pen is still used on a single image modality, slice-by-slice, to delineate the boundary between tumour and healthy tissue. Clinicians may indeed use some image fusion and toggle between 1 or 2 image modalities, or even consult another image modality on another screen at the same time to enrich their decision making, but the approach is mostly empirical and ad-hoc. The problem is that clinicians struggle in interpreting and utilizing multi-dimensional data in an effective way as it goes beyond the natural human experience. The rise of multimodality imaging for tumor delineation is an issue for clinicians as they need to understand which modalities to use and how to use them to accurately contour the tumour. Even though efforts have been made in standardizing the preferred image sequences per clinical site, there is surprisingly little development in advanced tools that physicians can use for this purpose. SUMMARY OF THE INVENTION The invention is defined by the claims. According to examples in accordance with an aspect of the invention, there is provided a method for prioritizing medical images for anatomical contouring. The method comprises: obtaining a plurality of medical images including the same anatomical region, each medical image being captured using a different imaging modality or imaging sequence; obtaining boundary data describing a region of at least one of the plurality of medical images; processing each of the plurality of medical images based on the boundary data to generate at least one visual characteristic value for each respective medical image; and dete