EP-4736106-A1 - IMAGE ENHANCEMENT OF MEDICAL IMAGES

EP4736106A1EP 4736106 A1EP4736106 A1EP 4736106A1EP-4736106-A1

Abstract

A computer-implemented method for enhancing a medical image, the method comprising: obtaining a first medical image of a subject, the first medical image having a plurality of voxels, the first medical image missing voxels in a portion of a region of interest; obtaining a second medical image of the subject, the second medical image having a plurality of voxels, the second medical image having voxels in the portion of the region of interest; and generating a combined medical image based on the first medical image and the second medical image, wherein the combined medical image is not missing voxels in the portion of the region of interest.

Inventors

URMAN, Noa
GLOZMAN, YANA

Assignees

Novocure GmbH

Dates

Publication Date: 20260506
Application Date: 20240618

Claims (15)

1. A computer-implemented method for enhancing a medical image, the method comprising: obtaining a first medical image of a subject, the first medical image having a plurality of voxels, the first medical image missing voxels in a portion of a region of interest; obtaining a second medical image of the subject, the second medical image having a plurality of voxels, the second medical image having voxels in the portion of the region of interest; and generating a combined medical image based on the first medical image and the second medical image, wherein the combined medical image is not missing voxels in the portion of the region of interest.
2. The method of claim 1, wherein at least one of: the first medical image is missing slices in the portion of the region of interest, wherein the second medical image has voxels in the portion of the region of interest where the first medical image is missing slices; the first medical image is truncated in the region of interest, wherein the second medical image has voxels in the portion of the region of interest where the first medical image is truncated; the first medical image has a lower resolution in a first direction than the second medical image; or the first medical image and the second medical image have a same magnetic resonance image (MRI) modality, wherein the first medical image and the second medical image have different acquisition orientations.
3. The method of claim 1, further comprising: padding the first medical image to replace the missing voxels in the portion of the region of interest to obtain a padded first medical image; padding the second medical image to replace any voxels missing in the portion of the region of interest to obtain a padded second medical image; resampling the padded first medical image and the padded second medical image with a same spacing size to obtain a resampled first medical image and a resampled second medical image; aligning the resampled second medical image to the resampled first medical image to obtain an aligned second medical image; and performing a same image processing on the resampled first medical image and the aligned second medical image to obtain a processed first medical image and a processed second medical image, wherein the combined medical image is generated based on the processed first medical image and the processed second medical image.
4. The method of claim 3, wherein performing the same image processing comprises: performing histogram matching on the resampled first medical image; and performing histogram matching on the aligned second medical image.
5. The method of claim 3, wherein the values for voxels in the processed first medical image are based on a histogram matching of the first medical image, and wherein the values for voxels in the processed second medical image are based on a histogram matching of the second medical image.
6. The method of claim 3, wherein performing the same image processing comprises: generating a first weighting map for the resampled first medical image; and generating a second weighting map for the aligned second medical image.
7. The method of claim 6, wherein the first weighting map has values representing a distance between a voxel in the first medical image and a same voxel in the resampled first medical image, and wherein the second weighting map has values representing a distance between a voxel in the first medical image and a same voxel in the aligned second medical image.
8. The method of claim 3, wherein performing the same image processing comprises: performing histogram matching on the resampled first medical image; performing histogram matching on the aligned second medical image; generating a first weighting map for the resampled first medical image; and generating a second weighting map for the aligned second medical image.
9. The method of claim 3, wherein generating the combined medical image comprises: for each voxel in the combined medical image, summing a first combination of a corresponding voxel in the processed first medical image and a voxel corresponding in the resampled first medical image, and a second combination of a corresponding voxel in the processed second medical image and a corresponding voxel in the aligned second medical image.
10. The method of claim 1, further comprising: generating at least one transducer location for delivering tumor treating fields to the subject based on the combined medical image.
11. A computer-implemented method for enhancing a medical image, the method comprising: obtaining a first medical image of a subject, the first medical image having a plurality of voxels, the first medical image missing slices in a portion of a region of interest; obtaining a second medical image of the subject, the second medical image having a plurality of voxels, wherein the second medical image has voxels in the portion of the region of interest where the first medical image is missing slices; performing histogram matching on the first medical image to obtain a first histogram image; performing histogram matching on the second medical image to obtain a second histogram image; and generating a combined medical image based on the resampled first medical image, the aligned second medical image, the first histogram image, and the second histogram image, wherein the combined medical image is not missing slices in the portion of the region of interest.
12. The method of claim 11, wherein generating the combined medical image comprises: for each voxel in the combined medical image, comparing a value for the voxel in the first histogram image and a value for the voxel in second histogram image to determine which histogram image has the value with a lowest gray level; if the first histogram image has the value with a lowest gray level, copying the corresponding voxel from the resampled first medical image into the combined medical image; and if the second histogram image has the value with a lowest gray level, copying the corresponding voxel from the aligned second medical image into the combined medical image.
13. A computer-implemented method for enhancing a medical image, the method comprising: obtaining a first medical image of a subject, the first medical image having a plurality of voxels; obtaining a second medical image of the subject, the second medical image having a plurality of voxels, wherein the first medical image has a lower resolution in a first direction than the second medical image; generating a first weighting map for the first medical image having values representing a distance between a voxel in the first medical image and a same voxel in a resampled first medical image; generating a second weighting map for the second medical image having values representing a distance between a voxel in the first medical image and a same voxel in a resampled second medical image; and generating a combined medical image based on the first medical image, the second medical image, the first weighting map, and the second weighting map, wherein the combined medical image has the lower resolution in the first direction.
14. The method of claim 13, wherein generating the first weighting map comprises generating a first checkerboard image for the first medical image, and wherein generating the second weighting map comprises generating a second checkerboard image for the second medical image.
15. The method of claim 13, wherein generating the combined medical image comprises: for each voxel, summing a product of the first weighting map and the resampled first medical image and a product of the second weighting map and the resampled second medical image.

Description

IMAGE ENHANCEMENT OF MEDICAL IMAGES CROSS-REFERENCE TO RELATED APPLICATIONS [0001] This application claims priority to U.S. Patent Application No. 18/744,991, filed June 17, 2024, and U.S. Provisional Application No. 63/524,574, filed June 30, 2023, which are incorporated herein by reference in their entirety. BACKGROUND [0002] Tumor treating fields (TTFields) are low intensity alternating electric fields within the intermediate frequency range (for example, 50 kHz to 1 MHz), which may be used to treat tumors as described in U.S. Patent No. 7,565,205. TTFields are induced non-invasively into the region of interest by transducers placed on the patient’s body and applying alternating current (AC) voltages between the transducers. Conventionally, transducers used to generate TTFields include a plurality of electrode elements including ceramic disks. One side of each ceramic disk is positioned against the patient’s skin, and the other side of each disc has a conductive backing. Electrical signals are applied to this conductive backing, and these signals are capacitively coupled into the patient’s body through the ceramic discs. Conventional transducer designs include arrays of ceramic disks attached to a subject’s body via a conductive skin-contact layer such as a hydrogel. AC voltage is applied between a pair of transducers for an interval of time to generate an electric field with field lines generally running in the front-back direction. Then, AC voltage is applied at the same frequency between at least another pair of transducers for another interval of time to generate an electric field with field lines generally running in the right-left direction. The system then repeats this two-step sequence throughout the treatment. BRIEF DESCRIPTION OF THE DRAWINGS [0003] FIG. 1 depicts an example method for generating a combined medical image. [0004] FIG. 2 depicts an example method for performing image matching based on histograms. [0005] FIG. 3 depicts an example method for performing image matching based on checkboard images. [0006] FIG. 4 depicts an example method for performing image matching based on checkboard images. [0007] FIG. 5 depicts an example flowchart for performing the example method described in FIGS. 1-4. [0008] FIG. 6 depicts an example apparatus to apply alternating electric fields to a subject’s body. [0009] FIGS. 7 A and 7B depict schematic views of exemplary design of a transducer for applying alternating electric fields. [0010] FIG. 8 depicts an example placement of transducers on a subject’s head. [0011] FIG. 9 depicts an example computer apparatus. DESCRIPTION OF EMBODIMENTS [0012] This application describes exemplary techniques utilizing computer algorithms for enhancing medical images missing information, such as having slices missing, being truncated, or having a lower resolution. [0013] When administering TTFields to a subject, such as a patient, one or more medical images are read and analyzed to determine a treatment plan for the subject. Traditionally, it may take approximately an hour to obtain a set of full resolution medical image(s) of the subject, while a rapid-taken medical image may only take several minutes. However, a rapid-taken medical image may not have complete information, such as having lower resolution or missing slices, and may not provide the most accurate determination. [0014] The inventors recognized that during a procedure of reading the medical images for TTFields treatment determination, a need exists for filling and enhancing medical images that have missing information or have a low quality based on existing medical images of the same subject with full or better information. [0015] The methods and systems described herein provide a practical application to fill in and/or enhance a medical image having missing voxels. By filling in and/or enhancing such a medical image, a more complete view of a subject can be obtained having more information about the subject. With a medical image having more information on the subject, a more accurate three-dimensional computational model of the subject can be obtained. With a more accurate three-dimensional computational model of the subject, a more accurate determination of where to place transducer arrays on the subject for delivering TTFields can be obtained. [0016] In particular, the inventors discovered computational techniques to fill in and/or enhance a medical image having missing voxels. Exemplary methods and systems provide for filling missing portions of a first medical image based on a second medical image that has voxels covering the missing portions and generating a combined medical image. In some embodiments, computer algorithm techniques may use histogram matching of the first and second medical images to generate the combined medical image. In some embodiments, computer algorithm technique may use a checkerboard algorithm that generates first and second weighting maps (based on first and second check