# US 2022/0284585 A1 · Methods and Apparatuses for Visualization of Tumor Segmentation

> Agents: this data is available as JSON from the patentgrep API and MCP server — see https://patentgrep.com/llms.txt

- Number: US 2022/0284585 A1
- Document ID: US20220284585A1
- Status: Published application
- Kind: A1
- Application number: 17/683,643
- Filed: 2022-03-01
- Published: 2022-09-08
- Earliest claimed: 2021-03-02
- Assignees: Novocure GmbH
- Applicants: Novocure GmbH
- Inventors: Michal HOLTZMAN GAZIT; Reuven Ruby SHAMIR
- Web page: https://patentgrep.com/patent/US20220284585A1

## CPC

- G06T 7/11: Region-based segmentation
- G06T 7/136: involving thresholding
- A61B 2034/105: {Modelling of the patient, e.g. for ligaments or bones}
- G06T 2207/20081: Training; Learning
- G06T 2207/20084: Artificial neural networks [ANN]
- G06T 2207/30096: Tumor; Lesion
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)

## Abstract

A computer-implemented method for tumor segmentation, the method comprises obtaining image data of a region of interest of a subject's body, wherein the region of interest corresponds to a tumor of the subject's body, generating two or more tumor segmentation predictions based on the image data, calculating a divergence between the two or more tumor segmentation predictions, and generating a visualization of tumor segmentation uncertainty based on the calculated divergence between the two or more tumor segmentation predictions.

## Claims

### Claim 1

A computer-implemented method for tumor segmentation, the method comprises:
obtaining image data of a region of interest of a subject's body, wherein the region of interest corresponds to a tumor of the subject's body;
generating two or more tumor segmentation predictions based on the image data;
calculating a divergence between the two or more tumor segmentation predictions;
and
generating a visualization of tumor segmentation uncertainty based on the calculated divergence between the two or more tumor segmentation predictions.

### Claim 2 (depends on 1)

The method of claim 1, further comprising:
obtaining anatomical information and surgical information of the subject's body, wherein the two or more tumor segmentation predictions are generated based on the image data, the anatomical information, and the surgical information.

### Claim 3 (depends on 2)

The method of claim 2, further comprising:
adding or removing anatomical structures from the visualization based on the anatomical information and surgical information of the subject's body.

### Claim 4 (depends on 1)

The method of claim 1, wherein the divergence is calculated through a symmetric Kullback-Leibler (KL) divergence loss.

### Claim 5 (depends on 1)

The method of claim 1, wherein the two or more tumor segmentation predictions are generated by segmentation networks, wherein each of the segmentation networks comprises a variational auto-encoder that reconstructs the image data from a shared encoder parameter.

### Claim 6 (depends on 1)

The method of claim 1, wherein the visualization comprises an image of the subject with a segmentation prediction and an uncertainty map, wherein the uncertainty map is based on the calculated divergence between the two or more tumor segmentation predictions.

### Claim 7 (depends on 6)

The method of claim 6, further comprising:
removing voxels from the segmentation prediction having an uncertainty greater than a threshold.

### Claim 8

A computer-implemented method for tumor segmentation, the method comprises
obtaining two or more segmentation networks trained using a common training set with a plurality of training losses, the training set comprising images of other subjects;
obtaining image data of a region of interest of a subject's body, wherein the region of interest corresponds to a tumor of the subject's body;
generating two or more tumor segmentation predictions based on the image data and the two or more segmentation networks;
calculating a divergence between the two or more tumor segmentation predictions;
and
generating a visualization of tumor segmentation uncertainty based on the calculated divergence between the two or more tumor segmentation predictions.

### Claim 9 (depends on 8)

The method of claim 8, wherein the segmentations networks are trained in parallel using a unified loss, wherein the unified loss comprises at least one of a segmentation loss or a divergence loss.

### Claim 10 (depends on 8)

The method of claim 8, further comprising:
applying augmentation to the training data set to obtain an augmented training set, wherein the augmentation comprises at least one of: intensity normalization, random shift, random scale-up of intensity value, random flip, or random scaling of an image patch; and
training the segmentation networks using the augmented training data set.

### Claim 11 (depends on 8)

The method of claim 8, further comprising:
obtaining anatomical information and surgical information of the subject's body, wherein the two or more tumor segmentation predictions are generated based on the image data, the two or more trained segmentation networks, the anatomical information, and the surgical information.

### Claim 12 (depends on 11)

The method of claim 11, further comprising:
adding or removing anatomical structures from the visualization based on the anatomical information and surgical information of the subject's body.

### Claim 13 (depends on 8)

The method of claim 8, further comprising:
determining locations on the subject's body to place transducers for applying tumor treating fields based on at least one of the visualization of tumor segmentation uncertainty or two or more tumor segmentation predictions.

### Claim 14

An apparatus comprising: one or more processors; and memory storing processor-executable instructions that, when executed by the one or more processors, cause the apparatus to:
obtain a training data set of images of other subjects;
training two or more segmentation networks using the training set and a unified loss based on a comparison between training outputs of the two or more segmentation networks; and
obtain image data of a region of interest of a subject's body, wherein the region of interest corresponds to a tumor of the subject's body;
generate two or more segmentation predictions based on the image data and the two or more trained segmentation networks; and
calculate a divergence between the two or more segmentation predictions; and
generate a visualization of tumor segmentation uncertainty based on the segmentation prediction, wherein the visualization of segmentation uncertainty is generated based on a calculated divergence between the two or more segmentation predictions.

### Claim 15 (depends on 14)

The apparatus of claim 14, wherein the unified loss comprises a segmentation loss, a divergence loss, and a reconstruction loss.

### Claim 16 (depends on 15)

The apparatus of claim 15, wherein the segmentation loss comprises a Dice coefficient and a cross entropy loss.

### Claim 17 (depends on 15)

The apparatus of claim 15, wherein the divergence loss comprises a Kullback-Leibler divergence loss.

### Claim 18 (depends on 15)

The apparatus of claim 15, wherein each of the segmentation networks comprises a variational auto-encoder, and the reconstruction loss comprises a reconstruction loss of the variational auto-encoder for each segmentation network.

### Claim 19 (depends on 15)

The apparatus of claim 15, wherein the unified loss comprises a parameter to balance the segmentation loss and the divergence loss.

### Claim 20 (depends on 14)

The apparatus of claim 14, wherein the memory storing processor-executable instructions that, when executed by the one or more processors, further cause the apparatus to:
obtain anatomical information and surgical information of the subject's body, wherein the two or more segmentation predictions are generated based on the image data, the two or more trained segmentation networks, the anatomical information, and the surgical information.
