# US 12,217,429 B2 · 3D radiomic platform for imaging biomarker development

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

- Number: US 12,217,429 B2
- Document ID: US12217429B2
- Status: Granted
- Kind: B2
- Application number: 17/515,052
- Filed: 2021-10-29
- Published: 2022-02-17
- Granted: 2025-02-04
- Earliest claimed: 2018-07-02
- Assignees: Tempus AI, Inc.
- Applicants: Tempus Ai, Inc.
- Inventors: Stephen Yip
- Web page: https://patentgrep.com/patent/US12217429B2

## CPC

- G06T 7/0016: {involving temporal comparison}
- A61B 5/0059: {using light, e.g. diagnosis by transillumination, diascopy, fluorescence (photoacoustic A61B5/0093; optical measurement of heart rate A61B5/02416; optical measurement of blood flow A61B5/0261; optical measurement of analytes A61B5/1455)}
- A61B 5/4848: {Monitoring or testing the effects of treatment, e.g. of medication}
- A61B 5/725: {using specific filters therefor, e.g. Kalman or adaptive filters (specific diagnostics methods using using bioelectric or biomagnetic signals A61B5/316)}
- A61B 5/726: {using Wavelet transforms}
- G06T 7/0012: {Biomedical image inspection}
- G06T 7/10: Segmentation; Edge detection (motion-based segmentation G06T7/215)
- G06T 7/174: involving the use of two or more images
- G06T 7/40: Analysis of texture (depth or shape recovery from texture G06T7/529)
- G06T 7/60: Analysis of geometric attributes
- G06T 17/00: Three-dimensional [3D] modelling for computer graphics
- G06T 2207/10081: Computed x-ray tomography [CT]
- G06T 2207/10088: Magnetic resonance imaging [MRI]
- G06T 2207/10104: Positron emission tomography [PET]
- G06T 2207/10116: X-ray image
- G06T 2207/30016: Brain
- G06T 2207/30028: Colon; Small intestine
- G06T 2207/30056: Liver; Hepatic
- G06T 2207/30061: Lung
- G06T 2207/30068: Mammography; Breast
- G06T 2207/30088: Skin; Dermal
- G06T 2207/30092: Stomach; Gastric
- G06T 2207/30096: Tumor; Lesion
- G06T 2210/41: Medical
- G06T: IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- A61B: DIAGNOSIS; SURGERY; IDENTIFICATION (usefulness limited to only animals A61D)

## Abstract

A platform is provided for generating 3D models of a tumor segmented from a series of 2D medical images and for identifying from these 3D models, radiomic features that may be used for diagnostic, prognostic, and treatment response assessment of the tumor. The radiomic features may be shape-based features, intensity-based features, textural features, and filter-based features. The radiomic features are compared to remove sufficiently redundant features, thereby producing a reduced set of radiomic features, which is then compared to separate genomic data and/or outcome data to identify clinically and biologically significant radiomic features for diagnostic, prognostic, and treatment response assessment, other applications.

## Claims

### Claim 1

A computer-implemented method to analyze medical image data, the method comprising:
obtaining, using one or more processors, a three-dimensional (3D) model of medical image data;
detecting, using the one or more processors, target tissue in the 3D model, the target tissue comprising tumor tissue and surrounding tissue;
identifying, using the one or more processors, a master set of three-dimensional (3D) radiomic features from the 3D model, the master set of 3D radiomic features corresponding to the target tissue and clinical outcome data;
associating, using the one or more processors, the master set of 3D radiomic features with one or more of clinical data and molecular data;
producing, using the one or more processors, one or more integrated radiomic-based indicators by characterizing each radiomic feature in the master set based on a machine learning-derived relevance score that quantitatively assesses the feature's predictive accuracy for patient-specific therapeutic outcomes; and
storing the one or more integrated radiomic-based indicators for assessing subsequent target tissue in subsequent medical image data.

### Claim 2 (depends on 1)

The method of claim 1, wherein the clinical outcome data is one or more of disease-free survival, tumor stage, response rate to therapeutics, absence of disease, and quality of life.

### Claim 3 (depends on 1)

The method of claim 1, wherein the molecular data is one or more of gene expression irregularities, irregular microRNA profile, splice variations, and DNA methylation.

### Claim 4 (depends on 1)

The method of claim 1, wherein the 3D radiomic features are imaging data that correlate to molecular or genomic data, treatment response data, diagnostic data, prognostic data, and/or classifiers of cancer patient risk stratification.

### Claim 5 (depends on 1)

The method of claim 1, wherein the 3D model of medical image data is generated from medical image data taken before a tumor therapy and medical image data taken after a tumor therapy.

### Claim 6 (depends on 1)

The method of claim 1, wherein the 3D model of medical image data is generated from a plurality of two-dimensional (2D) medical images.

### Claim 7 (depends on 1)

The method of claim 1, wherein identifying the master set of 3D radiomic features from the 3D model comprises identifying redundant 3D radiomic features using a pair-wise comparison and removing the 3D radiomic features to form the master set of 3D radiomic features.

### Claim 8 (depends on 1)

The method of claim 1, wherein the tumor tissue comprises breast cancer tissue, colon cancer tissue, gastric cancer tissue, endometrium cancer tissue, ovarian cancer tissue, hepatobiliary tract cancer tissue, urinary tract cancer tissue, lung cancer tissue, brain cancer tissue, or skin cancer tissue.

### Claim 9 (depends on 1)

The method of claim 1, further comprising:
identifying from among the one or more integrated radiomic-based indicators, a subset of integrated radiomic-based indicators significantly associated with clinical outcomes and/or genomic data.

### Claim 10 (depends on 9)

The method of claim 9, further comprising:
identifying from among the one or more integrated radiomic-based indicators, a subset of integrated radiomic-based indicators significantly associated with clinical outcomes and/or genomic data using a Cox proportional hazards model.

### Claim 11 (depends on 10)

The method of claim 10, wherein the subset of integrated radiomic-based indicators significantly associated with clinical outcomes and/or genomic have a p-value<0.005 using the Cox proportional hazards model.

### Claim 12 (depends on 1)

The computer-implemented method of claim 1, further comprising:
retrieving a three-dimensional (3D) model of a tumor;
identifying a set of integrated radiomic-based indicators based on the 3D model of the tumor, wherein the set of integrated radiomic-based indicators includes at least one of shape-based features, intensity-based features, textural features, and filter-based features that are correlative of a treatment response corresponding to a specific outcome;
training a model using the set of integrated radiomic-based indicators to assess the treatment response; and
providing an assessment of the treatment response based on the trained model.

### Claim 13

A computer-implemented method to analyze medical image data of a new subject, the method comprising:
obtaining, using one or more processors, a first three-dimensional (3D) model of medical image data of the new subject corresponding to a first time period before a therapy;
obtaining, using one or more processors, a second 3D model of medical image data of the new subject corresponding to a second time period after the first time period and associated with the therapy;
comparing, using the one or more processors, the first 3D model to the second 3D model, wherein the comparing includes:
(1) generating data indicating changes in 3D radiomic features between the first time period and the second time period, and
(2) determining changes in one or more integrated radiomic-based indicators between the first time period and the second time period, the one or more integrated radiomic-based indicators including a plurality of associations of the new subject between the 3D radiomic features and one or more of clinical data and molecular data; and
generating, by comparing the associations of the new subject to a plurality of associations of a cohort of different subjects, at least one of treatment response data, diagnostic data or prognostic data.

### Claim 14 (depends on 13)

The method of claim 13, wherein the therapy is a chemotherapy.

### Claim 15 (depends on 13)

The method of claim 13, wherein the therapy is a radiation.

### Claim 16 (depends on 13)

The method of claim 13, wherein the therapy is an immunotherapy.

### Claim 17 (depends on 13)

The method of claim 13, wherein the therapy is a poly ADP ribose polymerase (PARP) inhibitors therapy.

### Claim 18 (depends on 13)

The method of claim 13, wherein the therapy is a CAR T-cell therapy.

### Claim 19 (depends on 13)

The method of claim 13, wherein the therapy is a cancer vaccine.

### Claim 20 (depends on 13)

The method of claim 13, further comprising: determining efficacy of the therapy from the changes in the integrated radiomic-based indicators between the first time period and the second time period.

### Claim 21 (depends on 20)

The method of claim 20, further comprising: determining a next therapy in response to determining the efficacy of the therapy.

### Claim 22 (depends on 13)

The method of claim 13, further comprising:
identifying the one or more integrated radiomic-based indicators as a subset of integrated radiomic-based indicators from a master set of integrated radiomic-based indicators, the subset being significantly associated with clinical outcomes and/or genomic data.

### Claim 23 (depends on 22)

The method of claim 22, further comprising:
identifying the subset being significantly associated with clinical outcomes and/or genomic data using a Cox proportional hazards model.

### Claim 24 (depends on 23)

The method of claim 23, further comprising:
identifying the subset being significantly associated with clinical outcomes and/or genomic data as having a p-value<0.005 using the Cox proportional hazards model.

### Claim 25 (depends on 13)

The computer-implemented method of claim 13, further comprising:
determining, using the one or more processors, the plurality of associations of the cohort of the different subjects by:
for each of the subjects:
(a) identifying a respective master set of 3D radiomic features by processing at least two 3D models of medical image data each corresponding to a respective target tissue of the subject, and
(b) generating a respective selected radiomic feature set by determining which of the master set of 3D radiomic features of the subject are associated with clinical data and molecular data.

### Claim 26 (depends on 25)

The method of claim 25, wherein the target tissue comprises tumor tissue and surrounding tissue.