3d radiomic platform for imaging biomarker development

Patent metadata

Field	Value
Publication number	US-2020005461-A1
Application number	US-201916460975-A
Country	US
Kind code	A1
Filing date	Jul 2, 2019
Priority date	Jul 2, 2018
Publication date	Jan 2, 2020
Grant date	—

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.

First claim

What is claimed:

1 . A computer-implemented method to analyze medical image data, the method comprising:
obtaining, using one or more processors, the medical image data comprising a plurality of two-dimensional (2D) medical images; performing, using the one or more processors, target tissue detection and target tissue segmentation for each 2D medical image to produce a set of segmented target tissue images; generating, using the one or more processors, a three-dimensional (3D) model of detected and segmented target tissue; identifying, using the one or more processors, a master set of radiomic features for the 3D model of detected and segmented target tissue; comparing, using the one or more processors, at least some of the radiomic features in the master set to identify redundant radiomic features for the 3D model of detected and segmented target tissue; excluding, using the one or more processors, the redundant radiomic features from the 3D model of detected and segmented target tissue; and extracting, using the one or more processors, a selected set of radiomic features from the 3D model of detected and segmented target tissue.

2 . The method of claim 1 , wherein the target tissue is tumor tissue.

3 . The method of claim 2 , 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.

4 . The method of claim 1 , wherein the target tissue is tumor tissue and surrounding tissue.

5 . The method of claim 1 , further comprising:
identifying from among the selected set of radiomic features, radiomic features that are significantly associated with clinical outcomes and/or genomic data.

6 . The method of claim 5 , further comprising:
identifying from among the selected set of radiomic features, the radiomic features that are significantly associated with clinical outcomes and/or genomic data using a Cox proportional hazards model.

7 . The method of claim 6 , wherein the radiomic features significantly associated with clinical outcomes and/or genomic data have a p-value < 0 . 005 using the Cox proportional hazards model.

8 . The method of claim 6 , wherein the radiomic features are shape-based features.

9 . The method of claim 6 , wherein the radiomic features are intensity-based features.

10 . The method of claim 6 , wherein the radiomic features are textural features.

11 . The method of claim 6 , wherein the radiomic features are filter-based features.

12 . The method of claim 1 , further comprising identify redundant radiomic features of the 3D model using a pair-wise comparison.

13 . The method of claim 1 , further comprising:
performing the method of claim 1 (i) at a first time period before a tumor therapy and again (ii) at a second time period after the tumor therapy.

14 . The method of claim 13 , further comprising:
determining changes in radiomic features from the first time period before the tumor therapy to the second time period after the tumor therapy.

15 . The method of claim 14 , wherein the therapy is a chemotherapy.

16 . The method of claim 14 , wherein the therapy is a radiation.

17 . The method of claim 14 , wherein the therapy is an immunotherapy.

18 . The method of claim 14 , wherein the therapy is a poly ADP ribose polymerase (PARP) inhibitors therapy.

19 . The method of claim 14 , wherein the therapy is a CAR T-cell therapy.

20 . The method of claim 14 , wherein the therapy is a cancer vaccine.

21 . The method of claim 14 , further comprising: determining efficacy of the therapy from the changes in radiomic features from the first time period before the tumor therapy to the second time period after the tumor therapy.

22 . The method of claim 21 , further comprising: determining a next therapy in response to determining the efficacy of the therapy.

23 . A computer-implemented method for generating radiomic features for use in a 3D model of a tumor, the method comprising:
performing, using the one or more processors, using a convolution neural network, model target tissue detection and target tissue segmentation for a plurality of 2D medical images to produce a set of segmented target tissue images; generating, using the one or more processors, a 3D model of detected and segmented target tissue from a plurality of 2D medical images; identifying, using the one or more processors, using a convolution neural network, a master set of radiomic features for the 3D model of detected and segmented target tissue; identifying, using the one or more processors, using a statistical model, radiomic features that are significantly associated with clinical outcomes and/or genomic data; and storing or displaying the significantly associated radiomic features in an enhanced 3D model or in a digital report.

24 . A computing device comprising:
one or more processors; a user interface; and a computer-readable memory coupled to the one or more processors, the memory storing instructions that cause the one or more processors to: perform target tissue detection and target tissue segmentation for each of a plurality of 2D medical image and produce a set of segmented target tissue images; generate a 3D model of detected and segmented target tissue; identify a master set of radiomic features for the 3D model of detected and segmented target tissue; compare at least some of the radiomic features in the master set to identify redundant radiomic features for the 3D model of detected and segmented target tissue; exclude the redundant radiomic features from the 3D model of detected and segmented target tissue; extract a selected set of radiomic features from the 3D model of detected and segmented target tissue; and store or display the extracted set of radiomic features in an enhanced 3D model or in a digital report.

25 . The computing device of claim 24 , wherein the memory stores instructions that cause the one or more processors to: analyze the selected set of radiomic features and identify from the selected set of radiomic features, radiomic features that are significantly associated with clinical outcomes and/or genomic data.

26 . The computing device of claim 25 , wherein the memory stores instructions that cause the one or more processors to: identify the significantly associated radiomic features using a Cox proportional hazards model, wherein the significantly associated radiomic features have a p-value<0.005.

27 . The computing device of claim 25 , wherein the memory stores instructions that cause the one or more processors to: identify the significantly associated radiomic features correlated to tumor treatment effectiveness.

28 . A computer-implemented method to analyze medical image data, the method comprising:
obtaining, using one or more processors, a first plurality of 2D medical images captured at a first time period, performing, using the one or more processors, target tissue detection and target tissue segmentation on each of the first plurality of 2D medical images to produce a set of segmented target tissue images, and generating, using the one or more processors, a first 3D model of detected and segmented target tissue; identifying, using the

Assignees

• Tempus Labs Inc

Inventors

• Yip Stephen

Classifications

• G06T17/00

  Three-dimensional [3D] modelling for computer graphics · CPC title

• G06T2210/41

  Medical · CPC title

• G06T2207/10088

  Magnetic resonance imaging [MRI] · CPC title

• G06T2207/10081

  Computed x-ray tomography [CT] · CPC title

• G06T7/0016 Primary

  involving temporal comparison · CPC title