Quantitative predictors of tumor severity

Patent metadata

Field	Value
Publication number	US-10339648-B2
Application number	US-201414761688-A
Country	US
Kind code	B2
Filing date	Jan 21, 2014
Priority date	Jan 18, 2013
Publication date	Jul 2, 2019
Grant date	Jul 2, 2019

Abstract

Disclosed are methods for quantitatively predicting the severity of a tumor in a subject. In some embodiments, the methods further comprise selecting a course of therapy for the subject. In some embodiments, the tumor comprises is non-small cell lung cancer.

First claim

What is claimed is:

1. A method for quantitatively predicting the severity of a tumor in a subject, comprising
receiving, using a processor, an anatomical image acquired from a region of interest in a subject containing the tumor, wherein the anatomical image is a computed tomography (CT) image;
segmenting, using the processor, the anatomical image using a segmentation algorithm to define a volume of interest representing the tumor, thereby generating a segmented image of the tumor;
extracting, using the processor, one or more image features from the segmented image of the tumor, wherein the one or more image features are selected from the group consisting of entropy or a combination of convexity and entropy;
generating a quantitative score for the one or more image features, using the processor, wherein the quantitative score for the one or more image features is associated with tumor severity, and wherein the quantitative score for entropy is a ratio of the average entropy in a core region of the segmented image of the tumor to the average entropy in a boundary region of the segmented image of the tumor; and
presenting, using the processor, the quantitative score to a user for selecting a course of therapy for the subject based on the quantitative score for the one or more image features.

2. The method of claim 1 , wherein the one or more image features is entropy.

3. The method of claim 1 , wherein the ratio is calculated by an algorithm that performs steps comprising:
a) subdividing the segmented image of the tumor into the core region and the boundary region;
b) calculating an entropy coefficient for each pixel in the core region and the boundary region;
c) averaging the entropy coefficients for the core region and for the boundary region; and
d) comparing the average entropy coefficient of the core region to the average entropy coefficient of the boundary region to obtain the entropy ratio.

4. The method of claim 1 , wherein the one or more image features further comprises convexity.

5. The method of claim 4 , wherein the convexity is calculated by an algorithm that performs steps comprising:
a) identifying a border of the segmented image of the tumor;
b) creating a mask of the segmented image of the tumor;
c) calculating the area of the mask of the segmented image of the tumor;
d) creating a convex hull of the border of the segmented image of the tumor;
e) calculating the area of the convex hull; and
f) comparing the area of the mask to the area of the convex hull to calculate convexity.

6. The method of claim 5 , wherein the algorithm for convexity further comprises correcting for overlap with a structure comprising:
a) identifying a border of a structure;
b) comparing the border of the structure with the border of the segmented image of the tumor to determine overlap between the segmented image of the tumor and the structure; and
c) eliminating overlapping regions from the area of the mask of the segmented image of the tumor and the area of the convex hull used to calculate convexity.

7. The method of claim 1 , wherein the tumor comprises non-small cell lung cancer.

8. The method of claim 7 , wherein the tumor is a lung adenocarcinoma.

9. The method of claim 1 , wherein the method is combined with a tumor node metastasis (TNM) staging system to predict the severity of the tumor.

10. The method of claim 1 , wherein the anatomical image comprises a stack of two-dimensional image slices of the region of interest.

11. The method of claim 10 , wherein the method further comprises rendering the two-dimensional image slices into a three-dimensional image.

12. The method of claim 10 , wherein the wherein the quantitative score for convexity is calculated as the average convexity for each two-dimensional image slice.

13. The method of claim 1 , further comprising the step of acquiring the anatomical image of the region of interest in the subject containing the tumor.

14. The method of claim 1 , further comprising selecting a course of therapy for the subject based on the quantitative score for the one or more image features.

15. The method of claim 14 , wherein the quantitative score for convexity is an indication of either an irregular or a convex tumor shape, wherein the method comprises selecting adjuvant therapy for the subject if the quantitative score for convexity indicates an irregular tumor shape.

16. The method of claim 14 , wherein the quantitative score for entropy is an indication of tumor density, wherein the method comprises selecting adjuvant therapy for the subject if the quantitative score for entropy indicates a high tumor density.

17. A method of selecting a course of therapy for a subject based on the severity of a tumor in the subject, the method comprising:
receiving a quantitative score for one or more image features extracted from a segmented image of the tumor, wherein:
the quantitative score for the one or more image features is associated with tumor severity and the tumor comprises non-small cell lung cancer,
the one or more image features are selected from the group consisting of entropy or a combination of convexity and entropy,
the quantitative score for entropy is an indication of tumor density and is a ratio of the average entropy in a core region of the segmented image of the tumor to the average entropy in a boundary region of the segmented image of the tumor, and
the quantitative score for convexity is an indication of either an irregular tumor shape or a convex tumor shape; and

selecting a course of therapy for the subject based on the quantitative score for the one or more image features, wherein:
when the quantitative score for entropy indicates a high tumor density, the method comprises selecting surgery and post-surgery adjuvant therapy for the subject,
when the quantitative scope for entropy does not indicate a high tumor density, the method comprises selecting surgery for the subject,
when the quantitative score for convexity indicates an irregular tumor shape, the method comprises selecting surgery and post-surgery adjuvant therapy for the subject, and
when the quantitative score for convexity indicates a convex tumor shape, the method comprises selecting surgery for the subject.

Assignees

• H Lee Moffitt Cancer Ct & Res

Inventors

• Stringfield Olya
• Gillies Robert J

Classifications

• G06T5/50

  using two or more images, e.g. averaging or subtraction · CPC title

• G06T7/64

  of convexity or concavity · CPC title

• G06T2207/30064

  Lung nodule · CPC title

• G06T2207/20216

  Image averaging · CPC title

• G06T2207/30096

  Tumor; Lesion · CPC title