Artificial intelligence prediction of prostate cancer outcomes
One example method for AI prediction of prostate cancer outcomes involves receiving an image of prostate tissue; assigning Gleason pattern values to one or more regions within the image using an artificial intelligence Gleason grading model, the model trained to identify Gleason patterns on a patch-by-patch basis in a prostate tissue image; determining relative areal proportions of the Gleason patterns within the image; assigning at least one of a risk score or risk group value to the image based on the determined relative areal proportions; and outputting at least one of the risk score or the risk group value.
1. A computer-implemented method of assessing a prognosis of a prostate cancer patient, comprising the steps of:
receiving an image of prostate tissue;
assigning Gleason pattern values to one or more regions within the image using an artificial intelligence Gleason grading model, the model trained to identify Gleason patterns on a patch-by-patch basis in a prostate tissue image;
determining relative areal proportions of the Gleason patterns within the image;
generating a continuous risk score for the entire image based on the determined relative areal proportions;
assigning at least one of the continuous risk score or risk group value, the risk group value based on the continuous risk score; and
outputting at least one of the continuous risk score or the risk group value.
2. The method of claim 1 , wherein the assigning at least one of the continuous risk score or the risk group value comprises using a Cox proportional hazard regression model fitted to a set of Gleason patterns or proportions thereof assigned to a multitude of images in a data set by the artificial intelligence Gleason grading model and associated survival data associated with the data set.
3. The method of claim 1 , wherein the image of the prostate tissue comprises an image of a specimen obtained from a prostatectomy or a needle core biopsy.
4. The method of claim 1 , further comprising presenting at least one of the risk score or the risk group value on a display of a workstation.
5. The method of claim 4 , further comprising receiving a Gleason Grade for the image from the workstation.
6. The method of claim 5 , further comprising combining the risk group value with a Gleason Grade.
7. The method of claim 6 , wherein combining the risk group value with the Gleason Grade comprises determining an average of the risk group value and the Gleason Grade.
8. The method of claim 1 , wherein the risk group values comprise integers from 1 to 5, inclusive, and wherein a survival probability of each of the risk group values approximately correspond to a survival probability of prostate cancer patients having Gleason Grades of integers 1-5 inclusive, respectively.
9. The method of claim 1 , further comprising using a model trained to distinguish prostate tissue from non-prostate tissue, wherein receiving the image comprises receiving only prostate tissue regions of the image from the model.
10. A system for assessing a prognosis of a prostate cancer patient, comprising:
a non-transitory computer-readable medium; and
one or more processors communicatively coupled to the non-transitory computer-readable medium, the one or more processors configured to execute processor-executable instructions stored in the non-transitory computer-readable medium to:
receive an image of prostate tissue;
assign Gleason pattern values to one or more regions within the image using an artificial intelligence Gleason grading model, the model trained to identify Gleason patterns on a patch-by-patch basis in a prostate tissue image;
determine relative areal proportions of the Gleason patterns within the image;
generate a continuous risk score for the entire image based on the determined relative areal proportions;
assign at least one of the continuous risk score or risk group value, the risk group value based on the continuous risk score; and
output at least one of the risk score or the risk group value.
11. The system of claim 10 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to use a Cox proportional hazard regression model fitted to a set of Gleason patterns or proportions thereof assigned to a multitude of images in a data set by the artificial intelligence Gleason grading model and associated survival data associated with the data set.
12. The system of claim 10 , wherein the image of the prostate tissue comprises an image of a specimen obtained from a prostatectomy or a needle core biopsy.
13. The system of claim 10 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to present at least one of the risk score or the risk group value on a display of a workstation.
14. The system of claim 13 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to receive a Gleason Grade for the image from the workstation.
15. The system of claim 14 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to combine the risk group value with a Gleason Grade.
16. The system of claim 15 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to determine an average of the risk group value and the Gleason Grade.
17. The system of claim 10 , wherein the risk group values comprise integers from 1 to 5, inclusive, and wherein a survival probability of each of the risk group values approximately corresponds to a survival probability of prostate cancer patients having Gleason Grades of integers 1-5 inclusive, respectively.
18. The system of claim 10 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to use a model trained to distinguish prostate tissue from non-prostate tissue, wherein receiving the image comprises receiving only prostate tissue regions of the image from the model.
19. A non-transitory computer-readable medium comprising processor-executable instructions configured to cause one or more processors to:
receive an image of prostate tissue;
assign Gleason pattern values to one or more regions within the image using an artificial intelligence Gleason grading model, the model trained to identify Gleason patterns on a patch-by-patch basis in a prostate tissue image;
determine relative areal proportions of the Gleason patterns within the image;
generate a continuous risk score for the entire image based on the determined relative areal proportions;
assign at least one of the continuous risk score or risk group value, the risk group value based on the continuous risk score; and
output at least one of the risk score or the risk group value.
20. The non-transitory computer-readable medium of claim 19 , wherein the one or more processors are configured to execute further processor-executable instructions stored in the non-transitory computer-readable medium to use a Cox proportional hazard regression model fitted to a set of Gleason patterns or proportions thereof assigned to a multitude of images in a data set by the artificial intelligence Gleason grading model and associated survival data associated with the data set.