US 12,488,899 B2
Systems and methods for determining breast cancer prognosis and associated features
Charlie Saillard, Paris (FR); Benoit Schmauch, Paris (FR); Victor Aubert, Paris (FR); Aurelie Kamoun, Paris (FR); Magali Lacroix-Triki, Paris (FR); Ingrid Garberis, Paris (FR); Damien Drubay, Paris (FR); and Fabrice André, Paris (FR)
Assigned to OWKIN FRANCE SAS, Paris (FR); and OWKIN, INC., New York, NY (US)
Filed by OWKIN, INC., New York, NY (US); and OWKIN FRANCE SAS, Paris (FR)
Filed on Mar. 28, 2023, as Appl. No. 18/127,566.
Application 18/127,566 is a continuation of application No. 18/186,111, previously published as PCT/US2022/043692, filed on Sep. 15, 2022.
Prior Publication US 2023/0282362 A1, Sep. 7, 2023
Int. Cl. G16H 50/30 (2018.01); G06N 3/0464 (2023.01); G06N 3/0895 (2023.01); G06V 10/82 (2022.01); G16H 10/40 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G16H 50/70 (2018.01)
CPC G16H 50/30 (2018.01) [G06N 3/0464 (2023.01); G06N 3/0895 (2023.01); G06V 10/82 (2022.01); G16H 10/40 (2018.01); G16H 30/40 (2018.01); G16H 50/20 (2018.01); G16H 50/70 (2018.01)] 29 Claims
 
1. A computer-implemented method of predicting the likelihood that a subject having breast cancer will experience a relapse following treatment for identifying a targeted therapy option for breast cancer in the subject, said method comprising:
obtaining a digital image of a histologic section of a breast cancer sample derived from the subject;
obtaining one or more subject attributes derived from the subject;
computing an artificial intelligence (AI) risk score using a machine learning model, the machine learning model having been trained by processing a plurality of training images to predict a risk of relapse and the AI risk score being a first risk score for breast cancer;
computing a clinical risk score using a clinical model and the one or more subject attributes, the clinical model having been trained using one or more subject training attributes from different subjects and the clinical risk score being a second risk score for breast cancer;
stratifying a high risk group and a low risk group across the different subjects for respective clinical risk scores as output by the clinical model, wherein the high risk group has a higher likelihood for relapse of breast cancer than the low risk group;
computing a final risk score for the subject from the AI risk score and the clinical risk score by taking a weighted average between the AI risk score and the clinical risk score;
detecting that the AI risk score of the subject is within a range of the high risk group;
providing a diagnosis or prognosis based on the high risk of relapse for the subject comprising a personalized medicine or targeted therapy to treat the subject; and
administering a treatment to the subject to treat breast cancer based on the diagnosis or prognosis, wherein the treatment comprises administering a hormone therapy drug or a HER2 protein targeting drug.