Machine learning for optimizing ovarian stimulation

Abstract

Systems and methods for implementing machine-learning models for ovarian stimulation is described herein. In some variations, a computer-implemented method may include optimizing an ovarian stimulation process may include receiving patient-specific data associated with a patient, and predicting an egg outcome for the patient for each of a plurality of treatment options for an ovarian stimulation process based on at least one predictive model and the patient-specific data, where the at least one predictive model is trained using prior patient-specific data associated with a plurality of prior patients.

1 - 30 . (canceled) 
     
     
         31 . A method for optimizing an ovarian stimulation process for a patient, the method comprising:
 generating, via a processor, a predictive dose response curve for a patient based on one or more predictive models having received patient data comprising data associated with the patient, wherein the predictive dose response curve provides a predicted egg outcome for each of a plurality of candidate doses of ovarian stimulation medication for the patient, and wherein the ovarian stimulation medication is configured to promote follicle growth in the patient; and   administering a dose of ovarian stimulation medication to the patient based on the predictive dose response curve.   
     
     
         32 . The method of  claim 31 , wherein administering the dose of the ovarian stimulation medication comprises identifying the dose based on a shape of the predictive dose response curve. 
     
     
         33 . The method of  claim 32 , wherein the dose is equal to or less than a dose corresponding to a peak of the predictive dose response curve. 
     
     
         34 . The method of  claim 33 , wherein the peak of the predictive dose response curve is associated with a maximum predicted egg outcome for the patient. 
     
     
         35 . The method of  claim 31 , wherein the ovarian stimulation medication comprises one or both of follicle stimulating hormone (FSH) and luteinizing hormone (LH). 
     
     
         36 . The method of  claim 31 , wherein the predicted egg outcome comprises one or more of number of eggs retrieved, number of mature eggs, maturity yield, number of post-mature eggs, number of fertilized eggs, number of embryos, number of blastocysts, number of usable blastocysts, and number of euploid blastocysts 
     
     
         37 . The method of  claim 36 , wherein the predicted egg outcome comprises one or more of the number of eggs retrieved and the number of mature eggs. 
     
     
         38 . The method of  claim 31 , wherein the patient data comprises one or more of age, body mass index, ethnicity, diagnosis of infertility, prior pregnancy history, and prior birth history. 
     
     
         39 . The method of  claim 38 , wherein the patient data further comprises one or more of measurements of estradiol (E2), measurements of FSH, measurements of LH, measurements of progesterone (P4), measurements of anti-mullerian hormone (AMH), and measurements of antral follicle count (AFC). 
     
     
         40 . The method of  claim 38 , wherein the patient data further comprises one or more of data retrieved during ovarian stimulation, number of eggs retrieved, number of mature eggs retrieved, number of successfully fertilized eggs, number of blastocysts, number of usable blastocysts, pregnancy outcome, and live birth outcome. 
     
     
         41 . The method of  claim 38 , wherein the patient data further comprises one or more of a type of medication, a type of hormonal trigger injection to cause follicle maturation in the patient, and a number of IVF cycles associated with the patient. 
     
     
         42 . The method of  claim 31 , wherein generating the predictive dose response curve comprises generating a plurality of predictive does response curves. 
     
     
         43 . The method of  claim 42 , wherein generating a plurality of predictive dose response curves comprises generating a first predictive dose response curve and generating a second predictive dose response curve based on the first curve. 
     
     
         44 . The method of  claim 31 , wherein the plurality of candidate doses of ovarian stimulation medication comprises a plurality of starting doses of ovarian stimulation medication, a plurality of total doses of ovarian stimulation medication, a plurality of daily doses of ovarian stimulation medication, or combinations thereof. 
     
     
         45 . The method of  claim 31 , wherein the plurality of candidate doses of ovarian stimulation medication comprises a first candidate dose of ovarian stimulation medication and a second candidate dose of ovarian stimulation medication, and wherein the second candidate dose is greater than the first candidate dose. 
     
     
         46 . The method of  claim 31 , wherein the plurality of candidate doses comprises a first candidate dose having a first ratio of a plurality of ovarian stimulation medications and a second candidate dose having a second ratio of the plurality of ovarian stimulation medications, and wherein the first ratio is different than the second ratio. 
     
     
         47 . The method of  claim 31 , wherein the one or more predictive models are trained using prior patient data, the prior patient data comprising one or more of a baseline variable, information related to one or more prior IVF treatments, and a treatment variable for at least one prior patient of the plurality of prior patients. 
     
     
         48 . The method of  claim 47 , wherein the baseline variable comprises one or more of age, BMI, ethnicity, diagnosis of infertility, prior pregnancy history, prior birth history, measurements of E2, measurements of FSH, measurements of LH, measurements of P4, measurements of AMH, and measurements of AFC. 
     
     
         49 . The method of  claim 47 , wherein the information related to one or more prior IVF treatments comprises one or more of data retrieved during ovarian stimulation, number of eggs retrieved, number of mature eggs, number of successfully fertilized eggs, number of blastocysts, number of usable blastocysts, pregnancy outcome, and live birth outcome. 
     
     
         50 . The method of  claim 47 , wherein the treatment variable comprises one or more of a type of medication, a type of hormonal trigger injection to cause follicle maturation in the patient, and a number of IVF cycles associated with the patient. 
     
     
         51 . The method of  claim 47  further comprising using a similarity matching technique to identify a set of prior patients similar to the patient based on the prior patient data, wherein the predictive dose response curve is generated based on similar prior patient data comprising data associated with the set of similar prior patients. 
     
     
         52 . The method of  claim 51  wherein the plurality of candidate doses comprises doses of ovarian stimulation medication administered to at least one similar prior patient of the set of similar prior patients. 
     
     
         53 . The method of  claim 47 , wherein the set of similar prior patients comprises 100 similar prior patients. 
     
     
         54 . The method of  claim 53 , wherein the similarity matching technique comprises a K-nearest neighbors technique. 
     
     
         55 . The method of  claim 31 , wherein the one or more predictive models comprise one or more of a linear model and a neural network. 
     
     
         56 . The method of  claim 31 , wherein the patient is associated with a group of patients of a clinic, the method further comprising predicting a hormonal trigger day for each patient within the group of patients. 
     
     
         57 . The method of  claim 56  further comprising predicting medical staffing needs for one or both of egg retrievals and embryo biopsies on a future day based on the predicted trigger days. 
     
     
         58 . A method for optimizing an ovarian stimulation process, the method comprising:
 training one or more predictive models having received data for a patient using similar prior patient data associated with a set of prior patients similar to the patient, wherein the prior patient data comprises one or more of a baseline variable, information related to one or more prior IVF treatments, and a treatment variable for at least one similar prior patient within the set of similar prior patients; and   generating, via a processor, a predictive dose response curve that provides a predicted egg outcome for each of a plurality of candidate doses of an ovarian stimulation medication for the patient, wherein the predictive dose response curve is generated based on data associated with the set of similar prior patients, and wherein the ovarian stimulation medication is configured to promote follicle growth in the patient.   
     
     
         59 . The method of  claim 58 , wherein the baseline variable comprises one or more of age, BMI, ethnicity, diagnosis of infertility, prior pregnancy history, prior birth history, measurements of E2, measurements of FSH, measurements of LH, measurements of P4, measurements of AMH, and measurements of AFC. 
     
     
         60 . The method of  claim 58 , wherein the information related to one or more prior IVF treatments comprises one or more of data retrieved during ovarian stimulation, number of eggs retrieved, number of mature eggs, number of successfully fertilized eggs, number of blastocysts, number of usable blastocysts, pregnancy outcome, and live birth outcome. 
     
     
         61 . The method of  claim 58 , wherein the treatment variable comprises one or more of a type of medication, a type of hormonal trigger injection to cause follicle maturation in the patient, and a number of IVF cycles associated with the patient. 
     
     
         62 . The method of  claim 58 , wherein the predicted egg outcome comprises one or more of number of eggs retrieved, number of mature eggs, maturity yield, number of post-mature eggs, number of fertilized eggs, number of embryos, number of blastocysts, number of usable blastocysts, and number of euploid blastocysts. 
     
     
         63 . The method of  claim 58 , wherein the ovarian stimulation medication comprises one or both of FSH and LH. 
     
     
         64 . The method of  claim 58 , wherein the plurality of candidate doses comprises doses of ovarian stimulation medication administered to at least one similar prior patient within the set of similar prior patients. 
     
     
         65 . The method of  claim 58 , wherein generating the predictive dose response curve comprises using a similarity matching technique to identify the set of similar prior patients. 
     
     
         66 . The method of  claim 58 , wherein the set of similar prior patients comprises 100 similar prior patients.