Methods and systems for determining optimal decision time related to embryonic implantation
Methods and systems are for improvements to in-vitro fertilization using morpho-kinetic signatures. These improvements are achieved by analyzing a series of images of a developing embryo (e.g., time-lapse images) as opposed to a single static image. For example, due to the difficulty in identifying clear distinctions between morphological states based on static images, as well as the unpredictability of morpho-kinetic development of an embryo, the system analyzes the development of an embryo as a whole over a given time frame (e.g., fertilization to blastulation), which provides a better prediction of the viability of a given embryo. The analysis may take the form of a morpho-kinetic signature, which itself may be used to determine an optimal time to transfer and/or implant an embryo into a patient.
1 . A method of determining an optimal decision time related to embryonic implantation, the method comprising:
receiving, using control circuitry, a first morpho-kinetic signature of a first embryo, wherein the first morpho-kinetic signature is a representation of morpho-kinetic events in the first embryo as a function of time;
labeling, using the control circuitry, the first morpho-kinetic signature with a known time of peak implantation;
training, using the control circuitry, an artificial neural network to detect the known time of peak implantation based on the first morpho-kinetic signature;
receiving, using the control circuitry, a second morpho-kinetic signature of a second embryo with an unknown time of peak implantation, wherein the second morpho-kinetic signature is a representation of morpho-kinetic events in the second embryo as a function of time;
inputting, using the control circuitry, the second morpho-kinetic signature into the trained artificial neural network; and
receiving, using the control circuitry, a prediction from the trained artificial neural network that the second morpho-kinetic signature corresponds to the known time of peak implantation.
2 . The method of claim 1 , wherein the known time of peak implantation corresponds to a time when a probability of success of implantation is highest.
3 . The method of claim 1 , further comprising labeling the first morpho-kinetic signature with a known endometrial thickness, wherein the artificial neural network to is trained to detect the known time of peak implantation based on the first morpho-kinetic signature and the known endometrial thickness.
4 . The method of claim 1 , further comprising labeling the first morpho-kinetic signature with known clinical data, wherein the artificial neural network to is trained to detect the known time of peak implantation based on the first morpho-kinetic signature and the known clinical data.
5 . The method of claim 1 , further comprising:
receiving a first output from an initial artificial neural network indicating that the first embryo has a first classification at a first time point;
receiving a second output from the initial artificial neural network indicating the first embryo has a second classification at a second time point; and
aggregating the first output and second output to generate the first morpho-kinetic signature.
6 . The method of claim 1 , further comprising:
determining a first probability of successful implantation for the second embryo based on the prediction;
determining a second probability of successful implantation for a third embryo; and
determining a composite probability of success for implanting both the second embryo and the third embryo based on the first probability and the second probability.
7 . The method of claim 1 , further comprising:
determining a first probability of successful implantation for the second embryo based on the prediction;
determining a second probability of successful implantation for a third embryo;
normalizing the first probability and the second probability; and
selecting the second embryo based on the normalized first probability being higher than the normalized second probability.
8 . The method of claim 1 , wherein each morpho-kinetic event of the morpho-kinetic events is represented as a float value in a vector.
9 . The method of claim 1 , wherein the first morpho-kinetic signature is based on a series of time-lapse images of the morpho-kinetic events in the first embryo.
10 . The method of claim 9 , wherein the time of peak implantation corresponds to a time when a probability of success of implantation is highest.
11 . A method of determining an optimal decision time related to embryonic implantation, the method comprising:
receiving, using control circuitry, a morpho-kinetic signature of an embryo with an unknown time of peak implantation, wherein the morpho-kinetic signature is a representation of morpho-kinetic events in the embryo as a function of time;
inputting, using the control circuitry, the morpho-kinetic signature into an artificial neural network, wherein the artificial neural network is trained to detect to a time of peak implantation for the morpho-kinetic signature based on a set of training data comprising morpho-kinetic signatures labeled with known times of peak implantation;
receiving, using the control circuitry, a prediction, from the artificial neural network, for the time of peak implantation for the morpho-kinetic signature.
12 . The method of claim 11 , further comprising controlling a time-lapse imaging system to generate time-lapse images of the embryo and the morpho-kinetic signature is contained in the time-lapse images.
13 . The method of claim 11 , further comprising controlling an imaging system to generate z-stack images of the embryo and the morpho-kinetic signature is contained in the z-stack images.
14 . The method of claim 11 , further comprising controlling an imaging system to generate images of the embryo at multiple angles and the morpho-kinetic signature is contained in the images.
15 . A non-transitory computer-readable media comprising instructions that, when executed by one or more processors, cause operations comprising:
receiving a first morpho-kinetic signature of a first embryo, wherein the first morpho-kinetic signature is a representation of morpho-kinetic events in the first embryo as a function of time;
labeling the first morpho-kinetic signature with a known time of peak implantation;
training an artificial neural network to detect the known time of peak implantation based on the first morpho-kinetic signature;
receiving a second morpho-kinetic signature of a second embryo with an unknown time of peak implantation, wherein the second morpho-kinetic signature is a representation of morpho-kinetic events in the second embryo as a function of time;
inputting the second morpho-kinetic signature into the trained artificial neural network; and
receiving a prediction from the trained artificial neural network that the second morpho-kinetic signature corresponds to the known time of peak implantation.
16 . The non-transitory computer-readable media of claim 15 , wherein the known time of peak implantation corresponds to a time when a probability of success of implantation is highest.
17 . The non-transitory computer-readable media of claim 15 , wherein the instructions further cause operations comprising labeling the first morpho-kinetic signature with a known endometrial thickness, wherein the artificial neural network to is trained to detect the known time of peak implantation based on the first morpho-kinetic signature and the known endometrial thickness.
18 . The non-transitory computer-readable media of claim 15 , wherein the instructions further cause operations comprising labeling the first morpho-kinetic signature with known clinical data, wherein the artificial neural network to is trained to detect the known time of peak implantation based on the first morpho-kinetic signature and the known clinical data.
19 . The non-transitory computer-readable media of claim 15 , wherein the instructions further cause operations comprising:
receiving a first output from an initial artificial neural network indicating that the first embryo has a first classification at a first time point;
receiving a second output from the initial artificial neural network indicating the first embryo has a second classification at a second time point; and
aggregating the first output and second output to generate the first morpho-kinetic signature.
20 . The non-transitory computer-readable media of claim 15 , wherein the instructions further cause operations comprising:
determining a first probability of successful implantation for the second embryo based on the prediction;
determining a second probability of successful implantation for a third embryo; and
determining a composite probability of success for implanting both the second embryo and the third embryo based on the first probability and the second probability.
21 . The non-transitory computer-readable media of claim 15 , wherein the instructions further cause operations comprising:
determining a first probability of successful implantation for the second embryo based on the prediction;
determining a second probability of successful implantation for a third embryo;
normalizing the first probability and the second probability; and
selecting the second embryo based on the normalized first probability being higher than the normalized second probability.
22 . The non-transitory computer-readable media of claim 15 , wherein each morpho-kinetic event of the morpho-kinetic events is represented as a float value in a vector.
23 . The non-transitory computer-readable media of claim 15 , wherein the first morpho-kinetic signature is based on a series of time-lapse images of the morpho-kinetic events in the first embryo.
24 . The non-transitory computer-readable media of claim 15 , the operations further comprising:
generating pixel arrays from color images of the first morpho-kinetic signature, the pixel arrays comprising color components of the color images;
converting the pixel arrays from the color components to a grayscale representation; and
training the artificial neural network further based on the grayscale representation of the morpho-kinetic events.
25 . The non-transitory computer-readable media of claim 15 , the operations further comprising:
generating pixel arrays from grayscale images of the first morpho-kinetic signature;
converting the pixel arrays from grayscale pixels to color a color representation; and
training the artificial neural network further based on the color representation of the morpho-kinetic events.
26 . The non-transitory computer-readable media of claim 15 , the operations further comprising:
generating pixel arrays from images of the first morpho-kinetic signature;
adding an additional vector that includes clinical or demographic data related to the first embryo; and
training the artificial neural network further based on the pixel arrays of the morpho-kinetic events and also based on the additional vector.