FIELD OF THE INVENTION

This invention relates generally to assisted reproductive technology and in particular to prediction of a patient’s date of ovulation.

BACKGROUND OF THE INVENTION

In vitro fertilization (IVF) relates generally to medical techniques whereby an oocyte is fertilized outside a patient’s body and the fertilized embryo is transplanted into the patient’s uterus. The IVF process involves ovarian hormonal stimulation with blood tests and ultrasound monitoring, oocyte pick up from the patient’s ovaries and fertilization of the oocyte with the partner’s or a donor’s sperm in a culture medium in a laboratory. After the fertilized oocyte undergoes embryo culture for 2-6 days, it is then transferred into the uterus, with the intention of establishing a successful pregnancy. In many cases, surpluses of good quality embryos remain after the process. These embryos are frozen for future use. These embryos can be returned in a preparatory cycle supported by hormonal therapy or in a cycle based on the detection of natural ovulation

Ovulation is assumed to occur fourteen days before the start of the next menstrual cycle, which by definition starts at the first day of menstruation. After menstruation, the uterine lining begins to prepare for the possibility of a pregnancy in the current cycle becoming thicker and enriched in blood and nutrients. At the same time, follicles develop in the ovary, usually one of which develops towards ovulation. Around day 14 of the cycle, that leading follicle ovulates and releases an oocyte. The released oocyte begins its journey down the fallopian tubes to the uterus. Sperm cells in the fallopian tube present at this time, can fertilize the oocyte and the fertilized oocyte will travel to the uterus and attempt to implant in the uterine wall. If the oocyte was not fertilized or implantation does not occur, hormonal changes signal the uterus to prepare to shed its lining, and the oocyte breaks down and is shed along with lining thus commencing the next menstrual cycle. The above description is true for unassisted reproduction. However, during IVF, the fertilization occurs outside the body and the fertilized oocyte is transplanted in the uterus usually 3 or 5 days after fertilization. It is essential that the uterine lining will be properly prepared and that the body will have secreted the necessary hormones to render successful pregnancy viable. IVF requires the extraction of an oocyte that is fertilized outside the body and may be re-implanted in the same cycle or frozen for future use. Extraction of oocytes requires a mildly invasive surgical procedure and since implantation does not always succeed, may require extraction of additional oocytes. In order to increase success rates, multiple oocytes should be retrieved (the optimal number is about 15). All the extracted oocytes undergo a fertilization process, 1-2 of them are transferred to the uterus that month, and the others, depending on their quality as mentioned, are frozen for future use. This technique is known as frozen embryo transfer or FET.

When an embryo is transferred in the same cycle during which the oocyte was extracted, the uterine lining is ready and synchronized to embryos developmental stage. However, when a frozen embryo is transferred in a subsequent cycle, the uterine lining must be synchronized with the developmental stage of the embryo to allow successful implantation. One way to achieve this is to prepare the uterine lining using hormones and to transfer the embryo in an artificial cycle. However, in a natural cycle the embryo must be transferred after the same number of days following ovulation for which the embryo was grown prior to freezing to ensure that the uterine lining is properly developed to allow implantation. In practice, it is standard procedure to grow the embryos for 3 or 5 days prior to freezing and therefore to transplant the embryo precisely on the third or fifth day after ovulation of a subsequent cycle.

There are essentially two kinds of FET-IVF cycles: hormonal support cycles and “natural” cycles. The hormonally supported cycle may use the Estradiol hormone to build proper endometrial lining or uses drugs or hormones to time ovulation and thus allows easy control of the day of transfer. For this reason, the technique is popular, particularly when the patient is anovulatory. However, today, transfer of frozen embryos in a cycle based on the identification of the natural ovulation time is considered the preferred method of treatment, both due to better success rates and in terms of a lower rate of pregnancy complications. FET during the natural cycle, to which the present invention is directed, requires that the ovulation date be accurately predicted during the treatment cycle, without the use of drugs and using only blood and ultrasound tests. That way the embryo can be transferred the same number of days following ovulation as it was grown prior to freezing, and the chance for implantation to be successful is maximized

There is a vast literature relating to the prediction of a woman’s date of ovulation and more particularly her so-called fertile window, which is usually defined as the day of ovulation and the five days beforehand. The patent literature discloses many devices that include sensors for measuring body metrics, such as basal body temperature, and a processor responsive to the measured body metrics for predicting ovulation based on prior dates of ovulation and measurements for the same user. The literature does not appear to relate to the highly accurate prediction of ovulation date required for natural cycle FET.

WO2021037724 discloses a method and system for determining a time interval of fertility for a female with irregular menstrual cycles. A wearable device includes sensors for measuring physiological data; and a processor receives the measurements and determines an estimated time to ovulation, by use of a machine learning model and the physiological data; determining the time interval of fertility using pre-determined time thresholds; and generating a message for the female indicating the time interval of fertility. In an embodiment, determining the estimated time to ovulation comprises the processor training the machine learning model to determine the estimated time to ovulation using machine learning and a training dataset of physiological training data of a large group of females. The machine learning model can be trained prior to determining the estimated time to ovulation.

US20190103175 discloses an ovarian characteristic measurement method and system. A data interface is configured to receive a set of reproductive health data of a patient, which is processed using an ovarian characteristic data structure stored in memory. The data structure is generated from an ovarian characteristic data model that defines the ovarian characteristic. A value of the ovarian characteristic of the patient is determined based on the processing of the set of reproductive health data.

US20200395117 discloses adaptive image processing, image analysis, pattern recognition, and time-to-event prediction in various imaging modalities associated with assisted reproductive technology. The subject image is processed according to various computer vision techniques for object detection, recognition, annotation, segmentation, and classification of reproductive anatomy, such as follicles, ovaries and the uterus. An image processing framework may also analyze secondary data along with subject image data to analyze time-to-event progression of the subject image.

US20220036041 discloses a system and method of biological testing and deep learning to predict fertility based on ferning patterns and detecting white blood cells in cervical mucous samples.

WO/2022/098737 discloses a non-transitory computer-readable storage medium comprising instructions that when executed cause processor circuitry to generate a longitudinal dataset of a set of features from tracked physical measurements of a user received from sensor circuitry, apply at least one machine learning model to the longitudinal dataset of the set of features to identify a pattern within the set of features indicative of a probability of the user being in a state of menopause, predict a current state of menopause for the user based on the identified pattern, and communicate a data message indicative of the current state of menopause for the user.

US20130137940 discloses a fertility monitoring apparatus based on measurement of body temperature and having a central fertility platform that includes a database of fertility information from multiple users. Data from user accounts such as user’s age, cycle length, diagnosis of fertility, reproductive health information are stored in the database. The platform both provides feedback to the user and learns from the data collected to improve upon or self-modify calculations using machine learning, artificial intelligence, and data mining approaches.

US20200000441 discloses use of health information for a woman to predict timing of events related to the woman’s menstrual cycle. Historical cycle information for a woman can be used to predict upcoming cycle events, such as the start and stop of menstruation. To improve the accuracy of those predictions, one or more health metrics are monitored for the woman that can be correlated with the menstrual cycle. The metrics are monitored over time to determine patterns that can be correlated with menstrual cycle. This information can then be used to update the predictive model, as well as to update individual event predictions. Information about the predictions, and updates to the predictions, can be surfaced accordingly. Various other types of information can be utilized as well that may impact the timing of a cycle or relevance of different body data values such as variations in diet, stress, weight, body fat percentage, body mass index (BMI), medication, or other such factors that can be accounted for as well. Reference is also made to the following scientific publications:

Genuis, S.J., “High-tech family planning: reproductive regulation through computerized fertility monitoring.'' European Journal of Obstetrics & Gynecology and Reproductive Biology 153; 124-130, 2010. This paper relates to determination of the fertile window as a means for family planning and describes hormone-based computerized fertility monitors used by women for family planning, both in isolation and in conjunction with other methods of fertility monitoring. Hormone-based family planning using fertility monitors generally relies on the measurement of urinary E3G and LH to consistently delineate the fertile period.

Karen Clark el al. (SMU Scholar) “Open Cycle: Forecasting Ovulation for Family Planning’ (htps.7/schol ar. smu.edu/datasci encereview/vol 1/iss 1 /2/) proposes the use of machine learning methodologies to forecast ovulation date within one day of ovulation with a purported accuracy of 88% for use in family planning. The method is based on an assumption that a rise of 0.2 Fahrenheit in Basal Body Temperature (BBT) indicates ovulation. In fact, the scientific literature does not support this assumption since it is known that BBT can rise for many reasons other than ovulation. Thus, reference is made to Hsiu-Wei Su el al. “Detection of ovulation, a review of currently available methods" appearing in Bioeng Transl Med. 2017 Sep; 2(3): 238-246 (doi: 10.1002/btm2.10058), which lists many other factors affecting BBT, such as fever, alcohol consumption, emotional or physical stress, sleep disturbance, change of room temperature, change of waking time, change of climate, and recent start or discontinuation of birth control pills or anti-pyretics.

Tony Duan, el al. (Corenell University): NGBoost: Natural Gradient Boosting for Probabilistic Prediction (htpg://doi .org/ 1 ,0.48550/arxiv. 1910.0 225) discloses an algorithm for generic probabilistic prediction via gradient boosting. Probabilistic regression models output a full probability distribution over the outcome space, conditional on the covariates. This allows for predictive uncertainty estimation - crucial in applications like healthcare and weather forecasting.

None of these prior art publications appears to relate to accurate prediction of a patient’s next ovulation date for use by medical practitioners in the field of assisted reproduction. Even in this field of medical practice, there are two distinct needs: one requiring much higher accuracy than the other. Thus, for patients who ovulate normally but require artificial insemination, it is sufficient that semen be injected into the fallopian tube any time, but preferably as close to ovulation as possible, within the fertile window since sperm can survive for at least 2-3 days. This allows the patient’s oocyte to be fertilized by viable sperm that may have been injected even prior to or immediately subsequent to ovulation. In such cases, the fertile window provides quite literally a window of opportunity for fertilization to occur and obviates the need for precise determination or prediction of the date of ovulation.

However, this is very far from the case where a frozen embryo is to be transplanted into the patient’s uterus since this must be done in exact synchronism with the menstrual cycle in order to be optimally effective. This means that an oocyte that is removed from the patient on the day of ovulation, fertilized in vitro and then transplanted after, say, a further five days must be transplanted in a uterus that is precisely five days after ovulation. This is not a problem if the fertilized oocyte is transplanted in the same cycle as it was removed since whenever it is transplanted in the same cycle it will always be exactly in synchronism with the date of ovulation.

But an embryo that is frozen and then transplanted in a later (often much later) cycle must be transplanted on the correct day of the cycle to preserve the required synchronization. In other words, using the same example as above, the frozen embryo must be thawed and transplanted on day +5 after ovulation. This can be done for subsequent cycles only if the date of ovulation is known precisely. For the sake of abundant clarity, it should be noted that since the embryo is necessarily transplanted after the ovulation date, “prediction” of ovulation date does not require that the ovulation date be determined in advance of ovulation. It can also be determined shortly after ovulation but obviously in sufficient time to allow synchronized transplantation of the embryo.

Of course, it is well-known that the date of ovulation can be determined by detecting whether the ovaries house a developing oocyte they will later release during ovulation. Developing oocytes are called ‘follicles’ whose presence is established using ultrasound. Specifically, if a large follicle is detected in an initial ultrasound scan but is absent in a subsequent scan, then it can be inferred that ovulation took place between the first and second scan. In order to determine the optimal date for transplanting a frozen embryo, we need to pinpoint the exact date of ovulation. This would require performing multiple ultrasounds, which are not only expensive but require that the patient come to the clinic on successive days, which is highly disruptive. It should also be borne in mind that patients requiring assisted reproduction are invariably tense and vulnerable. Anything that can be done to minimize hospital visits is therefore to be welcomed. Even apart from this, sometimes the ovaries produce two or more oocytes in the same cycle and in such case the detection of a follicle in the ovary does not preclude the possibility that one oocyte has been released. If this happens, accurate detection of the date of ovulation for the current cycle will be impossible and the need to wait a whole cycle to allow for the possibility of successful transplantation of a frozen embryo is stressful for the patient and arduous for the physician.

SUMMARY OF THE INVENTION

It is an object of the present invention to provide a method and system for precise prediction of a patient’s ovulation date.

Within the context of the present description and appended claims the term “precise” means that the ovulation date must be predicted with sufficient accuracy to allow a fertilized embryo taken from an earlier cycle to be synchronously transplanted into the patient’s uterus during a subsequent cycle.

Within the context of the present description and appended claims the terms ‘synchronously transplanted’ and ‘synchronous transplantation’ denote that the fertilized embryo is transplanted in proper synchronization with the patient’s natural menstrual cycle i.e., that it is transplanted on the exact day in the cycle as the oocyte was originally extracted, which is always the ovulation date plus the number of days that elapsed prior to freezing the fertilized oocyte after extraction.

We also reiterate that prediction can also indicate precise determination of the ovulation date after ovulation has occurred and prior to synchronous transplantation of the embryo.

This object is realized in accordance with the invention by a two-part procedure. The first part is a learning method that creates a model for predicting ovulation date based on a large population of patients’ blood results and patient characteristics, while the second part provides a method for predicting ovulation date for a specific patient, using the model which has already been determined. The model can be regarded as a transfer function which receives as input measurements relating to a specific patient and produces as output a predicted next date of ovulation. Thus, in its first aspect there is provided in accordance with the invention a computer-implemented method for generating a model for predicting precise dates of ovulation from blood tests, the method comprising: obtaining patient characteristics and medical data pertaining to a plurality of patients, wherein the medical data includes, for each of the patients, blood test results each having associated timing information indicating the day in the patient’s natural menstrual cycle when the respective blood test was performed and historical data relating to past dates of ovulation for each patient; storing in a computer memory a dataset mapping past dates of ovulation with respective patient characteristics and blood test results and associated timing information for each patient; and applying deep learning/machine learning techniques to the dataset to create a model as a function of the respective patient characteristics and medical data pertaining to the plurality of patients.

In its second aspect, the invention provides a method for predicting a next date of ovulation for a target patient having known patient characteristics and blood test results and associated timing information indicating the day in the patient’s natural menstrual cycle when the respective blood test was performed using such a model, the method comprising: inputting the known blood test results and timing information together with the patient characteristics to the model so as to obtain for the target patient a next ovulation date.

The two procedures are mutually independent. Specifically, while the second procedure cannot be implemented until the model has been derived, once the model exists, the second procedure can be implemented in standalone manner. In this case, the two procedures are carried out sequentially. But they can also be carried out concurrently whereby the learning procedure is executed at the same time as implementing the second aspect either in respect of a different patient or to refine the model by continually gathering new data and applying deep learning techniques to a continually expanding dataset. For example, when a patient undergoes a treatment plan according to the invention on a possibly ongoing basis, new data is constantly being acquired and this can be used to supplement the dataset and improve the accuracy of the model. The invention also provides a treatment regimen and system for monitoring and guiding a patient through the process of medical analysis by taking an initial measurement, using the above-described algorithm to determine either an ovulation date if a sufficiently accurate prediction can be made or, if not, informing the patient when to perform a subsequent measurement. The system includes a patient device for receiving instructions from a fertility clinic and for entering blood test results and optionally ultrasound results, which are conveyed to the fertility clinic or to another medical facility for determination of the ovulation date if a sufficiently accurate prediction can be made or, if not, informing the patient when to perform a subsequent measurement. The treatment regimen is repeated repeatedly as required until an accurate ovulation date is determined, and which leaves adequate time to synchronously transplant an embryo. If the predicted ovulation date is in the future, then of course synchronous transplantation is always possible. If the date of ovulation has already occurred for the current cycle, synchronous transplantation during the current cycle will be possible only if there remain sufficient days since ovulation to factor in the number of days for which the embryo was grown prior to freezing, typically 3 or 5. So, in the case, where the algorithm predicts that ovulation occurred yesterday and the embryo was frozen on the third day after extraction, we still have two days left and synchronous transplantation during the current cycle is possible. There may, however, be rare occasions where the algorithm predicts that ovulation occurred more than three days ago even though the algorithm chooses the next days of testing in order to avoid such a scenario. In this case, synchronous transplantation during the current cycle is no longer possible and the patient will need to repeat the treatment regimen the following month.

Preferably, ongoing execution of the treatment plan is transparent to the physician since, in the absence of information on the patient’s average cycle length, the default day to call the patient for the initial blood test is day 8 of the menstrual cycle, where day 1 is where bleeding first occurs. If the average cycle length (ACL) is known, then the initial blood test is called for ACL-20 days. For a regular 28-day cycle, this is of course the same. In this mode, no intervention is required by the doctor other than to initiate the treatment plan. However, it should be noted that the default day is set only arbitrarily to 8 and, in practice, may be varied and may be even be set by the doctor for each patient or different clinics may adopt their own default day. Alternatively, the doctor can overrule the default initial test date. In either case, subsequent execution is transparent to the physician since the patient can perform a blood test or an ultrasound, if necessary, in an independent medical facility and the acting physician (obstetrician) can login to the fertility clinic or to a central facility in communication therewith to monitor the patient’s tests. Likewise, the fertility clinic or central facility can be configured to send the predicted ovulation date to the physician for coordination with the patient, who will then be invited to visit the fertility clinic for transplantation of the embryo.

BRIEF DESCRIPTION OF THE DRAWINGS

In order to understand the invention and to see how it may be carried out in practice, embodiments will now be described, by way of non-limiting example only, with reference to the accompanying drawings, in which:

Fig- 1 is a schematic diagram showing a system for implementing an embodiment of the invention;

Fig- 2 is a flow diagram showing the principal operations for creating a machine learning model for determining ovulation date in accordance with the invention;

Fig- 3 is a flow diagram showing the principal operations for using the model for determining ovulation date;

Fig. 4 is a flow diagram showing the principal operations for using the model to interact with a patient to obtain patient data and determine ovulation date;

Fig. 5 is a flow diagram showing the principal operations for using the model to interact with a patient to obtain patient data and determine ovulation date in a FET treatment system;

Figs. 6 is a pictorial representation showing a doctor’s smartphone client application prior to starting a treatment plan for a patient;

Figs. 7 and 8 are pictorial representations showing images of a patient’s smartphone client application during successive stages of a FET treatment plan; and

Figs. 9 is a pictorial representation showing the doctor’s smartphone client application after receiving the patient predicted ovulation date. DETAILED DESCRIPTION OF EMBODIMENTS

Fig. 1 is a schematic diagram showing a system 10 for monitoring and guiding a patient through a process of medical analysis prior to frozen embryo transfer in accordance with an embodiment of the invention. Before describing the system in detail, it is noted that this implementation is described as an example that requires accurate prediction of ovulation date, on the basis of which a doctor will decide the optimum day of the patient’s menstrual cycle for effecting frozen embryo transfer (FET). However, this is only one application of the invention, whose principal feature resides in predicting precise dates of ovulation from blood tests. As noted above, the degree of precision is such as to facilitate synchronous transplantation of a previously frozen fertilized embryo, but this does not require that the invention be used only within the context of FET. Nevertheless, the invention is intended for use with patients who have natural menstrual cycles, by which we mean that the menstrual cycle is not controlled artificially using drugs or hormones.

The system 10 comprises a computer 11 monitored by a doctor 12, typically working at a fertility clinic and providing assisted reproductive therapy over the Internet 13 to patients 14 and 15 having personal computer devices such as a smartphone 16 or a laptop 17 or another computer, preferably having a touchscreen interface. The doctor’s computer 11 is coupled to a central computer 18 that may be located in the same fertility clinic or external thereto in another facility and in communication with the doctor’s computer 11 over the Internet 13. Stored in the central computer 18 is a model derived in accordance with an aspect of the invention for predicting precise dates of ovulation from blood tests, and which is then used in combination with patient-specific data to predict the ovulation dates for each of the patients 14 and 15. In an embodiment of the invention relating to patients undergoing FET, the predicted ovulation may be used to determine when each patient should present herself at the fertility clinic for FET. In alternative embodiments, the central computer 18 may be located in the fertility clinic and connected to the doctor’s computer 11 either directly or via a VPN. Alternatively, the model may be stored in the doctor’s computer 11 in which case the central computer 18 may be omitted.

The algorithm according to the invention is implemented in two stages. In a first stage depicted in Fig. 2, a model is derived by obtaining patient characteristics and medical data pertaining to a plurality of patients. Patient characteristics relate to physical characteristics such as age, height, weight, BMI and may include ethnicity. The medical data includes, for each of the patients, blood test results each having associated timing information and historical data relating to past dates of ovulation for each patient. The timing information indicates the day in the patient’s natural menstrual cycle when the respective blood test was performed. For the purpose of description, it is assumed that the cycle commences on the day of menstruation when bleeding first occurs. Optionally, the blood results may be supplemented by ultrasound results containing the size of the dominant follicle taken at known times in the patients’ cycles. The compiled data are stored in a dataset, which maps past dates of ovulation with respective patient characteristics and blood test results for each patient. Deep learning/machine learning techniques are applied to the dataset to create a model of ovulation date as a function of the respective patient characteristics and medical data pertaining to the plurality of patients.

In the first stage the model is trained using historical data from a large population of patients (about 2,000 FET cycles for the ovulation prediction models). The raw data is subject to the following pre-processing operations:

■ Normalization: the data received from different sources are transformed into a common structure that can be treated the same by analysts and algorithms regardless of its source;

■ Cleaning: errors and outliers in the data are located and fixed or removed so the algorithms can train on clean and accurate data so they will produce accurate results. An example of an error would be an obvious typo in a blood test hormone value which reads 0,2 instead of 0.2. This is easily corrected. An example of an outlier would be a measured hormone that is ten times higher than all other data points, we assume that either it is a typing error, or the patient received medications without it being reflected in her medical file. Whatever, the reason, we remove these cycles.

■ The date of ovulation for each FET cycle is calculated using: o The date of the transfer operation performed by the attending physician; o The age in days of the embryo transferred, where the ovulation date is the transfer date minus the embryo age prior to freezing (an embryo that was grown for 5 days will be returned to the womb 5 days after the patient ovulated). It should be noted that in an embodiment of the invention, the historical data used to train the machine learning model used actual FET transfer dates for the sample population. The data were based on records for a sample of over 2,000 anonymous patients for whom physical characteristics, blood test results taken on known days of their menstrual cycles, the day in the cycle on which an embryo was actually transferred and the age of the embryo as defined above. In such a scenario the actual day of ovulation is not known categorically but is inferred from when the doctor elected to perform the FET transfer. For the purpose of training the algorithm, after calculating the ovulation day, both the transfer date and embryo age are no longer relevant and we do not use them, because it is of no further consequence when in that cycle a transfer happened, only when the ovulation itself happened.

However, the model can equally well be obtained for a statistically significant population of patients for whom blood tests and ultrasounds are performed on known days of their cycles to create a dataset containing their physical characteristics and their test results and timestamps. The timestamp for each blood test is the day in the patient’s menstrual cycle when the respective blood test was performed. Their respective dates of ovulation can be determined highly accurately to within one day by repeated ultrasounds thus permitting the model to be trained.

Figs. 3 and 4 are flow diagrams showing the principal operations for using the model for determining ovulation date, it being noted that use of the model is independent of how the model is derived. The patient is informed on what day of the menstrual cycle to do a blood test and optionally obtain ultrasound results containing the size of the dominant follicle. The patient conveys the test results to the fertility clinic which feeds the results and timestamps for the target patient together with physical characteristics into the model to predict ovulation date in the current cycle where possible - or to compute when in the current cycle or even in the next cycle to do another blood test. The call for additional tests is repeated as necessary until precise prediction of ovulation date becomes possible.

In one embodiment, the algorithm uses prior cycles of the same patient only to choose the initial blood test day. As noted above, in some embodiments, this may be set by default to day ACL-20 where ACL is the average cycle length. From that point, the most recent blood test is the best indicator for when ovulation will occur, and patient characteristics (age, weight, height, BMI) are used in addition to the blood test results to predict the ovulation date.

To summarize the algorithm uses the following data to determine ovulation after it has been trained:

■ Patient characteristics (age, weight, height, BMI, ethnicity);

■ Patient previous cycle statistics. In an embodiment reduced to practice, these include average cycle length, which is used to determine an initial blood test day;

■ Patient current cycle data (the two most recent blood tests which include their timings and the hormone values for Progesterone, Estradiol and LH);

■ Optionally ultrasound results showing development of the dominant follicle.

In a further aspect of the invention shown in Fig. 5, the model is used in a diagnostic method for monitoring and guiding a patient through a process of medical analysis prior to frozen embryo transfer, the method comprising:

(a) obtain data pertaining to physical characteristics of the patient;

(b) obtain measurements of the patient’s Progesterone, Estradiol and LH on a known day of the patient’s menstrual cycle;

(c) using a model obtained according to any one of the above-described methods to determine an ovulation date if a sufficiently accurate prediction can be made;

(d) if prediction of a sufficiently accurate ovulation date is impossible, determining when the patient should perform a subsequent blood test and informing the patient accordingly;

(e) repeating (b) to (d) as required until prediction of a sufficiently accurate ovulation date is possible;

(f) if the predicted date of ovulation has already occurred for the current menstrual cycle, and there remain insufficient days since ovulation to permit synchronous transplantation, inform doctor that transfer date was missed for the current cycle;

(g) otherwise, if the predicted ovulation date is in the future, or the predicted date of ovulation has already occurred for the current cycle and there remain sufficient days since ovulation to permit synchronous transplantation during the current cycle, conveying the ovulation date to the patient’s doctor. Upon being alerted to the predicted date of ovulation, the doctor will then make a determination as to when to call the patient to the fertility clinic for FET based on the age of the embryo as explained above. Typically, this calculation is made by the doctor and may be even be adjusted prior to conveying to the patient based on factors that are not predictable or programmable. For this reason, some embodiments allow for the doctor’s intervention before determining the optimal day for performing FET. But obviously this does not preclude the possibility of recording the age of the embryo to be transferred to the target patient, and computing the optimal day in the cycle for performing FET as the ovulation day + the embryo’s age.

It should be understood that Fig. 5 shows only one possible implementation of the invention and is not intended to be limiting. Thus, in the approach described above, we inform the doctor if the predicted date of ovulation has already occurred for the current menstrual cycle, and there remain insufficient days since ovulation to permit synchronous transplantation in the current cycle. This leaves it open for the doctor to decide how best to proceed and when to direct the patient to do a subsequent blood test and optionally ultrasound.

But it is also possible for the algorithm to determine a new day in the next or a subsequent menstrual cycle when the patient should perform a new blood test, and convey to the client device an instruction when to perform the subsequent blood test. Moreover, in the case where it is not possible to perform synchronous transplantation in the current cycle, the doctor may decide to skip one or more cycles before repeating the procedure or the algorithm may be configured to propose a suitable date in a future cycle.

Figs. 6 to 9 are pictorial representations showing images displayed on a doctor’s or a patient’s smartphone during communication between the patient and a fertility clinic during successive stages of a FET treatment plan. Reverting to Fig. 1, we will relate to the patient 14 whose assumed name is Janet Smith and who has downloaded a customized client application to her smartphone 16. Prior to using the application, Janet has registered at the fertility clinic and her details have been recorded including measurements of height, weight, BMI, age, previous menstrual cycle and other factors as discussed previously. Alternatively, she can register and enter all the relevant information using the client application. She is then connected with her doctor, who ultimately decides which treatment she receives (FET, IUI, etc.), and authorizes her to use the application. Upon being so authorized, she is given a username and password, allowing her to login to the application from home using her smartphone or laptop or another similar device.

Fig. 6 shows the doctor’s smartphone screen upon activating treatment. We see the patient’s name and the name of the fertility clinic of which she is a patient. It will be understood that in those cases where the server program is accessed from a central server, many different fertility clinics and doctors will typically access the central computer for feeding their respective patients’ data to a server program to obtain a prediction of the patient’s ovulation date. The central computer can maintain a database of all patients or alternatively the relevant data can be conveyed to the central computer from the fertility clinic’s computer 11. We also assume that the fertility clinic is able to provide different kinds of assisted reproduction, which may be selected from a dropdown menu 20. In our case, Janet is undergoing FET treatment and the dropdown menu 20 may default to FET based on the client’s medical record accessed from the central server 18 or from the doctor’s computer 11, depending on how the system is configured. Alternatively, the doctor can select FET from a number of options listed in the dropdown menu. In like manner, on accessing the system for the first time, the number of cycles is shown as 1 since this is Janet’s first menstrual cycle that is being monitored. This, too, will be updated based on data that is stored in the central server 18 or in the doctor’s computer 11, depending on where Janet’s records are maintained.

The doctor initializes the treatment plan by selecting the command button “Activate Treatment”. When a smartphone is used to run the doctor or client application, the interface shown in Fig. 6 is displayed on a touchscreen and the command buttons are actuated by touch. But obviously this is not a requirement when run on a laptop or other personal computer.

Fig. 7 shows the next stage when the patient logs in to the system after having been registered and assigned a treatment plan. A new instruction has been sent to Janet by her doctor requesting her to perform a blood test on Day 8 of her cycle in order to measure her levels of LH, Estradiol (E2) and progesterone. Janet confirms receipt of this instruction by touching the software button “Approve Instruction”.

Fig. 8 shows the next stage after she has done the blood test on Day 8 of her cycle. When she now logs in to the client application, she sees the “Add Test Results” screen and the current date. She can enter the relevant blood test results and if she had been directed to perform an ultrasound, she could now upload an image of the ultrasound results. Janet can then confirm that the data are correct by touching “Confirm”. In an alternative embodiment, Janet can upload a file, for example a PDF document, as received from the laboratory and the fertility clinic can enter the actual results for access by the server application. In yet a further embodiment, the blood test results could be conveyed to the server application directly by the laboratory for example when the laboratory and the fertility clinic are administered by a common health fund. In this case, the blood test results could be shown on the client application interface without the need for manual entry by the patient.

Pressing “confirm” conveys Janet’s blood test results (and optionally her ultrasound results) to the server application or prompts the server application to run the prediction program using Janet’s data, which it accesses from her medical record, depending on system configuration.

Fig. 9 is a pictorial representation showing an image of the doctor’s smartphone client application prior to issuance of an instruction to the patient. In this case, the doctor sees the Ovulation Tracking Table for Janet. The legend “UB” indicates that she received instructions to perform ultrasound and blood tests on Day 11 of her cycle, corresponding to 24/2/2022 and Janet’s levels of LH, Estradiol and progesterone are shown. In this example no ultrasound was performed, so the follicle size is empty. It is also seen the predicted ovulation date is Day 13 of Janet’s cycle, corresponding to 26/2/2022. The doctor can override the predicted ovulation date by selecting an earlier or later date, prior to confirming the displayed ovulation date. Upon confirmation the server application computes the optimal date for performing FET and sends an instruction to Janet’s smartphone application.

It will also be understood that the system according to the invention may be a suitably programmed computer. Likewise, the invention contemplates a computer program being readable by a computer for executing the method of the invention. The invention further contemplates a machine-readable memory tangibly embodying a program of instructions executable by the machine for executing the method of the invention.