Method for assessing likelihood of metformin therapy effectiveness in patients with polycystic ovary syndrome based on integral nomogram

Abstract

FIELD: medicine.SUBSTANCE: invention relates to clinical endocrinology, gynecology and personalized metabolic therapy and can be used to assess the likelihood of the effectiveness of metformin therapy in patients with polycystic ovary syndrome (PCOS). Before starting metformin therapy, the patient's insulin resistance index HOMA-IR, the level of soluble receptor for advanced glycation end products sRAGE in the blood serum, and body mass index BMI are determined. The values of HOMA-IR, sRAGE level and BMI are found on the corresponding scales of the nomogram presented in the figure, projected onto the point scale, converted into point values, summed up, the total value is found on the total point scale and projected onto the “Probability of therapy effectiveness” scale. The final probability of the effectiveness of therapy P is determined, which is interpreted as low if the P value is less than 0.40, as intermediate if the P value is from 0.40 to 0.70, and as high if the P value is more than 0.70.EFFECT: invention improves the accuracy of assessing the likelihood of the effectiveness of metformin therapy in patients with PCOS, reduces the risk of prescribing ineffective metabolic therapy, and shortens the time for selecting the optimal treatment strategy.1 cl, 1 dwg, 3 ex

Description

[0001] The invention relates to the field of medicine, primarily to clinical endocrinology, gynecology and personalized metabolic therapy, and can be used to assess the likelihood of the effectiveness of metformin therapy in patients with polycystic ovary syndrome before starting treatment.

[0002] Polycystic ovary syndrome (PCOS) is one of the most common endocrine-metabolic disorders in women of reproductive age and is characterized by pronounced heterogeneity of clinical, hormonal, and metabolic manifestations. One of the key pathogenetic mechanisms of PCOS is insulin resistance, the presence and severity of which vary significantly among patients and do not always correlate with traditional indicators of carbohydrate metabolism. [Teede HJ, Tau CT, Laven JJE, et al. Recommendations From the 2023 International Evidence-based Guideline for the Assessment and Management of Polycystic Ovary Syndrome. J Clin Endocrinol Metab. 2023;108(10):2447-2469. doi:10.1210 / clinem / dgad463].

[0003] International clinical guidelines for the diagnosis and management of PCOS consider metformin as a basic component of therapy, primarily in patients with metabolic disorders. They also emphasize the need for individualization of therapy, taking into account the phenotypic and metabolic heterogeneity of the disease, since there are no universal criteria for predicting the effectiveness of metformin before starting treatment [Melin J, Forslund M, Alesi S, et al. The impact of metformin with or without lifestyle modification versus placebo on polycystic ovary syndrome: a systematic review and meta-analysis of randomized controlled trials. Eur J Endocrinol. 2023;189(2):S37-S63. doi:10.1093 / ejendo / lvad098].

[0004] In clinical practice, the effectiveness of metabolic therapy, including metformin therapy, is typically predicted based on isolated glycemic parameters, such as fasting glucose, oral glucose tolerance test results, and glycated hemoglobin levels. These approaches reflect the current state of carbohydrate metabolism but do not assess the individual sensitivity of metabolic and hormonal systems to therapy and do not reliably predict treatment response before it begins [Fu J. et al. Predicting metformin efficacy in improving insulin sensitivity among women with PCOS and IR. 2024].

[0005] There are known methods for assessing insulin resistance based on the calculation of the insulin resistance index (HOMA-IR), as well as methods for stratifying metabolic risks based on body mass index (BMI). However, these methods are used primarily to diagnose metabolic disorders and are not used in combination to predict the effectiveness of metformin therapy in patients with PCOS [Sasaki N, Ueno Y, Higashi Y. Indicators of insulin resistance in clinical practice. Hypertens Res. 2024;47(4):978-980. doi:10.1038 / s41440-023-01566-7].

[0006] The literature describes the role of soluble receptor for advanced glycation end products (sRAGE) as a marker of antiglycation and protective metabolic reserve, but this indicator is used primarily for research purposes and is not used in clinical decision-making algorithms for the appropriateness of prescribing metabolic therapy [Tatone C, Di Emidio G, Placidi M, et al. AGEs-related dysfunctions in PCOS: evidence from animal and clinical research. J Endocrinol. 2021;251(2):R1-R9. Published 2021 Sep 28. doi:10.1530 / JOE-21-0143].

[0007] Thus, the prior art lacks methods for an integrated assessment of the various pathogenetic components of insulin resistance and for quantitatively predicting the likelihood of metformin therapy effectiveness in patients with PCOS before treatment. These solutions do not take into account the synergistic effects of insulin resistance, protective antiglycation reserve, and anthropometric characteristics on treatment outcome, leading to the prescription of standard treatment without regard for individual patient sensitivity.

[0008] The technical problem that the claimed invention is aimed at solving is the lack of a method that allows for the determination with sufficient accuracy and reproducibility of the individual probability of the effectiveness of metformin therapy in patients with PCOS before the start of treatment, taking into account the heterogeneity of the pathogenesis of insulin resistance.

[0009] The use of existing diagnostic and prognostic approaches based on isolated glycemic indicators or anthropometric data does not allow identifying patients with a high probability of resistance to standard metabolic therapy, which leads to loss of time, progression of metabolic disorders and reduced effectiveness of treatment measures.

[0010] Therefore, there is a need to create a method that provides an integrated assessment of the various metabolic mechanisms involved in the formation of the response to metformin therapy and allows for stratification of patients according to the likelihood of achieving a clinically significant effect before prescribing treatment.

[0011] The technical result of the invention is an increase in the accuracy of assessing the probability of the effectiveness of metformin therapy in patients with polycystic ovary syndrome, and a high information content of the method due to an integrated assessment of the most significant pathogenetic links of insulin resistance, assessed before the start of treatment.

[0012] An additional technical result is a reduction in the risk of prescribing obviously ineffective metabolic therapy by pre-stratifying patients according to the probability of achieving a clinically significant therapeutic effect, as well as reducing the time to select the optimal treatment strategy.

[0013] The specified technical result is achieved in a method for assessing the likelihood of the effectiveness of metformin therapy in patients with polycystic ovary syndrome, characterized by the fact that the patient's insulin resistance index HOMA-IR, the level of soluble receptor for advanced glycation end products sRAGE in the blood serum and body mass index BMI are determined before the start of metformin therapy, and based on the set of specified indicators, the individual probability of the effectiveness of therapy is calculated using a mathematical model of logistic regression.

[0014] P=1 / (1+e -z )),

[0015] where Z=-0.52 - 0.8961⋅HOMA-IR+0.0016⋅sRAGE+0.1002⋅BMI,

[0016] implemented in the form of a graphic nomogram consisting of 7 parallel linear scales:

[0017] - HOMA-IR insulin resistance index scale with values ​​from 6 to 1.5 units;

[0018] - soluble receptor for advanced glycation end products sRAGE scale in values ​​from 600 to 1600 pg / ml;

[0019] - Body Mass Index (BMI) scale with values ​​from 16 to 40 kg / m 2 ;

[0020] - scale for converting indicator values ​​into points (Points) from 0 to 100;

[0021] - Total Points scale from 0 to 180;

[0022] - Linear Predictor scale from “minus” 2 to 3.5;

[0023] - scale of probability of treatment effectiveness in values ​​from 0.2 to 0.95,

[0024] in this case, the values ​​of HOMA-IR, sRAGE level and BMI are found on the corresponding scales of the nomogram, projected onto the scale for converting the values ​​of the indicators into points, converted into point values, summed up according to the scale of total points to obtain the indicator of total points and, according to the scale of probability of treatment effectiveness of the nomogram, the final probability of the effectiveness of therapy P is determined, which is interpreted as low for a P value of less than 0.40, as intermediate for a P value from 0.40 to 0.70 and as high for a P value of more than 0.70.

[0025] The obtained probability value is used both to stratify PCOS patients according to the expected effectiveness of metformin therapy and to support clinical decision-making on the choice of management tactics before starting treatment.

[0026] The method is illustrated in Fig., which shows a nomogram.

[0027] The HOMA-IR index is used to quantify the degree of basal insulin resistance, with an increase in its value associated with a decreased likelihood of metformin therapy effectiveness. The sRAGE level is used as an indicator of protective antiglycation reserve, with higher sRAGE values ​​corresponding to a higher likelihood of a favorable response to therapy. BMI is used as an integral anthropometric indicator reflecting the severity of a metabolic component potentially sensitive to the insulin-sensitizing effect of metformin.

[0028] Based on certain indicators, the individual probability of the effectiveness of metformin therapy is calculated using a mathematical model of logistic regression, the results of which are presented in the form of a graphical nomogram.

[0029] The use of a nomogram involves converting the values ​​of each indicator into a point equivalent, summing up the points obtained, and determining the final probability of therapy effectiveness according to the corresponding scale.

[0030] The obtained probability is interpreted using threshold zones, whereby if the probability value is less than 0.40, a conclusion is made about a low probability of therapy effectiveness, if the value is from 0.40 to 0.70, an intermediate probability of therapy effectiveness is made, and if the value is more than 0.70, a high probability of therapy effectiveness is made.

[0031] The claimed method is used prior to the initiation of metformin therapy and is used to decide on the appropriateness of its use or the need to select alternative or more effective treatment approaches. However, the method is not based on an assessment of current glycemic levels or glycated hemoglobin levels and does not require the use of these parameters to calculate the likelihood of therapy effectiveness.

[0032] The set of features used and the order of their application do not follow obviously from the state of the art and ensure the achievement of the declared technical result, which confirms the compliance of the claimed solution with the criteria of novelty and inventive step.

[0033] The implementation of the claimed mathematical model in the form of a graphical nomogram ensures the achievement of the claimed technical result, which consists in increasing the accuracy and reproducibility of predicting the effectiveness of metformin therapy in patients with PCOS.

[0034] The use of a nomogram allows for an integrated assessment of the combined influence of independent pathogenetic factors—the degree of basal insulin resistance, the level of protective antiglycation reserve, and the anthropometric metabolic component—taking into account their nonlinear contribution to the likelihood of achieving the clinical effect of therapy.

[0035] The graphical implementation of the logistic model provides a standardized transformation of input indicator values ​​into point equivalents and their subsequent transformation into the probability of treatment effectiveness, which eliminates the variability of manual calculations and reduces the risk of subjective errors in clinical decision-making.

[0036] Thus, the use of the nomogram ensures the reproducibility of the calculation of the probability of a therapeutic response, the possibility of using the method in clinical practice without the use of computational tools and an increase in the validity of the choice of treatment tactics before the start of therapy.

[0037] To quantitatively determine the individual probability of therapy effectiveness, a prognostic mathematical model is used that links the values ​​of input metabolic parameters with the probability of achieving a clinically significant therapeutic response.

[0038] The probability of treatment effectiveness (P) is calculated using the logistic function

[0039] P=1 / 1+e -z ,

[0040] where

[0041] (e) is the base of the natural logarithm (Euler's number), approximately equal to 2.71828.

[0042] (Z) is a linear predictor of the model, which is calculated using the formula:

[0043] Z=a+b1⋅(HOMA-IR)+b2⋅(sRAGE)+b3⋅(MMT)

[0044] • b1, b2, b3 - regression coefficients,

[0045] • HOMA-IR - insulin resistance index,

[0046] • sRAGE - soluble receptor for advanced glycation end products,

[0047] • BMI - body mass index.

[0048] In a preferred embodiment of the invention, the model coefficients have the following values:

[0049] • b1=-0.8961 (for HOMA-IR),

[0050] • b2=+0.0016 (for sRAGE),

[0051] • b3=+0.1002 (for BMI).

[0052] • a is a constant (free term) of the model, obtained by reverse calculation based on the nomogram and is equal to 0.52 (determined at the stage of construction and calibration of the regression model and taken into account when forming the scale of final scores and the probability scale in the nomogram).

[0053] Thus, the linear predictor of the model is calculated using the formula

[0054] Z=-0.52 - 0.8961⋅HOMA-IR+0.0016⋅sRAGE+0.1002⋅BMI.

[0055] Clinical meaning of linear predictor:

[0056] The linear predictor (Z) is an integral quantitative characteristic of the combined influence of pathogenetically independent metabolic factors that determine the probability of response to therapy.

[0057] A negative value of the coefficient for HOMA-IR reflects a decrease in the probability of therapy effectiveness with an increase in the degree of basal insulin resistance.

[0058] A positive coefficient for the sRAGE level reflects an increase in the likelihood of therapy effectiveness with an increase in the individual antiglycation protective reserve.

[0059] The positive coefficient in the body mass index assessment reflects the contribution of the anthropometric metabolic component to the sensitivity to metformin therapy.

[0060] Thus, the linear predictor (Z) is an integral quantitative characteristic of the total influence of metabolic factors in a patient with polycystic ovary syndrome.

[0061] Using the logistic function transforms a linear predictor (Z) into a probability value that ranges from 0 to 1.

[0062] This ensures the clinical interpretability of the result and allows the obtained value to be used directly for stratification of patients with polycystic ovary syndrome.

[0063] Implementation of the model in the form of a nomogram: For practical application, the mathematical model is implemented in the form of a graphical nomogram (Fig. 1), which is a visual computational tool for determining the probability of therapy effectiveness without performing mathematical calculations.

[0064] The nomogram contains:

[0065] • upper scale for converting parameter values ​​into points (Points scale),

[0066] • HOMA-IR index scale;

[0067] • sRAGE level scale;

[0068] • body mass index scale;

[0069] • Total Points scale;

[0070] • scale of probability of therapy effectiveness (Probability).

[0071] Each input parameter value corresponds to a certain number of points, proportional to the contribution of this parameter to the linear predictor of the model. Point values ​​are generated based on the coefficients of the mathematical model and ensure equivalence between graphical calculations and analytical probability calculations.

[0072] The method is carried out, for example, as follows.

[0073] Before starting therapy, the patient's insulin resistance index (HOMA-IR), serum soluble receptor for advanced glycation end products (sRAGE) levels, and body mass index are determined. These parameters are determined using standard clinical and laboratory methods accepted in medical practice.

[0074] The obtained values ​​are used as input parameters for a logistic regression mathematical model, implemented as a graphical nomogram. The nomogram includes a scale for converting indicator values ​​into points, scales for the corresponding indicators, a scale for total points, and a scale for the probability of treatment effectiveness.

[0075] For each of the defined indicators, the corresponding value is found on the nomogram scale and projected onto the scale for converting indicator values ​​into points, obtaining a point equivalent. The resulting points are summed to form the total score. The final probability (P) of the effectiveness of biguanide metformin therapy is determined using the total score scale.

[0076] The resulting probability is interpreted using threshold zones. A P value of less than 0.40 indicates a low probability of treatment effectiveness; a P value in the range of 0.40 to 0.70 indicates an intermediate probability of treatment effectiveness; and a P value greater than 0.70 indicates a high probability of treatment effectiveness.

[0077] The assessment result is used before initiating metformin therapy to decide on the appropriateness of its use or the need for alternative or enhanced treatment strategies. This method does not require fasting blood glucose levels, an oral glucose tolerance test, or glycated hemoglobin levels and is not based on an assessment of the current state of carbohydrate metabolism.

[0078] The novelty of the method lies in the synergistic combination of markers reflecting different links in pathogenesis, which ensures the stability of the model: HOMA-IR (Coefficient: -0.8961, p<0.0001): The negative coefficient in the model indicates that the likelihood of success decreases as the HOMA-IR index increases. Clinically, this means that patients with severe basal insulin resistance require more effective or alternative treatment regimens, as standard therapy is highly likely to be ineffective. sRAGE (Coefficient: +0.0016, p=0.08): included in the model as a marker of protective antiglycation reserve. A positive coefficient indicates that, all other factors being equal, a patient with a higher sRAGE level has a better chance of restoring metabolic health. This allows for the identification of "hidden potential" for recovery.

[0081] BMI (Coefficient: +0.1002, p=0.013): In the context of this model, the positive association between BMI and efficacy reflects the presence of a "metabolic component" to therapy. Overweight patients often respond better to interventions aimed at weight loss and improving insulin sensitivity, compared to normal-weight individuals, whose pathogenesis is more complex and less responsive to standard treatment.

[0082] The developed model has undergone rigorous internal validation using the bootstrap method (B=200 iterations) to prevent overfitting:

[0083] - Discrimination: The adjusted C-index (AUC) was 0.714, confirming the ability of the model to discriminate between responders and non-responders.

[0084] - Calibration: The adjusted slope of the calibration curve was 0.937 (with an ideal slope of 1.0). This is a critically important parameter for clinical use. It ensures that the predicted probabilities are neither overestimated nor underestimated. If the nomogram predicts 70% success, then in reality, 70% of these patients achieve the target effect. This allows the physician to use the numerical value of risk to make informed decisions.

[0085] The claimed method is a scientifically based tool for personalized medicine, which allows, at the stage of therapy planning, to distinguish promising patients from those resistant to treatment with high accuracy, based on a set of metabolic and anthropometric data.

[0086] The method is implemented in the form of a graphical nomogram (Fig. 1) for ease of use in clinical practice.

[0087] The method is supported by the following clinical examples.

[0088] Example 1. A 27-year-old female patient with PCOS (phenotype C). She sought examination before planning a pregnancy; she had a history of miscarriage. The patient was found to have insulin resistance (HOMA-IR>2.5), and her BMI was within the normal reference range (BMI=22.4 kg / m2). 2 ). Objective: to assess the individual probability of the effectiveness of metformin therapy before the start of treatment. Step 1. Baseline parameters (before therapy): HOMA-IR=2.9, sRAGE=1380 pg / ml, BMI=22.4 kg / m 2 .

[0089] Step 2. Calculating HOMA-IR Points:

[0090] 1. On the HOMA-IR Index scale, find the value 2.9.

[0091] 2. From this point, draw a vertical line up to the Points scale.

[0092] 3. The Points scale reads: “69 points. Points for HOMA-IR: ≈69 points.

[0093] Step 3. Calculating sRAGE points:

[0094] 1. On the "sRAGE, pg / ml" scale, the value is 1380.

[0095] 2. Draw a vertical line up to the Points scale.

[0096] 3. Read: "31 points. sRAGE points: ≈31 points.

[0097] Step 4. Scoring for BMI:

[0098] 1. On the scale "BMI, kg / m 2 » find the value 22.4.

[0099] 2. Draw a vertical line up to the Points scale.

[0100] 3. Read: “16 points. BMI points: ≈16 points.

[0101] Step 5. Total Points Total Points:

[0102] Total Points=69+31+16= ≈116 points.

[0103] Step 6. Determining the probability of treatment effectiveness

[0104] 1. On the Total Points scale, find ≈116.

[0105] 2. The marked point is projected down onto the “Probability of Treatment Effectiveness” scale.

[0106] 3. Read the probability P: "0.77 (that is, ≈77%).

[0107] Final interpretation (by thresholds): at P ≈ 0.77, the patient belongs to the group with a high probability of treatment effectiveness (P>0.70).

[0108] Clinical interpretation: The individual probability of the effectiveness of metformin therapy, as monotherapy for the correction of carbohydrate metabolism disorders, is assessed as high.

[0109] Final conclusion (according to the completed algorithm): “The patient belongs to the group with a high probability of the effectiveness of metformin therapy according to the integrated nomogram”, which is used to decide on the advisability of prescribing therapy before its initiation.

[0110] The patient was prescribed metformin at a dose of 2000 mg / day for 12 months. High clinical efficacy of the therapy was achieved: after 12 months of metformin use, the insulin resistance index decreased to the reference value HOMA-IR = 2.1. Good tolerability of the drug was noted. The patient is planning a pregnancy.

[0111] Example 2. A 29-year-old female patient with PCOS (phenotype A) sought examination for menstrual irregularities (oligomenorrhea). The patient was found to have severe insulin resistance (HOMA-IR>2.5) and class I obesity (BMI=30.8 kg / m²). 2 ). The patient's medical history indicates that metformin was used 2 years ago at a dose of 1000-2000 mg / day for 3 months with poor tolerability (severe nausea, dyspeptic disorders) and failure to correct carbohydrate metabolism disorders (insulin resistance). Objective: To assess the individual likelihood of metformin therapy effectiveness before treatment initiation.

[0112] Step 1. Baseline (before therapy)

[0113] • HOMA-IR=5.5

[0114] • sRAGE=620 pg / ml

[0115] • BMI=30.8 kg / m 2 (obesity stage I)

[0116] Step 2. Calculating HOMA-IR Points

[0117] 1. On the HOMA-IR Index scale, find the value 5.5.

[0118] 2. From this point, draw a vertical line up to the Points scale.

[0119] 3. The Points scale reads: ≈11 points.

[0120] HOMA-IR Points: ≈ 11

[0121] Step 3. Calculating sRAGE Points

[0122] 1. On the "sRAGE, pg / ml" scale, find the value 620.

[0123] 2. Draw a vertical line up to the Points scale.

[0124] 3. Read: ≈1 point.

[0125] sRAGE points: ≈1.

[0126] Step 4. Scoring for BMI

[0127] 1. On the scale "BMI, kg / m 2 » find the value 30.8.

[0128] 2. Draw a vertical line up to the Points scale.

[0129] 3. Read: ≈37 points. BMI points: ≈16.

[0130] Step 5. Total Points

[0131] Total score:

[0132] Total Points=11+1.+37= ≈ 49 points

[0133] Step 6. Determining the probability of treatment effectiveness

[0134] 1. On the Total Points scale, find "49."

[0135] 2. The marked point is projected down onto the “Probability of Treatment Effectiveness” scale.

[0136] 3. Read the probability P: "0.27" (that is, "27%).

[0137] Final interpretation (by thresholds): at P ≈ 0.27, the patient belongs to the group with low probability of treatment effectiveness (P < 0.40).

[0138] Clinical interpretation: Despite the presence of severe insulin resistance, the detected low sRAGE level demonstrates an unfavorable protective potential for successful correction of carbohydrate metabolism disorders with this biguanide. The individual likelihood of the effectiveness of metformin therapy as monotherapy is assessed as low.

[0139] Final conclusion (according to the completed algorithm): “The patient belongs to the group of low probability of the effectiveness of metformin therapy according to the integrated nomogram”, which is used to decide on the advisability of prescribing therapy before its initiation.

[0140] The patient was recommended to undergo therapy with glucocorticoid-like peptide 1 agonists (semaglutide) for a year. After 12 months, a significant clinical effect was achieved: a decrease in the insulin resistance index to reference values ​​(HOMA-IR = 2.3), normalization of weight (BMI = 23.4 kg / m 2 ), restoration of a regular menstrual cycle.

[0141] Example 3. A 26-year-old female patient with PCOS (phenotype D) presented for examination due to menstrual irregularities (secondary amenorrhea) and primary infertility. The patient was found to have insulin resistance (HOMA-IR>2.5) and class I obesity (BMI=32.5 kg / m²). 2 ). Objective: to assess the individual probability of the effectiveness of metformin therapy before starting treatment. Step 1. Baseline parameters (before therapy):

[0142] HOMA-IR=3.8.

[0143] sRAGE=1,520 pg / ml.

[0144] BMI=32.5 kg / m 2 (obesity stage I).

[0145] Step 2. Calculating HOMA-IR Points

[0146] 1. On the HOMA-IR Index scale, find the value 3.8.

[0147] 2. From this point, draw a vertical line up to the Points scale.

[0148] 3. The Points scale reads: ≈ 52 points.

[0149] HOMA-IR Score: ≈52.

[0150] Step 3. Calculating sRAGE Points

[0151] 1. On the "sRAGE, pg / ml" scale, find the value 1520.

[0152] 2. Draw a vertical line up to the Points scale.

[0153] 3. Read: ≈36 points sRAGE points: ≈36.

[0154] Step 4. Scoring for BMI

[0155] 1. On the scale "BMI, kg / m 2 » find the value 32.5.

[0156] 2. Draw a vertical line up to the Points scale.

[0157] 3. Read: ≈ 41 points.

[0158] BMI score: ≈ 41

[0159] Step 5. Total Points

[0160] Total score:

[0161] Total Points=52+36+41= ≈129 points.

[0162] Step 6. Determining the probability of treatment effectiveness

[0163] 1. On the Total Points scale, find ≈129.

[0164] 2. The marked point is projected down onto the “Probability of Treatment Effectiveness” scale.

[0165] 3. Read the probability P: "0.82" (that is, "82%).

[0166] Final interpretation (by thresholds): at P ≈ 0.82, the patient is classified as having a high probability of treatment efficacy (P> 0.70). Clinical interpretation: A high sRAGE level reflects preserved protective potential for overcoming glycotoxicity, insulin resistance, and oxidative stress associated with impaired carbohydrate metabolism and obesity, ensuring a high expected response to the use of this biguanide. The individual probability of metformin monotherapy efficacy is assessed as high. Final conclusion (according to the completed algorithm): "The patient is classified as having a high probability of metformin therapy efficacy according to the integrated nomogram," which is used to decide on the appropriateness of prescribing therapy before its initiation.

[0167] The patient was prescribed metformin with a gradual increase in the dose to 2500 mg / day over the course of a year. After 12 months of therapy, a significant clinical effect was achieved: a decrease in the insulin resistance index to reference values ​​(HOMA-IR = 2.4), normalization of weight (BMI = 24.8 kg / m 2 ), restoration of the menstrual cycle. To overcome anovulatory infertility, ovulation stimulation is planned with metformin at a dose of 2500 mg per day.

[0168] The presented clinical examples demonstrate that the effectiveness of metformin therapy in patients with various PCOS phenotypes is determined by an integrated metabolic profile, including the degree of insulin resistance, BMI, and sRAGE level, which fundamentally distinguishes the claimed method from known solutions. The key differentiating factor is the sRAGE level, reflecting an individual's protective antiglycation reserve, which justifies the use of this indicator in the integrated nomogram. The presented embodiments of the invention illustrate the potential application of the claimed method for stratifying patients based on the likelihood of metformin therapy effectiveness before treatment initiation in patients with various PCOS phenotypes.

Images