patient
This retrospective study included 250 women who underwent elective fertility preservation therapy at Meir Medical Center in Israel from 2019 to 2022. Data extracted included (1) demographics (age, BMI), (2) clinical antral follicle count (AFC), and gonadotropin medications. , ovulation-inducing agent (GnRH agonist, hCG), number of stimulation days, starting dose and total gonadotropin dose, endometrial thickness on the day of induction, (3) Laboratory values: basal follicle-stimulating hormone (FSH), luteinizing hormone level (LH) )), progesterone, estradiol, and LH levels on the day of induction. At our medical center, the standard regimen for fertility preservation includes the use of an antagonist her protocol starting with 300 IU of gonadotropin.
Ethics approval
This study was approved by the Meir Medical Center Institutional Review Board (MMC-393-20). We confirm that all research was conducted in accordance with the Declaration of Helsinki. The Meir Medical Center Institutional Review Board, under the direction of Professor Ilan Cohen, approved a waiver of the informed consent requirement.
outcome measure
The primary outcome was the number of metaphase II oocytes obtained during treatment. Cobo et al. To obtain a reasonable success rate of live birth (40% for women under 35 years and 20% for women over 35 years), at least 8 MII oocytes should be stored, but 15 We proposed that additional oocytes beyond 20% contribute little to cumulative live births.Survival rate17.
Based on this information, we decided to classify the results into three classes according to oocyte number (OC): low (≤ 8 oocytes), medium (9–15 oocytes). , or high (≥ 16). We used various machine learning models and logistic regression to predict the outcome classes.
Analysis of pre-treatment and start date (post-treatment)
We performed two independent analyzes and trained the machine learning model with different subsets of parameters. Pre-treatment analyzes used only parameters available at the first visit before treatment initiation, including age, BMI, AFC, basal levels of estradiol, LH, and FSH. The results of this analysis were used to evaluate potential treatment outcomes before starting an IVF cycle. This may help decide whether to start treatment.
In the trigger date analysis, we included all the collected parameters to train the machine learning model. These include all parameters used in the pre-treatment analysis, plus estradiol, progesterone, LH, endometrial thickness on the day of induction, treatment protocol, starting dose, total dose, and number of days of stimulation. I was there. The results of this analysis were used to assess the effectiveness of treatments already initiated and to reduce uncertainty regarding patient outcomes regarding egg retrieval procedures.
data analysis
Discrete variables are presented as numbers and percentages, and continuous variables are presented as mean ± standard deviation (SD). p-values were calculated using the Student t test or χ.2. A p-value < 0.05 is considered significant. One-way analysis of variance (ANOVA) was used to compare the means of demographic, clinical, and laboratory variables between the three outcome groups. Bonferroni correction was applied to adjust for multiple comparisons.
