In Vitro Fertilization (IVF) Cumulative Pregnancy Rate Prediction From Basic Patient Characteristics

Bo Zhang, Yuqi Cui, Meng Wang, Jingjing Li, Lei Jin, Dongrui Wu

Abstract

Tens of millions of women suffer from infertility worldwide each year. In vitro fertilization (IVF) is the best choice for many such patients. However, IVF is expensive, time-consuming, and both physically and emotionally demanding. The first question that a patient usually asks before the IVF is how likely she will conceive, given her basic medical examination information. This paper proposes three approaches to predict the cumulative pregnancy rate after multiple oocyte pickup cycles. Experiments on 11,190 patients showed that first clustering the patients into different groups and then building a support vector machine model for each group can achieve the best overall performance. Our model could be a quick and economic approach for reliably estimating the cumulative pregnancy rate for a patient, given only her basic medical examination information, well before starting the actual IVF procedure. The predictions can help the patient make optimal decisions on whether to use her own oocyte or donor oocyte, how many oocyte pickup cycles she may need, whether to use embryo frozen, etc. They will also reduce the patient's cost and time to pregnancy, and improve her quality of life.

Introduction

In Vitro Fertilization (IVF) represents a critical medical intervention for millions of couples worldwide. However, the process is notoriously expensive, often exceeding $15,000 per cycle in some regions, and carries significant emotional and physical burdens. A major challenge for patients is the lack of early, reliable success predictions. Traditionally, accurate prognostic data is only available after the IVF process has begun—relying on factors like embryo morphology or ovarian response to stimulation. By this stage, patients have already committed substantial time and financial resources.

This research addresses the need for a "pre-flight" diagnostic tool. By utilizing basic biomarkers available during the very first clinical visit (Day 3), the study aims to predict the cumulative pregnancy rate—the probability of conceiving after multiple egg retrieval cycles—well before the actual procedure commences.

Methodology

The study utilized a dataset of 11,190 Chinese couples from Tongji Hospital. Through logistic regression, the researchers identified 11 key features predictive of success, including female age, BMI, infertility duration, Antral Follicle Count (AFC), Anti-Mullerian Hormone (AMH), and Follicle-Stimulating Hormone (FSH).

To handle the high degree of patient heterogeneity, the authors compared three distinct machine learning architectures:

  1. Clustering Model: Groups similar patients using $k$-means and predicts based on the group's average success rate.
  2. Support Vector Machine (SVM): A standard supervised model using a Radial Basis Function (RBF) kernel.
  3. C-SVM (Hybrid Model): This novel approach first clusters patients ($k=30$), merges statistically similar clusters via log-rank tests, and then trains a dedicated SVM for each resulting cluster.
The three proposed machine learning architectures: Clustering, SVM, and the hybrid C-SVM (C-SVM).

The pipeline ensures data consistency through one-hot encoding for categorical variables and z-normalization for numerical features.

Key Results and Efficiency

The hybrid C-SVM model emerged as the superior approach. While both SVM and C-SVM achieved an Area Under the Curve (AUC) of approximately 0.70, the C-SVM significantly reduced prediction bias, achieving a Root Mean Squared Error (RMSE) of 0.0267 compared to the standard SVM's 0.0794.

Beyond statistical accuracy, the practical advantages of this "Day 3" approach are substantial. As shown in the timeline comparison below, the proposed model provides a prediction within 1–3 days at a cost of approximately $50. In contrast, conventional methods require 2–3 months of treatment and cost roughly $4,500 before providing similar prognostic clarity.

Comparison of the proposed "Day 3" approach versus conventional IVF prediction timelines and costs.

Discussion and Conclusion

The success of the C-SVM model highlights the importance of accounting for patient heterogeneity; factors influencing success for a younger patient with PCOS differ significantly from those for an older patient with a diminished ovarian reserve. By clustering patients first, the model captures these nuances more effectively than a global model.

While the current model is optimized for a Chinese population and focuses on the first three cycles, it demonstrates that basic pre-treatment data is highly predictive. Future work will look to incorporate male-side factors and lifestyle variables to further refine these early-stage predictions, ultimately helping patients make more informed medical and financial decisions.

Summary updated

Abstract

Tens of millions of women suffer from infertility worldwide each year. In vitro fertilization (IVF) is the best choice for many such patients. However, IVF is expensive, time-consuming, and both physically and emotionally demanding. The first question that a patient usually asks before the IVF is how likely she will conceive, given her basic medical examination information. This paper proposes three approaches to predict the cumulative pregnancy rate after multiple oocyte pickup cycles. Experiments on 11,190 patients showed that first clustering the patients into different groups and then building a support vector machine model for each group can achieve the best overall performance. Our model could be a quick and economic approach for reliably estimating the cumulative pregnancy rate for a patient, given only her basic medical examination information, well before starting the actual IVF procedure. The predictions can help the patient make optimal decisions on whether to use her own oocyte or donor oocyte, how many oocyte pickup cycles she may need, whether to use embryo frozen, etc. They will also reduce the patient's cost and time to pregnancy, and improve her quality of life.
BibTeX
@inproceedings{zhang2019vitro,
  author = {Bo Zhang and Yuqi Cui and Meng Wang and Jingjing Li and Lei Jin and Dongrui Wu},
  title = {{In Vitro Fertilization (IVF) Cumulative Pregnancy Rate Prediction From Basic Patient Characteristics}},
  booktitle = {IEEE Access},
  year = {2019},
  url = {https://openreview.net/forum?id=aql2FeBRJ31},
}
Figure