In-vitro fertilization (IVF) failure rates currently remain above 65% and unfortunately, many cases involve unknown causes. But researchers now focus on uterine receptivity and they solve this problem because uterine peristalsis determines receptivity and success. Therefore, a multi-center study investigated IVF outcome prediction using machine learning and ultrasound strain analysis.

First, the researchers recruited 62 patients and then they performed transvaginal scans and also they used strain analysis. Second, the team extracted 25 features and they also used speckle tracking to ensure accuracy. Next, they analyzed frequency and amplitude and coordination features. Consequently, they evaluated uterine activity and then they compared algorithms and also they tested accuracy. This comprehensive approach allowed them to capture subtle uterine movements effectively.

Reliable IVF Outcome Prediction via Machine Learning

Notably, the study compared SVM and KNN and AdaBoost models. And the SVM model performed the best and also it reached an average area under the ROC curve of 0.81. In fact, it reached a high score so it distinguishes pregnancies effectively and also accurately. Specifically, coordination features showed links to success and frequency also showed strong associations. Therefore, these specific motion patterns serve as vital indicators for IVF outcome prediction in clinical settings.

Moreover, the SVM model demonstrated reliable performance and also it worked well across multiple fertility centers. Thus, clinicians can use this framework and also they make better decisions because the method is safe and also non-invasive. Similarly, it helps personalize treatment plans and also it enhances implantation success. Finally, this innovative approach might finally enhance IVF success rates and indeed, AI offers a promising path forward and also it improves patient care.

How does uterine peristalsis affect IVF success?

Uterine peristalsis determines the receptivity of the environment. Excessive or uncoordinated motion can prevent embryo implantation, so its assessment is vital for success.

Which machine learning model is most effective for prediction?

The Support Vector Machine (SVM) model demonstrated the best performance in this study, and it achieved an average AUC of 0.81 for predicting outcomes.

What features are most associated with clinical pregnancy?

Features related to the frequency and coordination of uterine peristalsis showed the strongest association with successful clinical pregnancy rates.

Disclaimer: This content is for informational and educational purposes only. It does not constitute medical advice or a professional relationship. Refer to the latest local and national guidelines for clinical practice.

References

• Cheng A et al. Prediction of in-vitro fertilization outcome by ultrasound strain analysis and machine learning: A multi-center study. Ultrasound J. 2026 Feb 24. doi: 10.5826/tuj.2026.18257. PMID: 41738337.


• Vidal A, et al. Lower pregnancy rate in women with high uterine peristalsis before embryo transfer: a systematic review and meta-analysis. Reprod Biol Endocrinol. 2025;23:49. doi: 10.1186/s12958-025-01380-5.