Recurrent spontaneous abortion (RSA) remains a significant clinical challenge for obstetricians and rheumatologists. While traditional markers identify some risks, many patients fall into a diagnostic gap. This research utilizes machine learning to improve antiphospholipid antibody prediction in patients with a history of recurrent spontaneous abortion. By integrating criteria and non-criteria antibodies, the study provides a more comprehensive risk profile for adverse pregnancy outcomes.

In this multicenter prospective study, researchers recruited 1,321 participants. They analyzed fifteen different antiphospholipid antibodies (aPLs) using chemiluminescence immunoassay. Subsequently, the team trained six machine learning algorithms to evaluate the data. Notably, the light gradient boosting machine (LGBM) and random forest (RF) models demonstrated superior accuracy compared to single indicators. Specifically, the RF model achieved an area under the curve (AUC) of 0.885 for predicting pregnancy outcomes. Therefore, these algorithms offer significant potential as auxiliary diagnostic tools for early warning.

The Role of Antiphospholipid Antibody Prediction in Clinical Practice

Modern diagnostic methods often struggle with seronegative cases. However, machine learning bridges this gap by identifying complex patterns across multiple biomarkers. The study showed that integrated models drastically outperform individual antibody tests. Furthermore, calibration and decision curve analysis confirmed the clinical utility of these tools. Consequently, doctors can now identify high-risk patients earlier and with greater precision. Although larger external validation is necessary, these findings represent a major step toward personalized reproductive medicine.

Frequently Asked Questions

What are non-criteria antiphospholipid antibodies?

Non-criteria aPLs are antibodies not included in the standard diagnostic criteria for Antiphospholipid Syndrome. However, they are clinically relevant as they help identify patients who experience RSA despite testing negative for traditional markers.

How does the LGBM model assist in RSA diagnosis?

The Light Gradient Boosting Machine (LGBM) model processes multiple antibody levels simultaneously. Consequently, it achieves higher predictive accuracy for RSA than single-marker tests, helping clinicians detect risks that standard protocols might miss.

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

Jiang X et al. Machine learning-based integration of antiphospholipid antibodies for predicting pregnancy outcomes in patients with recurrent spontaneous abortion: a multicenter prospective study. Clin Chem Lab Med. 2026 Mar 02. doi: 10.1515/cclm-2026-0088. PMID: 41764780.

Cervera R, et al. Antiphospholipid syndrome. Nat Rev Dis Primers. 2015;1:15010. doi: 10.1038/nrdp.2015.10.

Abrahams VM. Mechanism of antiphospholipid antibody-induced pregnancy complications. Thromb Res. 2020;151 Suppl 1:S99-S102. doi: 10.1016/S0049-3848(17)30077-4.