P-014 Laboratory medicine Poster

Artificial intelligence-based evaluation of homocysteine as a biomarker in traumatology patients

Audience Choice Award

To rate this poster, scan the QR code on the poster itself — voting only works on site. Every attendee has 3 stars and can give at most one per poster. A star is final and cannot be taken back.

Authors

N. Hasanova presenting

Department of Laboratory, Scientific Research İnstitute of Traumatology and Orthopedics, Baku, Azerbaijan

#0

Abstract

Background: Monitoring biochemical markers is essential for evaluating bone metabolism and healing in traumatology patients. Homocysteine (Hcy) has been associated with impaired bone quality and increased fracture risk. Artificial intelligence (AI) offers new opportunities for predictive analysis in laboratory medicine.

Objective: To assess changes in Hcy levels during treatment, examine their relationship with biochemical markers and bone mineral density, and evaluate AI-based models for predicting fracture risk.

Methods: A total of 128 participants were included: patients with fractures, non-fracture conditions, and healthy controls. Hcy levels were measured at baseline, day 10, and, one month using ELISA. Additional parameters included calcium, magnesium, phosphorus, vitamin D, ALP, and total protein.

Statistical analyses were performed using SPSS 20.0 with non-parametric tests and Spearman correlation. AI-based models, including Random Forest and Logistic Regression, were applied. Data were split into training (80%) and testing (20%) sets, and performance was evaluated using ROC analysis.

Results: Hcy levels were significantly higher in fracture patients at baseline and decreased after treatment (p = 0.004). AI models demonstrated strong predictive performance (AUC = 0.903, sensitivity 84%, specificity 93%). A significant negative correlation was observed between Hcy and bone mineral density (r = −0.441, p = 0.027). Elevated Hcy levels were also associated with reduced vitamin D and altered ALP activity.

Conclusion: AI-based models improve the interpretation of laboratory data and enable accurate prediction of Hcy dynamics and fracture risk. Homocysteine may serve as a valuable biomarker for monitoring and risk assessment in traumatology patients.