Early Prediction of Bronchopulmonary Dysplasia in Preterm Infants Using Clinical Data

Early Prediction of Bronchopulmonary Dysplasia Using Clinical Data From the First Three Postnatal Weeks in Preterm Infants: A Retrospective Study With Large Language Models

Early Prediction of Bronchopulmonary Dysplasia in Preterm Infants Using Clinical Data from the First Three Postnatal Weeks with Large Language Models: A Retrospective Study This retrospective, observational study aims to evaluate the early prediction of bronchopulmonary dysplasia (BPD) in preterm infants using clinical data from the first, second, and third postnatal weeks. The study includes infants born before 32 weeks of gestation or weighing less than 1,500 grams, followed at the Neonatal Intensive Care Unit of Konya City Hospital.

The study will compare the performance of different large language models (LLMs), including ChatGPT, Gemini, and Claude, in predicting BPD development. Clinical variables such as gestational age, birth weight, respiratory support, oxygen requirement, mechanical ventilation duration, and infection status will be used.

Primary outcome: Accuracy of BPD risk prediction by each AI model compared to actual clinical outcomes. Secondary outcomes: Sensitivity and specificity of predictions, weekly prediction performance, and comparative performance among AI models.

The results will provide insight into the potential clinical utility of AI-based approaches for early BPD risk assessment in preterm infants.

Study Overview

• Status: Not yet recruiting
• Conditions: Bronchopulmonary Dysplasia
• Intervention / Treatment: Other: Artificial Intelligence-Based Risk Prediction

Detailed Description

Premature birth remains a major risk factor for neonatal morbidity and mortality, with bronchopulmonary dysplasia (BPD) representing one of the most significant chronic pulmonary complications in very preterm infants. Despite advances in neonatal intensive care, early and accurate prediction of BPD remains challenging due to the multifactorial nature of its pathophysiology, involving respiratory support requirements, oxygen exposure, infection burden, and perinatal factors.

This retrospective study evaluates the feasibility of using large language models (LLMs) for early prediction of BPD based on structured clinical data extracted from neonatal intensive care unit (NICU) records. Clinical variables are organized into weekly datasets corresponding to the first, second, and third postnatal weeks to capture the dynamic evolution of respiratory status and clinical condition over time.

Standardized and anonymized patient-level datasets are formatted into structured prompts and provided to multiple LLMs (ChatGPT, Gemini, and Claude). Each model receives identical input variables to ensure comparability. The models are instructed to generate categorical risk stratification (low, medium, high) along with corresponding probability estimates for BPD development.

To ensure methodological consistency, prompt engineering is standardized across all models and time points. Outputs are recorded for each weekly time window, allowing temporal comparison of predictive performance and assessment of how early postnatal data influences model accuracy.

Model outputs are subsequently compared with confirmed clinical outcomes of BPD development in the study population. Performance evaluation focuses on discriminative ability and calibration of predictions across different time points and models.

This design enables a systematic assessment of the potential role of LLM-based approaches in neonatal risk stratification and provides insight into their applicability as supportive clinical decision-making tools in neonatal intensive care settings.

• Study Type: Observational
• Enrollment (Estimated): 108

Contacts and Locations

• Study Contact: Name: Melek Büyükeren, Assoc. Prof. Dr.; Phone Number: +90532-780-30-78; Email: melekbuyukeren@gmail.com

Study Locations

• Turkey (Türkiye)
  • Konya, Turkey (Türkiye), 42020
    • Konya City Hospital, İstiklal, Adana Çevre Yolu Cd. No:135/1

Participation Criteria

Eligibility Criteria

• Ages Eligible for Study: Child
• Accepts Healthy Volunteers: No

• Sampling Method: Non-Probability Sample

Study Population

A retrospective cohort of preterm infants (<32 weeks gestation or <1,500 g) admitted to the NICU of Konya City Hospital. Only infants with complete clinical data and documented BPD status are included. Both sexes are included.

Description

Inclusion Criteria:

• Preterm infants born before 32 weeks of gestation or with birth weight <1,500 grams
• Admitted and followed in the Neonatal Intensive Care Unit (NICU) of Konya City Hospital
• Availability of complete clinical data in hospital records
• Documented bronchopulmonary dysplasia (BPD) outcome status

Exclusion Criteria:

• Presence of major congenital anomalies
• Incomplete or missing clinical data
• Death shortly after birth with insufficient follow-up data to determine BPD status

Study Plan

How is the study designed?

Number of groups / cohorts

1

Cohorts and Interventions

Group / Cohort

Intervention / Treatment

Preterm Infants Cohort (<32 weeks or <1500 g); This cohort includes preterm infants born before 32 weeks of gestation or weighing less than 1,500 grams, followed at the Neonatal Intensive Care Unit of Konya City Hospital. Clinical data from the first, second, and third postnatal weeks are retrospectively collected for analysis. No interventions are applied; AI models are used to predict BPD risk based on existing clinical data.

Other: Artificial Intelligence-Based Risk Prediction; Different large language models (ChatGPT, Gemini, Claude) will analyze retrospective clinical data to predict the risk of bronchopulmonary dysplasia (BPD). This is an observational evaluation; no experimental treatment or therapy is administered.

What is the study measuring?

Primary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Accuracy of bronchopulmonary dysplasia (BPD) risk prediction by artificial intelligence (AI) models in preterm infants.	The primary outcome is the accuracy of different large language models (ChatGPT, Gemini, Claude) in predicting BPD development. AI-generated risk predictions will be compared to actual clinical outcomes to assess prediction correctness.	Postnatal weeks 1, 2, and 3

Secondary Outcome Measures

Outcome Measure	Measure Description	Time Frame
Sensitivity and specificity of AI predictions	Evaluate the true positive rate (sensitivity) and true negative rate (specificity) of each AI model's BPD risk predictions compared to actual outcomes.	Postnatal weeks 1, 2, and 3
Comparison of prediction accuracy across postnatal weeks	Compare AI model performance at different postnatal weeks to determine if prediction accuracy improves as more clinical data becomes available.	Postnatal weeks 1, 2, and 3
Comparative performance of different AI models	Compare AI model performance at different postnatal weeks to determine if prediction accuracy improves as more clinical data becomes available.	Postnatal weeks 1, 2, and 3

Collaborators and Investigators

• Sponsor: Konya City Hospital

Publications and helpful links

General Publications

• 1. Dai D, Chen H, Dong X, Chen J, Mei M, Lu Y, et al. Bronchopulmonary dysplasia predicted by developing a machine learning model of genetic and clinical information. Front Genet. 2021;12:689071. 2. Choi HJ, Lee G, Shin SH, Lee SM, Lee HC, Sohn JA, et al. Development and external validation of a machine learning model to predict bronchopulmonary dysplasia using dynamic factors. Sci Rep. 2025;15:13620. 3. Chen Y, Ma H, Liu X. Clinical and imaging data-based machine learning for early diagnosis of bronchopulmonary dysplasia: A meta-analysis. Curr Med Imaging. 2025;21:e15734056421036. 4. Özçelik G, Erol S, Korkut S, Köse Çetinkaya A, Özçelik H. Prediction of bronchopulmonary dysplasia using machine learning from chest X-rays of premature infants in the neonatal intensive care unit. Medicine (Baltimore).2025;104:e44322. 5. Akila K, Aravind Babu LR. Deep learning-driven early prediction of bronchopulmonary dysplasia using chest X-rays and clinical data. Electronics Communications and Computing Summit. 2025;3(3):90-97. 6. Zhang X, Wang Y, Li J, et al. Development and validation of machine learning models for predicting bronchopulmonary dysplasia risk in preterm neonates based on antenatal determinants. BMC Pediatr. 2025. 7. Li Y, Wang L, Chen Z, et al. Machine learning models combining oversampling techniques for prediction of bronchopulmonary dysplasia-associated pulmonary hypertension in very preterm infants. Respir Res. 2024;25:199.

Study record dates

Study Major Dates

• Study Start (Estimated): May 1, 2026
• Primary Completion (Estimated): December 1, 2026
• Study Completion (Estimated): December 31, 2026

Study Registration Dates

• First Submitted: April 6, 2026
• First Submitted That Met QC Criteria: April 6, 2026
• First Posted (Actual): April 13, 2026

Study Record Updates

• Last Update Posted (Actual): April 15, 2026
• Last Update Submitted That Met QC Criteria: April 12, 2026
• Last Verified: March 1, 2026

More Information

Terms related to this study

• Keywords: Artificial Intelligence; preterm infant; Machine Learning; Bronchopulmonary Dysplasia
• Additional Relevant MeSH Terms: Urogenital Diseases; Female Urogenital Diseases and Pregnancy Complications; Obstetric Labor, Premature; Obstetric Labor Complications; Pregnancy Complications; Respiratory Tract Diseases; Lung Diseases; Infant, Premature, Diseases; Infant, Newborn, Diseases; Lung Injury; Ventilator-Induced Lung Injury; Congenital, Hereditary, and Neonatal Diseases and Abnormalities; Premature Birth; Bronchopulmonary Dysplasia
• Other Study ID Numbers: KSH_MB_2026_1

Plan for Individual participant data (IPD)

• Plan to Share Individual Participant Data (IPD)?: NO
• IPD Plan Description: This study uses retrospective, de-identified clinical data from preterm infants admitted to the Neonatal Intensive Care Unit of Konya City Hospital. All patient information has been anonymized to protect privacy and confidentiality. Due to the sensitive nature of neonatal health data and institutional regulations, individual participant data (IPD) will not be shared with other researchers. The study results will be reported in aggregate form only, ensuring that no identifiable information is disclosed.

Drug and device information, study documents

• Studies a U.S. FDA-regulated drug product: No
• Studies a U.S. FDA-regulated device product: No