Site-Level Fine-Tuning with Progressive Layer Freezing: Towards Robust Prediction of Bronchopulmonary Dysplasia from Day-1 Chest Radiographs in Extremely Preterm Infants

2025-07-16

AI Summary

This study addresses the challenge of early prediction of bronchopulmonary dysplasia (BPD) in extremely low birth weight preterm infants, aiming to avoid high-risk prophylactic interventions for low-risk neonates. We propose a deep learning framework leveraging chest radiographs acquired within 24 hours of birth. Our method introduces a novel fine-tuning strategy combining progressive layer freezing with linear probing, applied to a ResNet-50 backbone pretrained in-domain on adult chest X-rays. To enhance generalizability and robustness under limited medical data, we integrate CutMix augmentation and discriminative learning rates. On moderate-to-severe BPD prediction, the model achieves an AUROC of 0.78±0.10, balanced accuracy of 0.69±0.10, and F1-score of 0.67±0.11—outperforming both ImageNet-initialized baselines and conventional radiographic scoring systems. The framework balances computational efficiency with predictive performance, enabling scalable multi-center deployment and federated learning.

Technology Category

• Machine Learning: Deep Neural Architectures and Foundation Models
• Computer Vision: Learning & Optimization for CV
• Search and Optimization: Learning to Search

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Web Mining and Content Analysis: Large pretrained models with web data
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications

Abstract

Bronchopulmonary dysplasia (BPD) is a chronic lung disease affecting 35% of extremely low birth weight infants. Defined by oxygen dependence at 36 weeks postmenstrual age, it causes lifelong respiratory complications. However, preventive interventions carry severe risks, including neurodevelopmental impairment, ventilator-induced lung injury, and systemic complications. Therefore, early BPD prognosis and prediction of BPD outcome is crucial to avoid unnecessary toxicity in low risk infants. Admission radiographs of extremely preterm infants are routinely acquired within 24h of life and could serve as a non-invasive prognostic tool. In this work, we developed and investigated a deep learning approach using chest X-rays from 163 extremely low-birth-weight infants ($leq$32 weeks gestation, 401-999g) obtained within 24 hours of birth. We fine-tuned a ResNet-50 pretrained specifically on adult chest radiographs, employing progressive layer freezing with discriminative learning rates to prevent overfitting and evaluated a CutMix augmentation and linear probing. For moderate/severe BPD outcome prediction, our best performing model with progressive freezing, linear probing and CutMix achieved an AUROC of 0.78 $pm$ 0.10, balanced accuracy of 0.69 $pm$ 0.10, and an F1-score of 0.67 $pm$ 0.11. In-domain pre-training significantly outperformed ImageNet initialization (p = 0.031) which confirms domain-specific pretraining to be important for BPD outcome prediction. Routine IRDS grades showed limited prognostic value (AUROC 0.57 $pm$ 0.11), confirming the need of learned markers. Our approach demonstrates that domain-specific pretraining enables accurate BPD prediction from routine day-1 radiographs. Through progressive freezing and linear probing, the method remains computationally feasible for site-level implementation and future federated learning deployments.

Problem

• Predicting bronchopulmonary dysplasia from early chest X-rays
• Improving BPD prognosis to avoid unnecessary infant toxicity
• Enhancing deep learning accuracy with domain-specific pretraining

Innovation

• Progressive layer freezing prevents overfitting
• Domain-specific pretraining improves BPD prediction
• CutMix augmentation enhances model performance