SIMFONI Pipe Cleaner 2
Main Applicant – Dr Wesley Yeung, Associate Consultant, National University Heart Centre, National University Health System (NUHS)

Building on the foundation established in Pipe Cleaner 1 (PC1), which demonstrated that AI model fine-tuning can be successfully conducted within the enTRUST platform using publicly available datasets, SIMFONI Pipe Cleaner 2 (PC2) takes the next critical step: adapting medical foundation AI models using real-world clinical data from Singapore’s healthcare system.

PC2 will transfer de-identified electronic health record (EHR) data from the National University Health System (NUHS) into the enTRUST platform and use this data to fine-tune a medical domain foundation model. This represents a significant advancement over PC1, as the model will now learn from clinical data that reflects Singapore’s unique patient demographics, disease profiles, and care practices rather than from publicly available overseas datasets.

The goal of the Singapore Medical Foundational AI Model Programme (SIMFONI) is to develop AI models that are genuinely adapted to Singapore’s healthcare context. A model trained on local clinical data is better positioned to support clinicians in making accurate, contextually relevant decisions. If successful, PC2 will demonstrate that enTRUST can serve as a trusted, secure environment for national-scale medical AI development, paving the way for broader adoption across Singapore’s public health clusters.