Multimodal phenotyping of ARDS: design and preliminary insights from the prospective BIOWARE cohort for precision critical management

BACKGROUND: Acute respiratory distress syndrome (ARDS) is a highly heterogeneous syndrome with substantial mortality, for which existing subphenotyping strategies based on single-modal data have limited ability to guide targeted therapies. A comprehensive, multimodal framework is urgently needed to decipher its complexity and advance precision care. METHODS: The BIOWARE study is a prospective, multicenter cohort aiming to enroll 2,000 ARDS patients across nine Chinese centers. The protocol integrates longitudinal data from clinical assessments, ventilator waveforms, advanced imaging (CT, EIT, lung ultrasound), and biospecimen analyses. FINDINGS: As of August 2025, 169 patients have been enrolled (median age 62 years, 74% male). Core clinical and imaging data were successfully acquired across all centers, confirming the feasibility of the multimodal collection protocol. While baseline biospecimen collection was complete (100% for plasma and BALF at Day 1), follow-up sampling decreased at later timepoints (Day 7 BALF: n = 24). Some specialized parameters (e.g., P0.1) showed higher missing rates due to clinical constraints, reflecting real-world implementation challenges. INTERPRETATION: The BIOWARE cohort provides a multidimensional framework that captures dynamic interactions among physiology, pulmonary structural changes, and host response, enabling identification of ARDS endotypes grounded in biological mechanisms. This integrative strategy represents a paradigm shift from syndrome-based classification toward mechanism-driven precision care, establishing a transformative platform for targeted therapeutic development.