A robotic wound care patient for evidence-based surgical site infection research

Surgical site infections (SSIs) are among the most common and preventable postoperative complications, yet existing preclinical models lack physiological realism and do not enable quantitative assessment of bacterial behavior. Wound pH critically modulates bacterial morphology and organization, underscoring the need for systems that replicate controlled wound environments.

To develop and validate a robotic wound care patient (RWCP) that reproduces SSI-relevant physical and biological conditions, and to quantify pH-dependent bacterial morphology and spatial organization on wound dressings using automated deep-learning image analysis.

A life-sized abdominal RWCP integrating layered soft-tissue simulants, respiration simulation, controlled exudate delivery, and a laparotomy incision was engineered. Simulated wound fluid inoculated with Lactobacillus delbrueckii subsp. bulgaricus was delivered at pH 5.8 (acidic) or pH 6.8 (mildly acidic). Dressing samples were imaged with SEM, and bacterial morphology and topology quantified using a Cellpose-based deep-learning model, FIJI macros, and Python algorithms. Outcome measures included bacterial count, area coverage, circularity, roundness, aspect ratio, chain number, and bacteria per chain.

Acidic pH increased bacterial counts by ∼45% and produced morphological elongation (circularity and roundness ↓; aspect ratio ↑). Topological analysis identified nearly fourfold more bacterial chains and larger assemblies under acidic conditions (p ≤ 0.02), indicating enhanced cooperative aggregation.

The RWCP provides a physiologically relevant, reproducible platform for SSI research, enabling sensitive detection of pH-driven bacterial morphological and organizational adaptations. This integrated mechanical-biological system offers a robust preclinical tool for evaluating wound care technologies and informing evidence-based SSI prevention strategies.