PISTA

We present PISTA, a curated and expert-reannotated dataset of pressure injury images comprising 1,970 de-identified photographs collected from publicly accessible sources. All images were re-annotated by certified wound care nurses under a unified protocol. The dataset provides comprehensive labels, including pressure injury stage based on the six-category international staging system, tissue characteristics, recommended debridement options, and primary and secondary dressing types, defined via expert consensus and guideline-informed rules. PISTA also includes metadata such as surgical wound indicators and Chinese free-text descriptions when available. To demonstrate its utility, we evaluate several computer vision baselines for pressure injury staging and treatment-related classification. Results indicate that PISTA can facilitate the development and benchmarking of automated pressure injury management models and supports education in wound care. The dataset is publicly available on Zonodo.