Pixaire1 study: evaluation of automated chronic wound surface measurement systems (Maxant et al. 2026)

A cross-sectional validation study conducted on 42 patients comparing two Pixacare AI modules — WoundTrack (semi-automated) and WoundSize (automated) — against the reference method (digitized planimetry). The semi-automated module shows excellent agreement with the reference (ICC: 0.96, p < 0.001), supporting the integration of a clinician validation step in AI-based wound measurement.

Publication

Study posted as a preprint on medRxiv on March 31, 2026. DOI: 10.64898/2026.03.30.26344793. Access the publication.

Objective

To evaluate two smartphone-based methods for measuring chronic wound surface area:

• WoundTrack (WT): semi-automated measurement
• WoundSize (WS): fully automated measurement

Both methods were compared with the reference technique: digitized planimetry (PL).

Study design

Cross-sectional, single-center, open-label study. Comparative validation versus reference method, conducted by two independent experts to assess both precision (inter-rater consistency) and accuracy (agreement with the reference).

Materials and methods

The study includes 42 patients enrolled from May to June 2023. Wound surfaces were measured using the three methods (WoundTrack, WoundSize, digitized planimetry) by two independent experts.

The statistical analysis was structured in four steps:

• Multivariate analysis of variance
• Assessment of precision: correlation between the two experts
• Assessment of accuracy: agreement between each evaluated method and the reference
• Analysis of non-conformities (differences greater than 20% in absolute value compared to planimetry), on a subgroup of wounds smaller than 8 cm²

Of the 42 patients included, 6 were excluded from the statistical analysis (4 multiplanar wounds, 2 cases with difficult edge delineation).

Results

Key findings are:

• Excellent agreement between the semi-automated WoundTrack module and planimetry: ICC = 0.96 (p < 0.001)
• Automated WoundSize module: strong overall performance, with limitations identified on small wounds
• No significant difference in the multivariate analysis of variance between methods
• Results support the integration of a clinician validation step in the automated measurement workflow

The study confirms that smartphone-based AI measurement of wound surface area is a reliable approach when combined with expert clinician validation, particularly for small wounds.

Keywords

Chronic wound, surface measurement, artificial intelligence, WoundTrack, WoundSize, planimetry, clinical validation, digital medical device