Three-dimensional stereophotogrammetry for periocular tumors: Reproducibility and clinical feasibility

This study aimed to evaluate the intra- and inter-rater reproducibility of periocular tumor measurements using three-dimensional (3D) stereophotogrammetry and to explore its potential clinical feasibility in periocular tumor assessment. Standardized 3D facial images were obtained from 150 patients with a total of 175 periocular tumors using the Vectra M3 imaging system. Surface areas were independently measured by two raters. Intra- and inter-rater reliability were assessed using intraclass correlation coefficients (ICC), mean absolute deviation (MAD), technical error of measurement (TEM), relative error of measurement (REM), and relative TEM (rTEM). Key factors influencing measurement reliability were identified through LASSO regression and random forest analysis. Overall measurement reliability was excellent, with intra-rater and inter-rater ICCs of 0.998 and 0.974, respectively. The mean intra-rater and inter-rater MAD were 0.63 mm² and 0.40 mm², while TEMs were 2.29 mm² and 7.81 mm². Intra-rater and inter-rater REM values were 1.94% and 1.22%, respectively. Tumors >5 mm showed significantly higher reliability than tumors ≤5 mm (p = 0.010). Tumors with well-defined boundaries had superior reproducibility compared to those with unclear margins (p = 0.026). Localization influenced reliability, with lateral canthus tumors showing the highest consistency (p = 0.008). Tumor color also affected reliability, with brownish-black tumors exhibiting the greatest reproducibility (p = 0.021). Three-dimensional stereophotogrammetry demonstrated high reproducibility for periocular tumor assessment. These findings suggest its potential clinical applicability as an adjunctive tool to support preoperative planning and longitudinal monitoring.