Reliability of periocular anthropometry using three-dimensional digital stereophotogrammetry

Purpose Non-invasive three-dimensional (3D) stereophotogrammetry is becoming increasingly popular in many fields. However, few studies have focused on its periocular application. We aimed to provide evidence for the periocular application of a novel anthropometric procedure using 3D stereophotogrammetry by evaluating its reliability. Methods Fifty-one Caucasians were recruited (102 eyes; mean age, 31.9 ± 13.6 years). Two sets of 3D images were acquired for each subject, and two measurement sessions were performed on each image by two raters. Fifty-two periocular landmarks were identified, and then 49 corresponding linear, curvilinear, and angular measurements were evaluated for intrarater, interrater, and intramethod reliability. Results Our findings showed highly reliable results for mean absolute difference (0.59 and 0.68 unit), relative error measurement (2.66% and 3.08%), technical error of measurement (0.59 and 0.66 unit), relative technical error of measurement (2.71% and 2.96%), and intraclass correlation coefficient (0.98) for intrarater 1 and intrarater 2 reliability; respectively 0.94 unit, 4.06%, 0.89 unit, and 3.94%, as well as 0.97 for interrater reliability; and respectively 0.98 unit, 4.66%, 0.96 unit, and 4.64%, as well as 0.96 for intramethod reliability. Conclusions This imaging system and the landmark identification protocol are highly reliable. The collected measurements and their errors can be applied for the comparison of reliability among various 3D imaging systems and populations. It could be utilized for planning surgeries and evaluating treatment outcomes for physicians in ophthalmology, plastic and esthetic surgery, and in the maxillofacial field where periocular morphology alterations are made.