Standardized Three-Dimensional Lateral Distraction Test: Its Reliability to Assess Medial Canthal Tendon Laxity

The medial canthal tendon (MCT) is a stripe of fibrous tissue that insert into the orbicularis oculi muscle as a ligament [1]. Aging may cause laxity of MCT, which eventually contribute to the etiology of lower eyelid ectropion, specifically medial ectropion and the accompanying epiphora [2, 3]. Proper assessment of MCT laxity is critical to the surgery options. A deficient diagnosis of neglecting the laxity [4] and thereby incomplete correction ectropion may lead to recurrent of epiphora and redness of the inferonasal conjunctiva as a result [5].

Lateral distraction test (LDT) was the gold traditional method for assessing the laxity of MCT, which was performed by pulling the lower eyelid laterally along a horizontal direction and observing how far the lower punctum can be pulled in relation to the cornea nasal limbus [6, 7] Results may vary depending on the subject of the issue, and quantitative analysis is difficult. Several studies [4, 6‐12] have investigated the grading method of MCT laxity. However, no universally grading scale or format is accepted at present for recording the laxity, and the result is usually just noted as being present or not.

Noninvasive three-dimensional (3D) digital photogrammetry has recently gained much interest in anthropometry and is beginning to replace the classical anthropometric techniques, including the caliper and two-dimensional (2D) imaging measurements [13]. Several 3D digital imaging systems have been developed and already used successfully in craniofacial centers all over the world [14‐20]. In addition to linear distances and angles calculation, 3D imaging system dramatically offers calculation of surface curves, surface areas, volumes and volume change from the human surface. Recently, the reliability and accuracy of applying VECTRA^M3 (Canfield Scientific, Inc., Parsippany, NJ), a type of 3D stereo-photography system, in maxillofacial anthropometry has been validated in several studies [14, 21‐26]. More specifically, the feasibility and reliability of employing this 3D imaging system in the periocular region of this 3D imaging system has also been verified by previous studies [13, 27, 28].

As far as we know, the MCT laxity has never been quantified by using 3D imaging system, and there is an increasing need for developing an appropriate way to quantify it. Hence, based on the successful application of this 3D imaging technique, we developed a novel 3D lateral distraction test (3D-LDT) to modify the assessment of MCT laxity. In addition, we want to investigate the reliability and reproductivity of this 3D-LDT, so as to utilize this new technique in the clinical routine for assessing the laxity of MCT. Therefore, the purposes of this study were to investigate the reliability of 3D camera when operating the LDT, and to assess the interobserver reproducibility of LDT using 3D stereophotogrammetry.

Each participant got four pictures. The former two pictures of NP were performed by the first author (X.H.), the latter two pictures of LDT were performed by the first author and the second observer (A.C.R.), separately. We calculated five statistics (Table 2) to evaluate inter-rater, intra-rater and inter-method reliabilities [29, 30]. Inter-rater and intra-rater reliability were performed on the same images. Inter-method reliability included comparing two neutral captures by one observer and comparing LDT performed by X.H. and A.C.R. The intraclass correlation coefficient (ICC) demonstrates high reliable result when close to 1 and low reliable result when close to 0 [18]. Mean absolute difference (MAD) was computed by equally dividing the absolute difference between two measurements. The relative error measurement (REM) was computed by averaging the MAD using the grand mean of two measurements for a specific variable and multiplying the result by 100 [13]. The technical error of measurement (TEM) was calculated as \(\left( {\sqrt {{\raise0.7ex\hbox{${\left( {\sum D^{2} } \right)}$} \!\mathord{\left/ {\vphantom {{\left( {\sum D^{2} } \right)} {2N}}}\right.\kern-\nulldelimiterspace} \!\lower0.7ex\hbox{${2N}$}}} } \right)\), in which D is the difference between independent measurements and N is the count of measured subjects [14]. This statistic can be explicated similar to the standard deviation [15]. The relative TEM (%TEM or rTEM) has been introduced to compare imprecision of different variables due to the positive association between the size of measurement and TEM. It was also computed by averaging the TEM for a specific variable by its grand mean and multiplying by 100. According to the scale initiated by Andrade et al., these results were defined into five categories: less than 1%, excellent; 1–3.9%, very good; 4–6.9%, good; 7–9.9%, moderate; and more than 10% poor [31, 32].

Microsoft Excel 2016 for MAC (Microsoft Corp., Redmond, WA, USA) was applied to record all the original data of inter-landmarks measurements. Subsequently, Statistical analysis was performed by SPSS version 23 software (IBM Corp., Armonk, NY, USA). Intra-rater, inter-rater, and inter-method differences of all the measurements were calculated using this system. Paired-sample t test was performed for normally distributed data and Wilcoxon signed rank tests for paired data were applied for non-normally distributed data. P values ≤ 0.05 were considered statistically significant.

Demographic features of the study subjects are shown in Table 3. A total of 48 participants was included in our study. Study participants were between 22 and 84 years of age (55.6 + 18.6 years old). The ratio of men to women was nearly equal, and most study participants were white, non-Hispanic. Results of reliability were divided into three categories including intra-rater, inter-rater and inter-method reliability. Table 4 demonstrated the ICC and mean differences of intra-rater, inter-rater, and intra-method for all inter-landmark measurements. Estimates of MAD, TEM, REM, and %TEM is shown in Table 5. Although statistically significant differences were found between intra-rater, inter-rater, or inter-method measurements, the different magnitudes for all of them were less than 1 mm and not clinically significant.

All the anthropometric measurements for the punctum position taken on 3D images by means of the VECTRA^M3 System in NP demonstrated high reliability (Table 5), consistent with the previous studies using this system in our research group for the periocular anthropometry [13, 27, 28]. All the measurements in the LDT observed by this indirect way of VECTRA^M3 System also demonstrated high reliability, despite the relatively poor result for the distance from limbus nasal center to the lower punctum (Table 4). Digital image processing techniques have been successfully used to examine the palpebral contour in different pathologies of the eyelid position [33]. Additionally, Daniella de Paiva Almeida Stuchi et al, investigated the intra- and inter-observer reliability of a modified distraction test based on the 3D imaging technique for assessing the horizontal tension of lower eyelid [34]. However, no studies used digital images to analyze the medial canthal tension laxity so far. Moreover, previous studies on the MCT laxity were only qualitatively graded [8, 35, 36]. To the best of our knowledge, this is the first study investigating the reliability of using 3D image for analyzing the MCT laxity. Furthermore, our study firstly demonstrated the mean NP and distraction position of the lower punctum in a quantitative way in a normal population.

Our study showed that almost all measurements originated from the NP had higher reliability (MAD, REM, TEM, %TEM and ICC) than those originated from the LDT (Table 5), particularly within the intra-observer and inter- method categories. Clearly, the 3D images taken in the NP avoid the bias that may be caused by the image torsion when performing the LDT. Images taken in NP provided the overall better precision than in the distraction test. These results suggest that the VECTRA^M3 System is capable of excellent reliability for the evaluation the punctum position for the normal population or patients with lower eyelids diseases. More specifically, the results could be introduced to further evaluation for lower eyelid diseases such as ectropion especially for patients with medial ectropion or symptom of epiphora, and so on.

For all the digital images, we set two different distance (Pc-Pu and Ln-Pu) according to two reference vertical lines, with one across the pupil center (Pc) and another across the corneal limbus nasal center (Ln). Previous studies [8, 11, 37] normally used the vertical line across the corneal limbus nasal center to define the resting position or the distraction distance of the lower punctum in the direct measurements. However, in the indirect way of using 3D images [13, 27, 28, 34], pupil center was the basement for almost all the landmarks. Hence, we developed a reliability comparison between the two reference vertical lines. In our study, measurements for both vertical lines demonstrated high score of ICC in intra- rater, inter-rater and inter-method, despite the significant difference of nPc-Pu for inter-methods reliability and n*Pc-Pu for intra-rater reliability, which were not clinically significant (≤ 1 mm) for all of them (Table 5). Additionally, no significant difference was found between them in NP for MAD, REM, TEM and %TEM in intra- rater, inter-rater and inter-method reliability (Table 4). Admittedly, the score of inter-methods reliability for n*Pc-Pu and dPc-Pu´ was relatively higher than those of intra- and inter-rater reliability, despite individual observers having trained together. Nevertheless, in the distraction test, the overall reliability of the distance according to limbus nasal center was lower than that of pupil center, which indicated that the distraction distance calculated from the pupil center was more reliable and repeatable for the evaluation of the MCT laxity than that from the limbus nasal center to the lower punctum (Table 4).

Furthermore, our study demonstrated the possibilities in the application for different pathological conditions. Evaluation of lower eyelid malpositions plays a vital role in the diagnosing of eyelid disease as well as the designation of surgeries, and the follow-up project is also essential for the evaluation of treatment effects. However, the main evaluation of lower eyelid malpositions, especially the medial canthal malposition is not quantitative due to the limited method [38]. In our study, with the application of the 3D stereophotography and our landmarks system, the pre-operative malposition could be evaluated quantitatively (Fig. 3), and the post-operative follow-up could be recorded for quantitative observation and comparison (Fig. 4). Additionally, the findings of the present study build the essential basis to introduce this novel 3D-LDT in the clinical routine and to conduct follow-up projects investigating variations of repair techniques and their failure rate.

The possible limitation of this study might be the challenge to keep the participants looking into the same position when performing the LDT, especially for the sensitive ones. For optimal image, an individual being captured must keep a consistent head position and visual focus direction. To minimize errors caused by the processing of LDT, repeated captures and cooperative with another operator are necessary.

In conclusion, the high intra-rater, inter-rater and inter-method reliability found in this study was due to the easy application of the technique and analysis of the results. Our novel standardized three-dimensional lateral distraction test (3D-LDT) seems to be a feasible and highly reliable method to assess medial canthal tendon (MCT) laxity. Furthermore, pathological eyelid conditions were not included in this study for the identification error of all the landmarks. Therefore, follow-up studies are necessary to investigate the reliability of this 3D-LDT on pathological cases as well as to evaluate the correlation of these measurements with subjective patients’ complaints and various disease severity of eyelid abnormalities. Lastly, to our best knowledge, this is the first study quantifying the reliability of the 3D-LDT for evaluating the MCT laxity based on digital image processing.