Repeatability and agreement analysis are fundamental investigative tools to assess the performance within and between measurement devices. Novel technologies are usually compared with the gold standard. The RSC technology is an established method to visualize the anterior segment of the eye. The RSC (Pentacam HR) has an extensive software package that allows numerous specialized analyses for keratoconus detection [19, 20] and progression [21]. The SS-OCT (ANTERION) on the other hand, is a novel device for anterior segment analysis, including a specific mode called “Cornea App” for corneal tomography measurements. This study aimed to compare both devices concerning repeatability and agreement of various corneal tomography parameters in normal and KC eyes.

The main finding of the current study revealed that the repeatability of anterior corneal parameters was comparable between RSC and SS-OCT in normal eyes with a slightly better tendency towards the RSC. In contrast, the repeatability was better using the SS-OCT in mild and moderate KC eyes. Further, the repeatability of most posterior corneal parameters was better with SS-OCT than with RSC in all subgroups. The offsets between both devices were mostly statistically significant and differed between the subgroups, ruling out interchangeability between the devices. The mean difference for Kmax was especially high in moderate KC eyes. Furthermore, wide LoA were also observed. To the best of our knowledge, this study is the first to compare RSC (Pentacam) and the novel SS-OCT (ANTERION) in normal and KC eyes with regards to the repeatability and agreement of corneal tomography parameters.

In this study, we found that higher keratoconic disorders affect repeatability of both devices negatively. In comparison to normal eyes, the repeatability of flat SimK and steep SimK was elevated in KC eyes for both devices. A similar finding was observed by Kreps et al. for RSC, in which repeatability was 0.16 D (flat SimK) and 0.2 D (steep SimK) in normal eyes and 0.63 to 0.87 D (flat SimK) as well as 0.56 to 0.82 D (steep SimK) in mild and moderate KC eyes, respectively. Here, similar criteria for subgroups were used. The repeatability of RSC was comparable for flat SimK (0.2 D) and steep SimK (0.2 D) in normal eyes. We also found elevated values for repeatability for mild (0.5 and 0.8 D) and moderate (0.5 and 0.8 D) KC eyes for flat SimK and steep SimK, respectively, which were like those reported by Kreps et al. [11].

Tañá-Rivero et al. investigated the repeatability of SS-OCT (ANTERION) in normal eyes and found the repeatability lower than 0.25 D for flat SimK, steep SimK and average SimK [22]. In this study, repeatability was found to be comparable (between 0.25 and 0.28 D). The repeatability of astigmatism was 0.18 D [22], which was lower than our findings (r = 0.23 D).

For Kmax, the repeatability was comparable for both devices in normal eyes. The repeatability of Kmax was 0.37 D, which was lower than reported values of Kosekahya et al. (r = 0.55 D) in normal eyes using the RSC [23]. Further studies which have used the RSC also found higher values for repeatability (0.47 and 0.8 D, respectively) [11, 17].

In the management of KC, the assessment of the maximum corneal curvature (Kmax) of the anterior corneal surface is commonly used to determine the disease’s progression. A change in Kmax of 1 D within 6 or 12 months is reported as a clinically significant progression [7, 8]. Therefore, the repeatability of Kmax should be lower than the defined border of clinical progression. In our study, repeatability increased for Kmax for both devices, however, the magnitude was much higher in RSC than in SS-OCT. An increase of repeatability indicates a poorer reliability. In mild KC eyes, the repeatability of RSC and SS-OCT were 0.62 and 0.5 D, respectively. In moderate KC eyes, the repeatability of RSC was 1.3 D compared to 0.8 D measured by SS-OCT indicating a lower reliability of the RSC in these higher stages of KC. Previously, several studies reported that as the repeatability increases, the higher the stage of KC is, when using RSC [11, 14, 23, 24]. In these studies, the values for repeatability ranged from 0.51 to 0.81 D, from 1.04 to 1.19 D, and from 1.34 to 1.66 D for early, moderate, and advanced KC, respectively [11, 14, 23]. The overall repeatability was reported between 0.99 [23] and 1.12 D [14]. Some of these studies concluded that the repeatability of Kmax in different stages of KC should be considered in clinical practice [11, 14]. In contrast, the repeatability of Kmax was found to be less than 1 D in KC eyes using SS-OCT in our study, leading us to conclude that a change of 1 D in Kmax as measured by SS-OCT can be used as an indicator of KC progression, but not in RSC. The steep SimK is another parameter that is considered to describe the disease’s progression [25,26,27]. Here, steep SimK showed repeatability lower than 1 D for mild and moderate KC for both devices. Moreover, there was no statistically significant offset observed in moderate KC leading to the assumption that this parameter could be useful for defining progression, even between different devices.

The Casia 2 (Tomey Corp., Nagoya, Japan) is another commercially available SS-OCT assessing corneal tomography. In a previous study, it was found that the standard deviation of five consecutive measurements increased with the severity of KC using both RSC and SS-OCT (Casia 2), while the outcome was in favor of RSC, except for very advanced stages of KC [24]. In contrast, our results revealed that repeatability of the anterior corneal parameters were in favor of the SS-OCT (ANTERION) in both KC subgroups with highest differences for Kmax.

The Bland-Altman analysis revealed that a (significant) negative offset was found for SimK values indicating higher SimK readings by SS-OCT in normal and mild KC eyes, except from Kmax. In contrast, SimK readings were lower for SS-OCT in moderate KC eyes than for RSC (positive offset). The LoA were extended the higher the stage of KC (greater than 1 D). The statistically significant offset does not allow interchangeability of anterior parameters in normal as well as in KC eyes.

The assessment of the posterior corneal surface became a major role in diagnosing KC and monitoring progression since it is measurable using Scheimpflug tomography [21, 28]. As mentioned, the posterior corneal curvature was considered an additional parameter for evaluating the progression of KC [10, 29]. Tellouck et al. and Fujimoto et al. showed that posterior steepening occurred earlier than anterior steepening, indicating an earlier detection of progression [30, 31]. Therefore, precise measurements of posterior corneal parameters are necessary for determining an ectatic progression.

Here, posterior corneal surface parameters demonstrated better repeatability when measured by SS-OCT than by RSC in all subgroups. The repeatability of RSC was consistently higher for flat SimK, steep SimK, average SimK, astigmatism, and Kmax in comparison to SS-OCT in normal and KC eyes. Szalai et al. found an increase of repeatability for flat SimK (0.16 vs. 0.55 D), steep SimK (0.18 vs. 0.51 D) and average SimK (0.18 vs. 0.55 D) between normal and KC eyes using RSC [32]. Furthermore, Kreps et al. reported repeatability values for central posterior SimK readings of 0.1 D in normal eyes and between 0.13 and 0.35 D in mild and moderate KC [11]. These observations were comparable to our study results, whereas repeatability values were lower compared to Szalai et al. For SS-OCT, Tañá-Rivero et al. observed similar values to those found in our study. Moreover, Flockerzi et al. found that posterior corneal parameters showed better repeatability in both healthy and KC eyes using the Casia 2 compared to the RSC [24].

The repeatability of BFS did not differ between the devices with respect to normal eyes, whereas lower values were found for SS-OCT in mild and moderate KC. The results were in tandem with previous investigations [11, 14, 22].

The Bland-Altman analysis revealed a non-significant offset as well as a range of LoA close to 1 D between both devices in mild and moderate KC eyes for posterior flat SimK, steep SimK, and Kmax, indicating good agreement. However, the range of LoA of Kmax was 1.66 and 2.27 D for mild and moderate KC, respectively, which cannot be assumed as interchangeable.

A high repeatability was found for CCT and MCT in normal and KC eyes for both devices. The maximum of repeatability was 8.4 and 4.8 µm for RSC and SS-OCT, respectively. The CV of the measurements was < 1%. Three studies considered a corneal thickness reduction as a factor for the disease’s progression if the loss of MCT was more than 20 µm [33], 10 µm [34], or more than 5% [25]. These limits seem to be appropriate, as the measurement variability was found to be lower than 10 µm and 1% for both devices in the current study. Previously, several studies reported higher values of repeatability in normal eyes [32, 35] and KC eyes [11, 14, 36, 37], where the 10 µm criteria would not be met. Overall, the SS-OCT showed lower values of repeatability indicating higher reliability. Similar results were found comparing the SS-OCT (Casia 2) with the RSC [24]. Therefore, it is assumed that SS-OCT might be more appropriate for assessing corneal thinning in the progression period of KC. An offset between 1.7 and 7.5 µm was observed in this study between RSC and SS-OCT, where RSC measured a higher corneal thickness consistently, which was statistically significant in normal and moderate KC eyes. In addition, the wide range of more than 20 µm of LoA excludes interchangeability between the devices. Moreover, there is currently no consensus, if different technologies, such as OCT, Scheimpflug or ultrasound, when assessing corneal thickness, are interchangeable or if conversion factors are applicable. Several reports demonstrated a higher measured corneal thickness by RSC compared to OCT devices [24, 38, 39], whereas interchangeability is reported as well [40]. These discrepancies might be due to the different investigated populations and dependent on age and the presence of ocular pathologies [41]. All in all, assessing corneal thickness should be performed using the same device to evaluate changes accurately in patients.

Scheimpflug (RSC) technology was also compared with SS-OCT in corneas with specific properties following INTACS implantation as well as Fuchs endothelial corneal dystrophy (FECD). Matar et al. reported after INTACS implantation, in KC eyes, there was a higher repeatability of Kmax using the RSC (> 1 D) than using the Casia 2 SS-OCT (< 1 D) [42]. Furthermore, both the SimK and corneal thickness were found to be higher with RSC than SS-OCT, which we attributed to the different measuring methods and the insufficient detection of corneal curvature due to higher reflectivity of the implant [42]. We concluded that both devices are eligible to assess corneal tomography for long-term follow-ups, however, interchangeability is not advised [42]. Comparing the results of our study, SimK values and corneal thickness were found to be higher in KC eyes, especially in moderate KC, measured by RSC compared to SS-OCT. The findings of the repeatability of Kmax were equivalent to the findings of Matar et al. comparing both technologies. Moreover, a steeper anterior corneal curvature, a flatter posterior curvature, and a thinner corneal thickness were determined by SS-OCT (ANTERION) in comparison to RSC in FECD eyes [43].

According to the technical aspects, the better repeatability of the SS-OCT compared to the RSC might be not explained by the available scans or data points (16,640 scans vs. 138,000 elevation points). Instead, it can be assumed that the number of radial B-scans (65 vs. 25 images of the RSC), the shorter scanning time (< 1 s vs. 1 s of the RSC), and the live eye tracking system result in a more precise measurement of the SS-OCT.

Based on the results of the current study, the SS-OCT seems to be a better tool for assessing corneal tomography (SimK and corneal thickness) in follow-up examinations in normal eyes as well as mild and moderate KC eyes. Currently, the SS-OCT (ANTERION) is limited in providing a specific ectasia screening tool. As mentioned before, RSC (Pentacam) provides a wide range of screening tools, especially for corneal ectasia, keratoconus detection as well as progression. For these applications, the RSC is very helpful in clinical practice. Further, the repeatability is adequate for measuring normal and mild KC eyes.

The study is limited by a non-randomized order of the performed measurements. RSC was always measured before SS-OCT indicating that a certain learning effect could not be avoided, vice versa, patients could also become more tired after these many repeated measurements. Moreover, the age was not matched properly between all subgroups along with the normal eyes and mild KC group involving more female cases (37%) than the moderate KC group (10%). This might influence the results.