Evaluation of epithelial, stromal, and corneal thickness maps in the distinction between pellucid marginal degeneration and inferior keratoconus (pellucid-like) Authors: Büşra Dilara Yıldırım Erdal, Burcu Kazancı Published: 2026 ABSTRACT BackgroundTo distinguish between these two ectasias, which can be confused with each other, using epithelial, stromal, and total corneal thickness maps in patients with pellucid marginal degeneration (PMD) and inferior keratoconus (I-KC) that have similar topographic features.MethodsThis prospective cross-sectional study included 114 eyes of 73 patients (56 eyes with PMD, 58 eyes with I-KC). All measurements were obtained using MS-39 AS-OCT. The values in the epithelial, stromal, and corneal thickness maps were recorded.ResultsThere was no significant difference in mean age between the two groups (p = 0.061). In PMD patients, minimum epithelial, stromal, and total corneal thickness values in the inferotemporal (IT), inferior (I), inferonasal (IN), and nasal (N) zones within 6–8 mm and 8–10 mm areas were significantly thinner than in the I-KC group (p < 0.05). In the I-KC group, minimum and average epithelial thickness values in the central 2 mm area, in the IT and temporal (T) zones within 2–4 mm areas, and in the IT zone within 4–6 mm areas were significantly thinner than in the PMD group (p < 0.05). To distinguish PMD from I-KC, the IT 2–4 min./IN 8–10 min. difference in stromal thickness had an AUC of 0.878, with 79.5% sensitivity and 76% specificity; in epithelial thickness, the IT 2–4 min./IN 6–8 min. ratio reached the highest diagnostic accuracy, with an AUC of 0.855, 78.6% sensitivity, and 79.7% specificity (p < 0.001).ConclusionsPatients exhibiting crab-claw patterns on the sagittal curvature map can differentiate between PMD and I-KC using pachymetric values and combinations of epithelial, stromal, and corneal thickness maps obtained with wide-field AS-OCT. Introduction Pellucid marginal degeneration (PMD) is a non-inflammatory progressive corneal ectasia of unknown etiology [1]. It is an ectasia with bilateral and asymmetric involvement, characterized by a band-like corneal thinning in the inferior cornea between 4 and 8 o’clock, usually 1–2 mm from the limbus [2]. Differentiating PMD from other corneal ectasias is necessary for treatment, follow-up, and prognosis. Compared to PMD, keratoconus has a higher frequency, an earlier onset, rapid progression, and a higher risk of hydrops [1]. However, studies claim the two are different entities with the same disease [3, 4]. The crab-claw, butterfly, or kissing dove patterns typical for PMD can be seen in patients with inferior keratoconus (I-KC) [4]. These cases have been described as pellucid-like keratoconus, a subtype of I-KC [5]. However, evaluating only the sagittal curvature map may lead to errors in distinguishing these two entities. In the literature, biomicroscopic examination findings, corneal topography and tomographic maps, corneal density, corneal biomechanical properties, and corneal aberrations have been evaluated to date for distinguishing PMD from I-KC (pellucid-like keratoconus) [3, 6–8]. However, other analyses and imaging methods with high specificity and sensitivity that provide definitive distinctions are needed. Anterior segment optical coherence tomography (AS-OCT) is a unique technology that provides a wide range of morphological and quantitative analyses of the corneal layers, in addition to the topographic and tomographic data from Scheimpflug-based corneal tomography devices, which have been increasingly used in recent years. AS-OCT provides epithelial and stromal maps and is an important guide for evaluating disease pathophysiology, differential diagnosis, clinical course, and planning surgical and laser interventions. Reinstein et al. reported that corneal epithelial thickness is increased in the area surrounding the cone when examined at the cone apex in patients with keratoconus, thus creating a classic pattern (doughnut pattern) on the epithelial map [9]. However, many studies in the literature show that epithelial thickness maps in patients with keratoconus differ from those in normal eyes and exhibit distinctive features across different stages and subtypes, including subclinical and forme fruste keratoconus [10–12]. A limited number of studies have compared epithelial thickness between PMD patients and those with keratoconus. However, these studies did not include a sufficient number of PMD patients [13]. Our study aimed to examine the differences between PMD and I-KC, which have similar sagittal curvature maps. We will examine not only the mean epithelial thickness but also the minimum epithelial and stromal thicknesses, and use these measurements to distinguish the two entities. Materials and methods Our single-center, observational, cross-sectional study was conducted among patients followed in the Cornea-Contact Lens Unit of the Ophthalmology Department at a tertiary referral center hospital. The study was conducted within the framework of the Helsinki Declaration, and the Ethics Committee approved the study protocol (no: 754/2024). Written informed consent was obtained from all participants. Study design Patients who were monitored in our cornea-contact lens unit from 2014 to 2024 due to PMD and I-KC were invited for examination and anterior segment OCT evaluation. This study included 114 eyes of 73 patients. Among these patients, 34 had PMD (56 eyes), while 39 had I-KC (58 eyes). In patients diagnosed with PMD, the following factors were considered during evaluation: absence of a Fleischer ring, Vogt striae, vascularization, or lipid accumulation on biomicroscopic examination; presence of crab-claw, kissing doves, or butterfly patterns in the sagittal curvature map on corneal tomography; presence of band-like corneal thinning inferiorly on pachymetric evaluation; the appearance of a protrusion just above the thinning area; presence of a bell-like appearance in the pachymetry map; and patients not being in the first two decades of life at diagnosis [2, 7]. In patients diagnosed with I-KC, the following factors were considered during evaluation: age 30 years or older (to obtain an age group compatible with PMD), a Fleischer ring, Vogt striae, or a conical protrusion observed during biomicroscopic examination, pellucid-like crab-claw, kissing doves, or butterfly patterns observed in the sagittal curvature map on tomography, corneal protrusion from the thinning area, and absence of a bell-like appearance on the pachymetry map [4, 5]. In addition, patients with stage 2 or higher were included according to the Amsler-Krumeich classification [14]. Patients’ past hospital records indicated that they had been diagnosed with keratoconus within the first three decades. In both groups, patients with corneal scars (sequelae of hydrops or apical scars), moderate or severe dry eye, a history of corneal surgery (including corneal crosslinking), and a history of corneal disease other than PMD or keratoconus (corneal dystrophies, sclerosing conjunctivitis, inflammatory corneal diseases) were excluded. Patients in the I-KC group without crab-claw, kissing doves, or butterfly patterns on their axial sagittal curvature maps were also excluded. In addition, patients who used contact lenses were re-examined and reimaged after at least three weeks. All patients underwent a complete ophthalmologic examination. Best-corrected visual acuity (with glasses) was assessed using the logMAR chart. General corneal topographic and tomographic features were assessed with MS-39. Analyses of topographic, tomographic and corneal epithelial and stromal maps of the cornea with MS-39 combined placido disc and high-resolution AS-OCT MS-39 AS-OCT (software Phoenix version 4.1.1.5;CSO, Frienze, Italy), which combines a Placido disc and spectral domain optical coherence tomography (SD-OCT) system. MS-39 acquires slices with an axial resolution of 3.6 μm from a 16 mm area using an infrared superluminescent diode (SLD) source emitting light at a wavelength of 850 nm. SD-OCT scanning consists of 25 B scans, consisting of 1024 A scans in 16 mm slices at a time interval of approximately 1 s. In addition, during the OCT scan, two frontal images of the eye, one for keratoscopy and the other for the iris, are obtained with a field of view of approximately 14.1 × 10.6 mm. In this way, axial sagittal curvature map, pachymetry, and sublayer corneal thickness maps, anterior and posterior elevation maps can be obtained. In our study, all AS-OCT images were taken in the “12 × 5@10 mm” mode because it has better resolution. Also, kmax, k1, k2, and simk values; maximum anterior and posterior elevation values; and frontal astigmatism values in the 6 mm and 3 mm areas were obtained from these maps and the summary data. Some symmetry, curvature, and ectasia indices used in keratoconus screening; symmetry index front (SIf), symmetry index back (SIb), Central-Surrounding index front (CSIf), Central-Surrounding index back (CSIb), ectasia index front (EIf), ectasia index back (EIb) were also obtained from the keratoconus screening data section of the MS-39 device. Corneal epithelial thickness, stromal thickness, and total corneal thickness maps, created from MS-39 AS-OCT sections, divide a 10 mm circular region into 33 areas, with automatic segmentation in the improved version. After dividing the device into 5 areas of central 2 mm, 2–4 mm, 4–6 mm, 6–8 mm, 8–10 mm, it divides the other four areas except the central zone into eight more zones. These are superior (S), nasal (N), temporal (T), inferior (I), superonasal (SN), superotemporal (ST), inferonasal (IN), and inferotemporal (IT) zones. In each zone, 3 values are automatically calculated as minimum (min.), average (avg.), and maximum (max.) (Fig. 1). Fig. 1Representative corneal tomography maps comparing Pellucid Marginal Degeneration (PMD) and Inferior Keratoconus (I-KC/Pellucid-like Keratoconus). (A) Axial sagittal curvature map, epithelial thickness map, stromal thickness map, and total corneal thickness map of a representative PMD patient, demonstrating the characteristic “crab claw” / “kissing doves” pattern on the axial sagittal curvature map with inferior peripheral steepening and relative central flattening, accompanied by diffuse epithelial and stromal thinning localized to the inferior peripheral cornea. (B) Corresponding tomographic maps of a representative I-KC (pellucid-like keratoconus) patient, showing a similar inferior steepening pattern on the axial sagittal curvature map; however, the epithelial thickness map reveals focal central epithelial thinning with compensatory peripheral thickening, while the stromal and total corneal thickness maps demonstrate inferior paracentral thinning — a distribution pattern distinctly different from the peripheral thinning profile seen in PMD. The comparative evaluation of these curvature and thickness maps plays a critical role in the differential diagnosis of PMD and I-KC (pellucid-like keratoconus) Representative corneal tomography maps comparing Pellucid Marginal Degeneration (PMD) and Inferior Keratoconus (I-KC/Pellucid-like Keratoconus). (A) Axial sagittal curvature map, epithelial thickness map, stromal thickness map, and total corneal thickness map of a representative PMD patient, demonstrating the characteristic “crab claw” / “kissing doves” pattern on the axial sagittal curvature map with inferior peripheral steepening and relative central flattening, accompanied by diffuse epithelial and stromal thinning localized to the inferior peripheral cornea. (B) Corresponding tomographic maps of a representative I-KC (pellucid-like keratoconus) patient, showing a similar inferior steepening pattern on the axial sagittal curvature map; however, the epithelial thickness map reveals focal central epithelial thinning with compensatory peripheral thickening, while the stromal and total corneal thickness maps demonstrate inferior paracentral thinning — a distribution pattern distinctly different from the peripheral thinning profile seen in PMD. The comparative evaluation of these curvature and thickness maps plays a critical role in the differential diagnosis of PMD and I-KC (pellucid-like keratoconus) Statistical analysis SPSS version 22 (IBM SPSS Statistics V22.0) was used to evaluate the values obtained from the study to form tables. The normality analysis of the data was determined according to the Kolmogorov-Simirnov test or the Shapiro-Wilk test. In parametrically distributed data, the independent sample t test was applied to compare two groups. As a result of this test, the parameters with significant p values were determined. These parameters were evaluated with ROC curve analysis and their diagnostic values were determined in the distinction between PMD and inferior keratoconus. The parameters with high diagnostic values were combined as difference and ratio and the parameters with higher definitiveness were re-evaluated with ROC curve analysis. The maximum summarization of sensitivity and specificity was used to determine appropriate clinical cut-off values. The data were presented as mean standard deviation (SD). P values less than 0.05 were considered significant. Sample size estimation The estimated sample size was calculated assuming an effect size of 0.6 and found to be at least a total sample size of 98 for each group with at least 49 cases with 90% power and 95% confidence level (G*Power version 3.1.9.4 computer software, Universität Düsseldorf, Germany). Results In the PMD group, 23 of 34 patients were male and 11 were female; in the I-KC group, 22 of 39 patients were male and 17 were female (p = 0.324). The mean age of the patients was 54.28 ± 7.34 in the PMD group and 50.89 ± 11.35 in the I-KC group, with no significant difference (p = 0.061). The demographic, topographic, and tomographic data of the patients are listed in Table 1. Table 1Comparison of demographic characteristics and corneal tomographic featuresPMD groupI-KC groupP valueAge (year)54.28 ± 7.3450.89 ± 11.350.061GenderMale23 (67.6%)22 (56.9%)0.324&(x2= 0.97)Female11 (32.3%)17 (43.5%)BCVA(Logmar)0.50 ± 0.340.40 ± 0.210.058K1 (D)41.98 ± 3.3244.51 ± 2.47 < 0.001 θ K2 (D)49.71 ± 4.0248.34 ± 3.510.056Kavg. (D)45.33 ± 2.8546.30 ± 2.960.117Kmax. (D)64.70 ± 12.6454.86 ± 6.13 < 0.001 Thinnest corneal pachymetry (µm)463.73 ± 68.65453.08 ± 40.840.315Anterior elevation98.48 ± 54.3050.15 ± 22.39 < 0.001 θ Posterior elevation140.41 ± 80.3973.86 ± 31.17 < 0.001 Astigmatism (3 mm) (D)7.96 ± 4.294.67 ± 2.33 < 0.001 Astigmatism (6 mm) (D)7.74 ± 4.193.23 ± 1.78 < 0.001 SIf2.73 ± 3.054.72 ± 2.77 < 0.001 SIb0.48 ± 0.821.41 ± 0.81 < 0.001 CSIf0.02 ± 1.502.18 ± 1.94 < 0.001 CSIb0.58 ± 0.750.88 ± 0.74 0.014 θ EIf28.20 ± 21.9134.09 ± 18.010.119EIb41.38 ± 31.4673.04 ± 38.56 < 0.001 *: Mann-Whitney U-test was used for non-parametric dataPMD: pellucid marginal degeneration, I-KC: inferior keratoconus, BCVA: best corrected visual acuity, K1: flat keratometry, K2: steep keratometry, Kavg.: average keratometry, Kmax.: maximum keratometry, SIf: symemetry index front, SIb: symemetry index back, CSIf: central-surrounding index front, CSIb: central-surrounding index back, EIf: ectasia index front, EIb: ectasia index back Table 1Comparison of demographic characteristics and corneal tomographic featuresPMD groupI-KC groupP valueAge (year)54.28 ± 7.3450.89 ± 11.350.061GenderMale23 (67.6%)22 (56.9%)0.324&(x2= 0.97)Female11 (32.3%)17 (43.5%)BCVA(Logmar)0.50 ± 0.340.40 ± 0.210.058K1 (D)41.98 ± 3.3244.51 ± 2.47 < 0.001 θ K2 (D)49.71 ± 4.0248.34 ± 3.510.056Kavg. (D)45.33 ± 2.8546.30 ± 2.960.117Kmax. (D)64.70 ± 12.6454.86 ± 6.13 < 0.001 Thinnest corneal pachymetry (µm)463.73 ± 68.65453.08 ± 40.840.315Anterior elevation98.48 ± 54.3050.15 ± 22.39 < 0.001 θ Posterior elevation140.41 ± 80.3973.86 ± 31.17 < 0.001 Astigmatism (3 mm) (D)7.96 ± 4.294.67 ± 2.33 < 0.001 Astigmatism (6 mm) (D)7.74 ± 4.193.23 ± 1.78 < 0.001 SIf2.73 ± 3.054.72 ± 2.77 < 0.001 SIb0.48 ± 0.821.41 ± 0.81 < 0.001 CSIf0.02 ± 1.502.18 ± 1.94 < 0.001 CSIb0.58 ± 0.750.88 ± 0.74 0.014 θ EIf28.20 ± 21.9134.09 ± 18.010.119EIb41.38 ± 31.4673.04 ± 38.56 < 0.001 *: Mann-Whitney U-test was used for non-parametric dataPMD: pellucid marginal degeneration, I-KC: inferior keratoconus, BCVA: best corrected visual acuity, K1: flat keratometry, K2: steep keratometry, Kavg.: average keratometry, Kmax.: maximum keratometry, SIf: symemetry index front, SIb: symemetry index back, CSIf: central-surrounding index front, CSIb: central-surrounding index back, EIf: ectasia index front, EIb: ectasia index back Comparison of demographic characteristics and corneal tomographic features 0.324& (x2= 0.97) *: Mann-Whitney U-test was used for non-parametric data PMD: pellucid marginal degeneration, I-KC: inferior keratoconus, BCVA: best corrected visual acuity, K1: flat keratometry, K2: steep keratometry, Kavg.: average keratometry, Kmax.: maximum keratometry, SIf: symemetry index front, SIb: symemetry index back, CSIf: central-surrounding index front, CSIb: central-surrounding index back, EIf: ectasia index front, EIb: ectasia index back When the epithelial thickness maps of the patients were evaluated, the central 2 mm min. and avg. values in the I-KC group; T min. and avg., IT min. and avg., I min. and avg. values in the 2–4 mm area; IT min. and avg., I min. values in the 4–6 mm area were significantly thinner than in the PMD group (all p < 0.05). In the PMD group, IN min. and avg., N min. and avg. in the 4–6 mm area; T min. and avg., IT min., I min., IN min. and avg., N min. and avg. in the 6–8 mm area, IT min. and avg., I min. and avg., IN min. values in the 8–10 mm area were significantly thinner compared to the I-KC group (all p < 0.05). No significant difference was observed in other areas (p > 0.05) (Table 2). Table 2Comparison of the thinnest and average epithelial thickness in the central and lower halves of the corneaPMD groupI-KC groupP valueCentral 2 mm circular area Min. (µm) 50.14 ± 4.5645.86 ± 6.85 < 0.001 Avg. (µm) 54.81 ± 4.8151.25 ± 6.68 0.002 Max. (µm) 59.30 ± 5.2357.31 ± 6.570.0772–4 mm annular area T min. (µm) 48.71 ± 5.0045.86 ± 7.18 0.016 T avg. (µm) 51.55 ± 4.9349.19 ± 6.96 0.039 IT min. (µm) 50.87 ± 4.4845.25 ± 7.31 < 0.001 IT avg. (µm) 53.73 ± 4.7947.79 ± 7.19 < 0.001 I min. (µm) 52.73 ± 4.5147.17 ± 6.41 < 0.001 I avg. (µm) 55.53 ± 4.4550.25 ± 6.16 < 0.001 IN min. (µm) 51.67 ± 3.7750.10 ± 6.010.096 IN avg. (µm) 55.08 ± 4.3953.53 ± 6.120.123 N min. (µm) 50.25 ± 3.4752.43 ± 6.41 0.026 N avg. (µm) 53.32 ± 3.5755.63 ± 6.100.064*4–6 mm annular area T min. (µm) 48.16 ± 4.5347.51 ± 4.680.458 T avg. (µm) 51.16 ± 4.3052.46 ± 5.070.142 IT min. (µm) 50.14 ± 3.9746.65 ± 5.15 < 0.001 IT avg. (µm) 53.41 ± 3.8850.06 ± 5.31 < 0.001 I min. (µm) 51.23 ± 5.1848.89 ± 4.60 0.012 I avg. (µm) 54.42 ± 5.3253.12 ± 4.110.144 IN min. (µm) 48.89 ± 4.9851.81 ± 3.47 < 0.001 IN avg. (µm) 52.42 ± 4.0155.39 ± 4.36 < 0.001 N min. (µm) 48.87 ± 4.5352.91 ± 3.98 < 0.001 N avg. (µm) 52.25 ± 4.0455.98 ± 4.91 < 0.001 6–8 mmannular area T min. (µm) 47.25 ± 4.8250.94 ± 5.01 < 0.001 T avg. (µm) 51.05 ± 4.3454.74 ± 4.63 < 0.001 IT min. (µm) 46.69 ± 6.1250.27 ± 5.76 0.002 IT avg. (µm) 52.05 ± 4.7753.91 ± 5.380.054 I min. (µm) 47.57 ± 6.8250.05 ± 6.32 0.047 I avg. (µm) 52.44 ± 5.8853.40 ± 5.580.377 IN min. (µm) 45.71 ± 7.0150.79 ± 5.76 < 0.001 IN avg. (µm) 51.64 ± 4.9554.27 ± 4.86 0.005 N min. (µm) 48.07 ± 7.2352.63 ± 4.69 < 0.001 N avg. (µm) 52.80 ± 4.7955.46 ± 4.42 0.003 8–10 mmannular area T min. (µm) 44.77 ± 6.3747.27 ± 7.020.055 T avg. (µm) 51.57 ± 5.1753.81 ± 5.460.075* IT min. (µm) 43.11 ± 8.6647.78 ± 6.59 0.003 IT avg. (µm) 51.88 ± 5.3554.39 ± 6.20 0.030 I min. (µm) 40.61 ± 8.6444.98 ± 7.45 0.007 I avg. (µm) 49.00 ± 7.4051.88 ± 6.73 0.043 IN min. (µm) 43.00 ± 9.3846.97 ± 6.86 0.029 IN avg. (µm) 54.22 ± 8.5255.80 ± 5.820.292 N min. (µm) 48.95 ± 9.7950.47 ± 8.220.977* N avg. (µm) 57.69 ± 7.7456.90 ± 6.610.582*: Mann-Whitney U-test was used for non-parametric dataPMD: pellucid marginal degeneration, I-KC: inferior keratoconus, T: temporal, IT: inferotemporal, I: inferior, IN: inferonasal, N: nasal, min.: minimum, avg.: average Table 2Comparison of the thinnest and average epithelial thickness in the central and lower halves of the corneaPMD groupI-KC groupP valueCentral 2 mm circular area Min. (µm) 50.14 ± 4.5645.86 ± 6.85 < 0.001 Avg. (µm) 54.81 ± 4.8151.25 ± 6.68 0.002 Max. (µm) 59.30 ± 5.2357.31 ± 6.570.0772–4 mm annular area T min. (µm) 48.71 ± 5.0045.86 ± 7.18 0.016 T avg. (µm) 51.55 ± 4.9349.19 ± 6.96 0.039 IT min. (µm) 50.87 ± 4.4845.25 ± 7.31 < 0.001 IT avg. (µm) 53.73 ± 4.7947.79 ± 7.19 < 0.001 I min. (µm) 52.73 ± 4.5147.17 ± 6.41 < 0.001 I avg. (µm) 55.53 ± 4.4550.25 ± 6.16 < 0.001 IN min. (µm) 51.67 ± 3.7750.10 ± 6.010.096 IN avg. (µm) 55.08 ± 4.3953.53 ± 6.120.123 N min. (µm) 50.25 ± 3.4752.43 ± 6.41 0.026 N avg. (µm) 53.32 ± 3.5755.63 ± 6.100.064*4–6 mm annular area T min. (µm) 48.16 ± 4.5347.51 ± 4.680.458 T avg. (µm) 51.16 ± 4.3052.46 ± 5.070.142 IT min. (µm) 50.14 ± 3.9746.65 ± 5.15 < 0.001 IT avg. (µm) 53.41 ± 3.8850.06 ± 5.31 < 0.001 I min. (µm) 51.23 ± 5.1848.89 ± 4.60 0.012 I avg. (µm) 54.42 ± 5.3253.12 ± 4.110.144 IN min. (µm) 48.89 ± 4.9851.81 ± 3.47 < 0.001 IN avg. (µm) 52.42 ± 4.0155.39 ± 4.36 < 0.001 N min. (µm) 48.87 ± 4.5352.91 ± 3.98 < 0.001 N avg. (µm) 52.25 ± 4.0455.98 ± 4.91 < 0.001 6–8 mmannular area T min. (µm) 47.25 ± 4.8250.94 ± 5.01 < 0.001 T avg. (µm) 51.05 ± 4.3454.74 ± 4.63 < 0.001 IT min. (µm) 46.69 ± 6.1250.27 ± 5.76 0.002 IT avg. (µm) 52.05 ± 4.7753.91 ± 5.380.054 I min. (µm) 47.57 ± 6.8250.05 ± 6.32 0.047 I avg. (µm) 52.44 ± 5.8853.40 ± 5.580.377 IN min. (µm) 45.71 ± 7.0150.79 ± 5.76 < 0.001 IN avg. (µm) 51.64 ± 4.9554.27 ± 4.86 0.005 N min. (µm) 48.07 ± 7.2352.63 ± 4.69 < 0.001 N avg. (µm) 52.80 ± 4.7955.46 ± 4.42 0.003 8–10 mmannular area T min. (µm) 44.77 ± 6.3747.27 ± 7.020.055 T avg. (µm) 51.57 ± 5.1753.81 ± 5.460.075* IT min. (µm) 43.11 ± 8.6647.78 ± 6.59 0.003 IT avg. (µm) 51.88 ± 5.3554.39 ± 6.20 0.030 I min. (µm) 40.61 ± 8.6444.98 ± 7.45 0.007 I avg. (µm) 49.00 ± 7.4051.88 ± 6.73 0.043 IN min. (µm) 43.00 ± 9.3846.97 ± 6.86 0.029 IN avg. (µm) 54.22 ± 8.5255.80 ± 5.820.292 N min. (µm) 48.95 ± 9.7950.47 ± 8.220.977* N avg. (µm) 57.69 ± 7.7456.90 ± 6.610.582*: Mann-Whitney U-test was used for non-parametric dataPMD: pellucid marginal degeneration, I-KC: inferior keratoconus, T: temporal, IT: inferotemporal, I: inferior, IN: inferonasal, N: nasal, min.: minimum, avg.: average Comparison of the thinnest and average epithelial thickness in the central and lower halves of the cornea 6–8 mm annular area 8–10 mm annular area *: Mann-Whitney U-test was used for non-parametric data PMD: pellucid marginal degeneration, I-KC: inferior keratoconus, T: temporal, IT: inferotemporal, I: inferior, IN: inferonasal, N: nasal, min.: minimum, avg.: average When the stromal and total corneal thickness maps were evaluated for minimum thickness, no significant differences were found in the central 2 mm area (stromal min. thickness p = 0.068 and corneal min. thickness p = 0.148). In the I-KC group, IT min. and I min. in the 2–4 mm area were significantly thinner than in the PMD group (all p < 0.05). In the PMD group, IT min., I min., IN min., and N min. values were significantly thinner in both the 6–8 mm and 8–10 mm areas (all p < 0.05). No significant differences were observed in other areas (p > 0.05) (Table 3). Table 3Comparison of the thinnest stromal thickness in the central and lower corneal halvesStromal thicknessTotal corneal thicknessPMDI-KCP valuePMDI-KCP valueCentral 2 mm circular area Min. (µm) 423.30 ± 41.67409.96 ± 35.600.068471.57± 49.46459.53 ± 38.260.1482–4 mm annular area T min. (µm) 431.64 ± 41.86419.50 ± 35.740.098480.23± 43.91467.27 ± 38.550.97 IT min. (µm) 431.05 ± 40.49409.27 ± 35.49 0.003 482.02 ± 40.17456.86 ± 38.58 0.001 I min. (µm) 428.62 ± 46.62412.87 ± 35.30 0.045 481.14± 47.27461.39 ± 38.26 0.016 IN min. (µm) 438.78 ± 50.74431.81 ± 32.950.388492.38± 49.80482.15 ± 36.610.216 N min. (µm) 452.85 ± 43.82450.55 ± 32.740.752506.57± 41.87504.39 ± 35.630.7664–6 mm annular area T min. (µm) 450.08 ± 40.23440.84 ± 33.170.183498.38± 38.84491.05 ± 34.770.291 IT min. (µm) 444.62 ± 44.77431.77 ± 31.28 0.044 496.61± 43.77481.01 ± 33.19 0.035 I min. (µm) 437.33 ± 62.40439.67 ± 31.080.849*488.39± 66.93489.63 ± 33.770.901 IN min. (µm) 452.69 ± 69.47467.15 ± 29.730.325*504.21± 72.51520.34 ± 33.310.133 N min. (µm) 483.92 ± 43.80484.71 ± 31.380.914534.79± 44.09540.98 ± 32.750.3956–8 mm annular area T min. (µm) 480.82 ± 42.62481.60 ± 32.880.913529.00± 40.82536.03± 33.910.318 IT min. (µm) 450.35 ± 67.78475.10 ± 30.77 0.045* 502.89± 66.89528.72± 31.91 0.011 I min. (µm) 428.92 ± 63.71483.28 ± 31.92 0.002* 481.20± 82.77536.42 ± 34.42 < 0.001 IN min. (µm) 454.66 ± 63.94509.32 ± 32.11 0.001* 505.04± 80.98562.39 ± 35.12 0.001* N min. (µm) 500.78 ± 66.06525.43 ± 28.89 0.041* 554.86± 68.42577.98 ± 34.60 0.026 8–10 mm annular area T min. (µm) 513.44 ± 56.39520.57 ± 35.340.742*562.54± 50.81574.76± 36.980.144 IT min. (µm) 467.44 ± 58.98519.03 ± 39.49 < 0.001 530.94± 58.47575.30± 36.95 < 0.001 I min. (µm) 456.26 ± 68.87537.78 ± 44.44 < 0.001 514.76± 66.15592.45± 44.05 < 0.001 IN min. (µm) 495.86 ± 66.55568.04 ± 41.45 < 0.001 553.69± 65.17622.27± 43.70 < 0.001 N min. (µm) 539.52 ± 65.29566.86 ± 35.87 0.011 594.02± 63.46624.22± 37.42 0.003 *: Mann-Whitney U-test was used for non-parametric dataPMD: pellucid marginal degeneration, I-KC: inferior keratoconus, T: temporal, IT: inferotemporal, I: inferior, IN: inferonasal, N: nasal, min.: minimum Table 3Comparison of the thinnest stromal thickness in the central and lower corneal halvesStromal thicknessTotal corneal thicknessPMDI-KCP valuePMDI-KCP valueCentral 2 mm circular area Min. (µm) 423.30 ± 41.67409.96 ± 35.600.068471.57± 49.46459.53 ± 38.260.1482–4 mm annular area T min. (µm) 431.64 ± 41.86419.50 ± 35.740.098480.23± 43.91467.27 ± 38.550.97 IT min. (µm) 431.05 ± 40.49409.27 ± 35.49 0.003 482.02 ± 40.17456.86 ± 38.58 0.001 I min. (µm) 428.62 ± 46.62412.87 ± 35.30 0.045 481.14± 47.27461.39 ± 38.26 0.016 IN min. (µm) 438.78 ± 50.74431.81 ± 32.950.388492.38± 49.80482.15 ± 36.610.216 N min. (µm) 452.85 ± 43.82450.55 ± 32.740.752506.57± 41.87504.39 ± 35.630.7664–6 mm annular area T min. (µm) 450.08 ± 40.23440.84 ± 33.170.183498.38± 38.84491.05 ± 34.770.291 IT min. (µm) 444.62 ± 44.77431.77 ± 31.28 0.044 496.61± 43.77481.01 ± 33.19 0.035 I min. (µm) 437.33 ± 62.40439.67 ± 31.080.849*488.39± 66.93489.63 ± 33.770.901 IN min. (µm) 452.69 ± 69.47467.15 ± 29.730.325*504.21± 72.51520.34 ± 33.310.133 N min. (µm) 483.92 ± 43.80484.71 ± 31.380.914534.79± 44.09540.98 ± 32.750.3956–8 mm annular area T min. (µm) 480.82 ± 42.62481.60 ± 32.880.913529.00± 40.82536.03± 33.910.318 IT min. (µm) 450.35 ± 67.78475.10 ± 30.77 0.045* 502.89± 66.89528.72± 31.91 0.011 I min. (µm) 428.92 ± 63.71483.28 ± 31.92 0.002* 481.20± 82.77536.42 ± 34.42 < 0.001 IN min. (µm) 454.66 ± 63.94509.32 ± 32.11 0.001* 505.04± 80.98562.39 ± 35.12 0.001* N min. (µm) 500.78 ± 66.06525.43 ± 28.89 0.041* 554.86± 68.42577.98 ± 34.60 0.026 8–10 mm annular area T min. (µm) 513.44 ± 56.39520.57 ± 35.340.742*562.54± 50.81574.76± 36.980.144 IT min. (µm) 467.44 ± 58.98519.03 ± 39.49 < 0.001 530.94± 58.47575.30± 36.95 < 0.001 I min. (µm) 456.26 ± 68.87537.78 ± 44.44 < 0.001 514.76± 66.15592.45± 44.05 < 0.001 IN min. (µm) 495.86 ± 66.55568.04 ± 41.45 < 0.001 553.69± 65.17622.27± 43.70 < 0.001 N min. (µm) 539.52 ± 65.29566.86 ± 35.87 0.011 594.02± 63.46624.22± 37.42 0.003 *: Mann-Whitney U-test was used for non-parametric dataPMD: pellucid marginal degeneration, I-KC: inferior keratoconus, T: temporal, IT: inferotemporal, I: inferior, IN: inferonasal, N: nasal, min.: minimum Comparison of the thinnest stromal thickness in the central and lower corneal halves 471.57 ± 49.46 480.23 ± 43.91 481.14 ± 47.27 492.38 ± 49.80 506.57 ± 41.87 498.38 ± 38.84 496.61 ± 43.77 488.39 ± 66.93 504.21 ± 72.51 534.79 ± 44.09 529.00 ± 40.82 536.03 ± 33.91 502.89 ± 66.89 528.72 ± 31.91 481.20 ± 82.77 505.04 ± 80.98 554.86 ± 68.42 562.54 ± 50.81 574.76 ± 36.98 530.94 ± 58.47 575.30 ± 36.95 514.76 ± 66.15 592.45 ± 44.05 553.69 ± 65.17 622.27 ± 43.70 594.02 ± 63.46 624.22 ± 37.42 *: Mann-Whitney U-test was used for non-parametric data PMD: pellucid marginal degeneration, I-KC: inferior keratoconus, T: temporal, IT: inferotemporal, I: inferior, IN: inferonasal, N: nasal, min.: minimum All parameters that showed significant differences were examined using ROC curve analysis. The highest AUC values were observed for epithelial thickness parameters IT 2–4 min., IN 6–8 min., and IN 4–6 min. (0.797, 0.707, and 0.727, respectively), for stromal thickness parameters IT 2–4 min., IN 8–10 min., and I 8–10 min. (0.644, 0.811, and 0.804, respectively), and for corneal thickness parameters IT 2–4 min., IN 8–10 min., and I 8–10 min. (0.668, 0.823, and 0.807, respectively) (all p < 0.05) (Table 4). Table 4ROC curve analysis in the distinction between PMD and I-KCRisk factorAUC (95% CI)Cut-off valueP valueSensitivity %Specifity %Epithelial thickness IT 2–4 min. (µm) 0.797 (0.713–0.882)47.50 < 0.001 75.0%69.5% IN 6–8 min. (µm) 0.707 (0.611–0.803)48.50 < 0.001 74.1%66.7% IN 4–6 min. (µm) 0.706 (0.608–0.804)50.50 < 0.001 72.4%67.9% IT 2–4 min./ IN 6–8 min. ratio 0.855 (0.785–0.925)0.970 < 0.001 78.6%79.7% IT 2–4 min.-IN 6–8 min. differance 0.811 (0.729–0.888)-0.500 < 0.001 75.0%74.1%Stromal thickness IT 2–4 min. (µm) 0.644 (0.542–0.746)418.50 0.008 57.1%56.9% IN 8–10 min. (µm) 0.811 (0.721–0.902)543.50 < 0.001 74.0%75.0% I 8–10 min. (µm) 0.804 (0.712–0.896)511. < 0.001 72.3%73.9% IT 2–4 min./ IN 8–10 min. ratio 0.869 (0.799–0.939)0.773 < 0.001 77.3%78.0% IT 2–4 min.-IN 8–10 min. difference 0.878 (0.811–0.944)-126 < 0.001 79.5%76.0%Total corneal thickness IT 2–4 min. (µm) 0.668 (0.569–0.767)469.50 0.002 58.9%60.3% IN 8–10 min. (µm) 0.823 (0.741–0.904)594.00 < 0.001 76.4%75.0% I 8–10 min. (µm) 0.807 (0.716–0.898)566 < 0.001 75.5%76.1% IT 2–4 min/ IN 8–10 min. ratio 0.864 (0.798–0.930)0.792 < 0.001 72.9%72.7% IT 2–4 min.-IN 8–10 min. difference 0.874 (0.810–0.938)-122.50 < 0.001 75.0%72.7%PMD: pellucid marginal degeneration, I-KC: inferior keratoconus, AUC: area under curve, 95% CI: 95% confidence interval, IT: inferotemporal, I: inferior, IN: inferonasal, min.: minimum Table 4ROC curve analysis in the distinction between PMD and I-KCRisk factorAUC (95% CI)Cut-off valueP valueSensitivity %Specifity %Epithelial thickness IT 2–4 min. (µm) 0.797 (0.713–0.882)47.50 < 0.001 75.0%69.5% IN 6–8 min. (µm) 0.707 (0.611–0.803)48.50 < 0.001 74.1%66.7% IN 4–6 min. (µm) 0.706 (0.608–0.804)50.50 < 0.001 72.4%67.9% IT 2–4 min./ IN 6–8 min. ratio 0.855 (0.785–0.925)0.970 < 0.001 78.6%79.7% IT 2–4 min.-IN 6–8 min. differance 0.811 (0.729–0.888)-0.500 < 0.001 75.0%74.1%Stromal thickness IT 2–4 min. (µm) 0.644 (0.542–0.746)418.50 0.008 57.1%56.9% IN 8–10 min. (µm) 0.811 (0.721–0.902)543.50 < 0.001 74.0%75.0% I 8–10 min. (µm) 0.804 (0.712–0.896)511. < 0.001 72.3%73.9% IT 2–4 min./ IN 8–10 min. ratio 0.869 (0.799–0.939)0.773 < 0.001 77.3%78.0% IT 2–4 min.-IN 8–10 min. difference 0.878 (0.811–0.944)-126 < 0.001 79.5%76.0%Total corneal thickness IT 2–4 min. (µm) 0.668 (0.569–0.767)469.50 0.002 58.9%60.3% IN 8–10 min. (µm) 0.823 (0.741–0.904)594.00 < 0.001 76.4%75.0% I 8–10 min. (µm) 0.807 (0.716–0.898)566 < 0.001 75.5%76.1% IT 2–4 min/ IN 8–10 min. ratio 0.864 (0.798–0.930)0.792 < 0.001 72.9%72.7% IT 2–4 min.-IN 8–10 min. difference 0.874 (0.810–0.938)-122.50 < 0.001 75.0%72.7%PMD: pellucid marginal degeneration, I-KC: inferior keratoconus, AUC: area under curve, 95% CI: 95% confidence interval, IT: inferotemporal, I: inferior, IN: inferonasal, min.: minimum ROC curve analysis in the distinction between PMD and I-KC PMD: pellucid marginal degeneration, I-KC: inferior keratoconus, AUC: area under curve, 95% CI: 95% confidence interval, IT: inferotemporal, I: inferior, IN: inferonasal, min.: minimum In terms of epithelial thickness, the cut-off value of IT 2–4 min./IN 6–8 min. ratio in distinguishing PMD from I-KC cut-off was ≥ 0.97, sensitivity was 78.6%, specificity was 79.7%, and AUC (95%CI) value was 0.855 (0.785–0.925) (p < 0.001). Similarly, IT 2–4 min.-IN 6–8 min. difference cut-off value was ≥-0.50 μm, sensitivity was 75%, specificity was 72.9%, and AUC (95%CI) value was 0.809 (0.729–0.888) (p < 0.001) (Table 4). For stromal thickness, the IT 2–4 min./IN 8–10 min. ratio had a PMD and I-KC cut-off value of ≥ 0.773, with sensitivity of 77.3%, specificity of 78.0%, and AUC (95% CI) of 0.869 (0.799–0.939) (p < 0.001). Similarly, the IT 2–4 min.-IN 8–10 min. difference had a cut-off value of ≥-126 μm, with sensitivity of 79.5%, specificity of 76%, and AUC (95% CI) of 0.855 (0.785–0.925) (p < 0.001) (Table 4) (Fig. 2). Fig. 2ROC curve analysis for differentiating PMD from I-KC using stromal and total corneal thickness; s_IT 2–4 min./IN 8–10 min. ratio (stromal inferotemporal 2–4 mm zone minimum thickness /inferonasal 8–10 mm zone minimum thickness ratio), and s_IT 2–4 min.-IN 8–10 min. difference (stromal inferotemporal 2–4 mm zone minimum thickness - inferonasal 8–10 mm zone minimum thickness difference), c_ IT 2–4 min./IN 8–10 min. ratio (corneal inferotemporal 2–4 mm zone minimum thickness /inferonasal 8–10 mm zone minimum thickness ratio) and, c_IT 2–4 min.-IN 8–10 min. difference (corneal inferotemporal 2–4 mm zone minimum thickness - inferonasal 8–10 mm zone minimum thickness difference) indices ROC curve analysis for differentiating PMD from I-KC using stromal and total corneal thickness; s_IT 2–4 min./IN 8–10 min. ratio (stromal inferotemporal 2–4 mm zone minimum thickness /inferonasal 8–10 mm zone minimum thickness ratio), and s_IT 2–4 min.-IN 8–10 min. difference (stromal inferotemporal 2–4 mm zone minimum thickness - inferonasal 8–10 mm zone minimum thickness difference), c_ IT 2–4 min./IN 8–10 min. ratio (corneal inferotemporal 2–4 mm zone minimum thickness /inferonasal 8–10 mm zone minimum thickness ratio) and, c_IT 2–4 min.-IN 8–10 min. difference (corneal inferotemporal 2–4 mm zone minimum thickness - inferonasal 8–10 mm zone minimum thickness difference) indices For the IT 2–4 min./IN 8–10 min. ratio in terms of total corneal thickness, the PMD cut-off value was ≥ 0.792, with sensitivity 72.9%, specificity 72.7%, and AUC (95% CI) 0.864 (0.798–0.930) (p < 0.001). Similarly, for the IT 2–4 min. -IN min. difference, the cut-off value for an 8–10 min. difference was ≥-122.5 μm, with sensitivity 75%, specificity 72.7%, and AUC (95% CI) 0.874 (0.810–0.938) (p < 0.001) (Table 4) (Fig. 2). Discussion PMD and I-KC; the differentiation of progression rates, frequency, and types of complications that may develop, responses to keratoplasty or corneal ring treatment, and applications are very important from a clinical perspective [15]. While lamellar wedge resection is an important option for visual rehabilitation before keratoplasty in PMD, it is not a primary treatment option for I-KC, highlighting the importance of distinguishing these two ectasias [16]. Sheimpflug camera-based corneal tomography, frequently used in clinical practice, does not provide sufficient information to differentiate these two entities [13]. However, AS-OCT, which is increasingly used and important in clinical practice, can be a priority diagnostic method due to its detailed analysis of corneal layers. Our study used the MS- 39 Placido disk combined with AS- OCT to evaluate epithelial, stromal, and total corneal maps in eyes with PMD and I- KC. MS- 39 has been shown to have good reproducibility for epithelial thickness in both keratoconic and healthy eyes across the central and peripheral cornea in previous studies [12, 17]. We found that epithelial, stromal, and total thinnest thickness values were significantly reduced in the 6–8 mm and 8–10 mm areas, especially from the inferonasal to the inferotemporal zones, in PMD patients compared with I- KC patients. This thinning pattern is consistent with the band- shaped thinning beyond the 1–2 mm area from the inferior limbus, as described for PMD [2]. However, in areas where stromal and total corneal thickness is reduced, epithelial thinning may be a response to the existing ectatic condition, as epithelial thickness normally thickens inferiorly [18, 19]. In I- KC patients, the epithelial map was significantly thinner in the central 2 mm and 2–4 mm and in the inferior 4–6 mm, especially in the inferior and inferotemporal zones, and the stromal and total corneal thickness maps were thinner in the 2–4 mm and 4–6 mm areas, especially in the inferotemporal zone. This inferior thinning, especially in the inferotemporal zone within the 6- mm circular area, was consistent with epithelial thinning over the cone area reported in KC patients in the literature [9]. Similar to our study, Mohr et al. found significant epithelial, stromal, and corneal thickness thinning in the 7–9 mm area in the inferotemporal zone in PMD eyes and especially in the 2–5 mm area in the inferior and inferotemporal zone in the KC group in their AS- OCT (Avanti/RTVue) study comparing 10 PMD eyes with 49 KC eyes [13]. However, unlike us, they did not detect significant changes in the 5–7 mm area they defined as the middle zone, which may be due to their inability to distinguish KC patients, especially those with I-KC, whose sagittal curvature maps resembled PMD, as in our study. Another advantage of our study is that we included 56 PMD eyes and tried to neutralize age-related changes by selecting the I-KC group from the older population. Previous studies have shown that epithelial thickness decreases with age [20]. Many studies in the literature have examined the distinction of subclinical and forme fruste keratoconus from normal eyes, keratoconus from corneal warpage, and keratoconus from healthy eyes using epithelial and stromal thickness maps [11, 21, 22]. In these studies, very high AUC, specificity, and sensitivity values were obtained for descriptive performance using neural networks, decision trees, and logistic regression analyses of multiple parameters [11, 21, 22]. When we examined PMD and I-KC cases that could show the same patterns in the sagittal curvature map, which were very confusing, the best results were found with a single parameter for epithelial thickness; IT 2–4 min. and IN 6–8 min. (AUC values 0.797 and 0.707, respectively). We obtained better AUC values in combinations of these parameters, specifically the IT 2–4 min.-IN 6–8 min. difference and the IT 2–4 min./IN 6–8 min. ratio (AUC values 0.855 and 0.809, respectively). For single parameters, we obtained the best results in stromal and total corneal thicknesses in the IN 8–10 min. area (AUC values 0.811 and 0.823, respectively). The selection of the IT 2–4 mm and IN 8–10 mm zones for parameter comparison from stromal and corneal thickness maps is based on both the thinning patterns observed in our dataset and the known pathophysiological features of PMD and I-KC. In PMD, corneal thinning is not limited to the directly inferior region but tends to extend toward the inferonasal and inferotemporal periphery, with the most pronounced thinning exhibiting a band-shaped peripheral thinning pattern 1–2 mm from the limbus, characteristic of this ectasia [2, 3]. In contrast, in I-KC, the cone apex is typically located in the inferotemporal paracentral region [4, 7]. Therefore, these two regions may represent the zones with the highest diagnostic value in distinguishing PMD from I-KC, as they reflect different thinning patterns specific to each disease. In stromal thickness, the AUC values of the combination of IT 2–4 min.-IN 8–10 min. difference and IT 2–4 min./IN 8–10 min. ratio were 0.869 and 0.878, whereas in total corneal thickness they were 0.864 and 0.874. Mohr et al. obtained higher AUC values (0.910–0.977) in the differentiation of PMD and KC using combination parameters in the form of the difference and ratio of the inferotemporal 2–5 mm zone with the inferior 7–9 mm zone (IT 2–5/I7-9 difference and IT 2–5/I7-9 ratio) [13]. Our study obtained lower AUC and sensitivity values because the cone position of eyes with I-KC in the crab-claw pattern is closer to the areas where PMD thins than central or more central KCs. Another reason is that we included more eyes with PMD in the study than Mohr et al., thus allowing us to evaluate PMDs at different stages. In distinguishing I-KC patients with crab-claw and kissing doves patterns on the sagittal curvature map from real PMD patients, Koç et al. found that corneal densitometry values in the 6–10 mm and 10–12 mm zones were significantly higher in PMD patients than in I-KC patients [6]. This increase in densitometry was attributed to increased backscatter of light due to disruptions in the epithelial and stromal layers [6, 23]. In our study, we showed that band-shaped deterioration in the peripheral and inferior zones affected both the epithelium and the stroma. In addition, the same study found that ectasia indices were significantly higher in the I-KC group than in PMD. Although we used a different device in our study, we found that the symmetry and ectasia indices (SIf, SIb, CSIf, CSIb, and EIb) were significantly higher in the I-KC group than in the PMD group. This is because the ectasia and symmetry indices were evaluated in the central 3 mm and 6 mm circular areas [24]. While the strength of our study is that it is the largest study using corneal sublayer analysis to differentiate patients with PMD and I-KC who show similar patterns on the sagittal curvature map, it has some limitations. First, we could only evaluate a 10-mm area with MS-39 AS-OCT, which meant we could not assess some peripheral parts of the cornea, which has an average diameter of 11–12 mm. Second, our sample size was limited despite PMD being a rare form of corneal ectasia. As a result, this study enables a high-quality distinction by using epithelial, stromal, and total corneal thickness maps in PMD and I-KC patients who display similar patterns in the sagittal curvature map. Particularly in stromal and total corneal thickness maps, combinations of differences and ratios between IT 2–4 mm and IN 8–10 mm zones show good accuracy and can assist clinicians in differentiating between PMD and I-KC, which show similar patterns in axial sagittal curvature maps. Further research with larger sample sizes, devices capable of better analyzing corneal layers, and parameters that can improve discrimination is needed.