ABCD progression display for keratoconus progression: a sensitivity-specificity study.

OBJECTIVE: To evaluate the sensitivity and specificity of the ABCD progression display for keratoconus progression. METHODS: Data was collected from patients that underwent at least two Pentacam assessments 6 months apart. Sensitivity and specificity were calculated for the ABCD progression display. Progression was defined by criterion 1: change in two ABCD parameters above 80% confidence interval (CI) or criterion 2: change in one ABCD parameter above 95%CI. Receiver operating characteristic analysis compared the area under the curve (AUC) of all ABCD parameter combinations. RESULTS: Thirty eyes were evaluated over a median time of 10.3 months. Progression by criterion 1 resulted in a sensitivity of 61.9% and specificity of 88.9%. Progression by criterion 2 resulted in higher sensitivity (80.9%) and specificity (100%). Pairwise comparisons of the ROC curves show that the AUC achieved by criterion 2 was significantly higher than criterion 1 (0.905 vs. 0.754, p = 0.0332). Evaluation of all ABCD combinations with a significant change of 80% or 95% CI did not show superiority over criterion 1 or 2 regarding progression detection. The D parameter had a very low AUC (0.5-0.556). CONCLUSIONS: The ABCD progression display can assess keratoconus progression with high sensitivity and specificity, thus assisting the patients' decision-making process. The D parameter did not contribute to the sensitivity or specificity of this classification.

Validation of the multi-metric D-index change in the assessment of keratoconus progression.

PURPOSE: To validate the effectiveness of the multi-metric D-index by Pentacam in detecting keratoconus (KC) progression. MATERIALS AND METHODS: This was a retrospective study of KC patients at the Rabin Medical Center, Petah Tikva, Israel, during 2016-2018 with at least two corneal tomography examinations six months apart. Agreement between clinical diagnosis of progression (1.5D increase in mean keratometric value, 1D increase in Kmax, a 5% decrease in central corneal thickness (CCT), worsening of visual acuity by more than one line, or deterioration of manifest corneal astigmatism > 1.5D) and the D-index was evaluated. Receiver operating characteristic (ROC) analysis was used to find the D-index's optimal cutoff value to show progression. RESULTS: We included KC eyes in the stable group (N = 7) and the progression group (N = 54). Patient demographics and tomographic parameters at baseline were similar between the groups. The D-index change was significantly higher in the progression group than in the stable group (median + 1 and 0.0, respectively, p = 0.024). Based on the ROC analysis, the optimal D-index cutoff change within at least six months was 0.32 (59.3% sensitivity and 100% specificity (area under the curve [AUC] = 0.771, Youden = 0.592). Subjects with a D-index change above this value had a 21.1-fold increase in odds for corneal ectasia progression requiring CXL (OR: 21.1, 95%CI 1.17-398.8, p = 0.038). CONCLUSION: The multi-metric D-index can serve as a clinically feasible parameter to detect KC progression and guide patients' referral for further interventions.

Thickness measurements taken with the spectralis OCT increase with decreasing signal strength.

BACKGROUND: Optical coherence tomography (OCT) is used worldwide by clinicians to evaluate macular and retinal nerve fiber layer (RNFL) characteristics. It is frequently utilized to assess disease severity, progression and efficacy of treatment, and therefore must be reliable and reproducible. OBJECTIVE: To examine the influence of signal strength on macular thickness parameters, macular volume measurement and RNFL thickness measured by spectral-domain optical coherence tomography (SD-OCT). METHODS: Macular thickness parameters, macular volume measurement and RNFL thickness were measured by the Spectralis® OCT (Heidelberg Engineering, Heidelberg, Germany). In each eye, the focusing knob was adjusted to obtain 4 images with different signal strengths - Low (below 15), Moderate (15-20), Good (20-25) and Excellent (above 25). The relationship between signal strength and measured data was assessed using the mixed model procedure. RESULTS: A total of 71 eyes of 41 healthy subjects were included. Central macular thickness, macular volume and mean RNFL thickness increased with decreasing signal strength. Specifically, eyes with excellent signal strength showed significantly thinner central macular thickness (p = 0.023), macular volume (p = 0.047), and mean RNFL thickness (p = 0.0139). CONCLUSIONS: Higher signal strength is associated with lower macular thickness, macular volume and RNFL thickness measurements. The mean differences between excellent and low-quality measurements were small implicating that SD-OCT is a reliable imaging tool even at low quality scans. It is imperative that the physician compares the signal strength of all scans, as minute differences may alter results.

Optical Coherence Tomography Biomarkers Predicting Visual Acuity Change after Intravitreal Bevacizumab Injections for Macular Edema Secondary to Branch Retinal Vein Occlusion.

PURPOSE: This study aimed to identify baseline optical coherence tomography (OCT) factors in branch retinal vein occlusion (BRVO) that predict response to bevacizumab injections. METHODS: It is a retrospective case series of consecutive patients that underwent OCT at diagnosis, and the central macular thickness (CMT), extent of disorganized retinal inner layers, and outer retinal layers including external limiting membrane, ellipsoid zone (EZ), and cone outer segment tips (COST) were measured. Patients received 3 consecutive monthly injections of bevacizumab followed by pro re nata treatment. The main outcome measure was improvement in best-corrected visual acuity (BCVA) after 1 year. Results are expressed as mean ± SD, and p value <0.05 was considered statistically significant. RESULTS: Overall, 66 eyes of 66 patients, with an average age of 68.5 ± 11.4 years, were included. The mean BCVA improved significantly from 0.68 logMAR at baseline to 0.50 logMAR at 1 year (p < 0.001). Baseline logMAR BCVA (r = 0.41, p < 0.001) and CMT (r = 0.23, p = 0.04) were associated with improvement while EZ (r = -0.24, p = 0.05) and COST (r = -0.32, p = 0.01) disruption with deterioration in BCVA. CONCLUSION: In patients with naïve BRVO treated with bevacizumab, BCVA improvement at 1 year can be predicted from baseline BCVA, CMT, extent of COST disruption, and EZ disruption.

Predicting response of ischemic central retinal vein occlusion to bevacizumab injections: 1 year follow-up.

PURPOSE: To identify baseline optical coherence tomography (OCT) factors in ischemic central retinal vein occlusion (CRVO) that predict response to bevacizumab injections. METHODS: Patients underwent OCT at diagnosis and the central macular thickness, subretinal fluid width and height as well as the presence of intraretinal cysts were measured. The extent of disorganized retinal inner layers, outer plexiform layer (OPL), external limiting membrane, ellipsoid zone and cone outer segment tips (COST) was recorded. Patients received three consecutive monthly injections of bevacizumab followed by pro re nata treatment. RESULTS: Overall 32 eyes of 32 patients aging 69.4 ± 12.7 years were included. In univariate correlational analyses baseline OPL disruption (- 41, p = 0.02), EZ disruption (r = - 0.36, p = 0.05) and COST disruption (r = - 0.37, p = 0.04) correlated with deterioration in BCVA at 1 year. Partial least squares demonstrated that the factors most largely associated with deterioration in BCVA were COST disruption and OPL disruption followed by EZ disruption, while worst vision at baseline was associated with improvement in vision. CONCLUSION: In patients with ischemic CRVO treated with bevacizumab, BCVA improvement at 1 year can be partially predicted from baseline OCT measurements by the extent of COST, OPL and EZ disruption.