Preoperative Imaging Model Can Accurately Predict Edema Course After Keratoplasty Procedure

From left, Kay Coulson, Dr. Jim Montgomery and Marlene Beckett perform a CK (conductive keratoplasty) procedure on a patient at Insight Lasik in Lafayette, Colo. Montgomery is the only eye surgeon in Colorado currently certified to perform CK. (The Denver Post / Jerry Cleveland) (Photo By Jerry Cleveland/The Denver Post via Getty Images)
Scheimpflug imaging helps illuminate corneal edema resolution after Descemet membrane endothelial keratoplasty.

Researchers have developed and externally validated a model to accurately predict edema resolution following Descemet membrane endothelial keratoplasty (DMEK) surgery in patients with advanced Fuchs’ dystrophy, according to research published in JAMA Ophthalmology.

Using participants from 2 separate prospective studies, researchers developed the model using Scheimpflug tomographic imaging. 

Two cohorts of patients with advanced Fuchs’ dystrophy — a deviation cohort (n=100) and a validation cohort (n=32) — were created. The deviation cohort was used to develop a predictive model utilizing linear least absolute shrinkage; selection operator regression was also conducted in this group. In the validation cohort, investigators tested the overall performance, discrimination, and calibration. 

Up to 3 attempts were made to obtain high-quality Scheimpflug images in participants. Three independent ophthalmologists specializing in the cornea evaluated the tomographic features, including irregularity in the lines of equal corneal thickness, displacement of the thinnest point of corneal thickness, and the absolute amount of focal posterior corneal depression. 

A predictive regression model was developed for corneal edema resolution. Investigators considered the following Scheimpflug parameters as potential predictors of edema resolution: tomographic features of nonparallel isopachs, displacement of the thinnest point, focal posterior corneal depression, preoperative central corneal thickness, and standardized anterior and posterior corneal backscatter. 

The model was then externally validated in the validation cohort to examine the overall performance, discrimination, and calibration. 

A total of 88 patients (61% women; median age, 68 years) were included in the analysis. Prior to DMEK, median corneal thickness was 586 µm (interquartile range [IQR], 555 µm to 611 µm) and 603 µm (IQR, 579 µm to 625 µm) in the derivation and validation cohorts, respectively. Seventy-seven and 19 eyes in each cohort had nonparallel isopachs, and 73 and 21 eyes in each cohort had focal posterior depression (median absolute depression, 18 µm and 13 µm, respectively). 

Median time between pre- and postoperative Scheimpflug image was 12 months and 14 months in the derivation and validation cohorts, respectively (IQR, 9 months to 16 months and 10 months to 18 months). Following DMEK, median corneal thickness was 77 µm thinner in the derivation cohort and 75 µm thinner in the validation cohort. Eyes in each cohort that had no tomographic features prior to DMEK had a median edema resolution of 42 µm and 59 µm, resulting in a mean difference of 29 µm (95% CI, 17 µm to 41 µm). Eyes with more edema resolution experienced a higher gain in best-corrected visual acuity following DMEK. 

Goodness-of-fit statistics were calculated in the derivation cohort to internally validate the model. Overall model performance was high and area under the curve was 0.84. Overall model performance was also high in the validation cohort with a mean difference between predicted and observed edema of 3.3 µm, indicative of good calibration without over- or underestimation. Area under the curve was 0.97 in distinguishing between patients with more than 50 µm of edema and those with less edema resolution. 

Study limitations include the use of data from a single academic center and the use of an intermediate endpoint. 

Future studies, the researchers wrote, must demonstrate whether the implementation of this model for clinical use results in improved decision-making and better long-term visual outcomes. 

“The developed and validated model described…appears to predict edema resolution after DMEK in eyes with Fuchs’ dystrophy,” the research says. “Applying this model in clinical practice and in research settings in conjunction with subjective, morphologic, and optical parameters of disease severity may allow for more precise and personalized counseling on outcomes.” 

Disclosure: One study author declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.


Zander D, Grewing V, Glatz A, et al. Predicting edema resolution after Descemet membrane endothelial keratoplasty for Fuchs dystrophy using Scheimpflug technology. JAMA Ophthalmol. Published online February 18, 2021. doi:10.1001/jamaophthalmol.2020.6994