Working out the mystery of ectasia risk with artificial intelligence

This article was reviewed by Renato Ambrósio, Jr, MD, PhD

Ectasia is an intriguing and mysterious complication of laser-vision-correction (LVC) procedures.

The potentially devastating problem underscores the importance of determining the susceptibility of the cornea for developing progressive ectasia, and of going beyond detecting just mild or subclinical keratoconus.

The corneal structure as well as the potential impact of LVC should be considered to predict ectasia risk in every patient.

Related: Teaching AI algorithms to identify corneal pathology: The future is now

“The LVC procedure and eye rubbing are the primary environmental culprits in the development of ectasia in any cornea,” said Renato Ambrósio, Jr, MD, PhD. “So, a basic factor for avoiding ectasia is educating the patient not to rub the eye.”

Ambrósio is an adjunct professor of ophthalmology at the Federal University of the State of Rio de Janeiro, and director of refractive surgery, VisareRIO Refracta Personal Laser, Rio de Janeiro, Brazil.

Ambrósio cited the Ectasia Risk Score System study by Randleman and colleagues that improved the identification of the risk for ectasia by combining topographic pattern, residual stromal bed, age, and preoperative pachymetry. He noted it also manifests refraction.¹

The risk stratification factor scale had a specificity of 91% and a sensitivity of 96%.

In addition, as early as 2001, it was recognized that more than half of the original corneal thickness should be preserved to ensure the corneal integrity; this was counter to the classic “250 μm rule” for the residual stromal bed.²

This concept was further expanded by Santhiago and colleagues with their study of percentage of tissue altered (PTA),³ which determined that more than 40% is a risk factor for ectasia in cases with normal corneal topography.

Related: CXL imperative in treatment of pediatric keratoconus

Going beyond
Ectasia progression occurs due to biomechanical failure that is related to the preoperative structure of the cornea, and due to the impact of the LVC procedure.

Considering ectasia has been reported even in low-risk patients with no signs of keratoconus, better presurgical assessment is needed.

“Identifying mild or subclinical keratoconus is essential but not enough. The quest is to characterize the susceptibility forbiomechanical decompensation or ectasia progression,” Ambrósio said.

The deviation from normality in the Belin-Ambrosio Enhanced Ectasia Display (Oculus Pentacam; Wetzlar, Germany) was the first attempt to predict ecstasia using linear regression analysis. Furthermore, the Pentacam Random Forest Index was developed by Lopes and colleagues⁴ in a multicenter case-control study that applied more sophisticated machine learning tools to enhance tomographic characterization of ectasia susceptibility.

The integration of tomography data and corneal biomechanical parameters enabled further improvement of the accuracy for detecting mild (subclinical) ectasia as demonstrated in normal topography eyes from patients with very asymmetric ectasia.

Related: ASCRS 2020: Returning to baseline post-corneal collagen crosslinking for keratoconus

To understand the impact on the cornea from LASIK when assessing ectasia risk, Ambrósio and colleagues conducted a trial to develop an enhanced novel data-driven method for doing so: the relational thickness/tissue altered (RTA).

Reoperative data were available from 3278 patients with stable LASIK outcomes and from 105 patients who developed ectasia after LASIK; all were followed for more than 2 years.

The patients’ central corneal thickness (CCT) (apex) and at the thinnest point (TP),age at surgery, manifest refraction, maximal ablation depth, and flap thickness were included.

The residual stromal bed (RSB) and the PTA were calculated based on either CCT and TP.

The first results showed that TP is a better, more accurate parameter for calculating the RSB and the PTA and assessing ectasia risk than is CCT.

Investigators developed the RTA as a logistic regression formula that considers age, the relationship between the flap and the TP, the relationship between the ablation and the TP, the TP of the RSB, and the difference between the apex and the minimal thickness.

Related: Improving LASIK outcomes with biomechanical analysis

According to Ambrósio, the RTA was better (P <.001) than the minimal stromal bed and the PTA values.

“Progressive ectasia occurs because of corneal biomechanical decompensation,” Ambrósio concluded. “This is related to the preoperative innate ectasia susceptibility and to the impact of the LVC. We must extend the role of machine learning in ectasia. The RTA is a better metric to determine the impact of LASIK. Future studies involving finite element modeling are needed.”

Read more by Lynda Charters

---

Renato Ambrósio, Jr, MD, PhD
e:dr.renatoambrosio@gmail.com
Ambrósio is a consultant for Oculus and a founder of the Brazilian Study Group of Artificial Intelligence and Corneal Analysis.

References

1. Randleman JB, Woodward M, Lynn MJ, Stulting RD. Risk assessment for ectasia after corneal refractive surgery. Ophthalmology. 2008;115(1):37-50. doi:10.1016/j.ophtha.2007.03.073

2. Ambrósio R Jr, Wilson SE. Complications of laser in situ keratomileusis: etiology, prevention, and treatment. J Refract Surg. 2001;17(3):350-379.

3. Godfrey KJ, Korn BS, Kikkawa DO. Blepharitis following ocular surgery: identifying risk factors. Curr Opin Ophthalmol. 2016;27(1):311-315.

4. Lopes BT, Ramos IC, Salomão MQ, et al. Enhanced tomographic assessment to detect corneal ectasia based on artificial intelligence.Am J Ophthalmol 2018;195:223-232. doi: 10.1016/j.ajo.2018.08.005.