Official ESCRS | European Society of Cataract & Refractive Surgeons

 

Next-generation mapping of 3D epithelium thickness: a tool to assess wound healing after refractive surgery

Search Title by author or title

Session Details

Session Title: Advanced Tools for Outcome Assessment

Session Date/Time: Tuesday 17/09/2019 | 08:30-10:30

Paper Time: 09:32

Venue: Free Paper Forum: Podium 4

First Author: : G.Kundu INDIA

Co Author(s): :    R. Shetty   A. Roy   M. Francis                       

Abstract Details

Purpose:

To profile 3-D distribution of epithelium thickness before and after refractive surgery with Zernike polynomials.

Setting:

Narayana Nethralaya Superspeciality Eye Hospital, Bangalore,India

Methods:

The study included 50 SMILE (Small Incision Lenticule Extraction), 50 LASIK (Laser-Assisted In Situ Keratomileusis), 42 PRK (Photorefractive Keratectomy) and 40 Trans-PRK (Transepithelial PRK) eyes of 124 patients. Optical coherence tomography (OCT, RTVue, Optovue Inc) scan was performed before and after surgery. The 3-dimentional epithelial thickness map was generated by interpolating thickness between eight 2-D scans of 6mm diameter. Epithelial Zernike indices (EZI) were used to describe the asymmetric spatial distribution of epithelium thickness. Pre-minus-post (pre-to-post) EZI were included in decision tree based artificial intelligence (AI) model to obtain feature unique to postoperative wound healing.

Results:

Root mean square (RMS) 3rd and 4th order EZI showed significant increase in LASIK (-0.53±0.43 and -0.41±0.37 µm; p<0.0001) and SMILE (-0.25±0.38 and -0.35 ±0.48 µm; p<0.005), respectively.Further,pre-minus-post 3rd order EZI was greater in PRK (manual: 0.03±0.61 and Trans-PRK:-0.15±0.32 µm) compared to LASIK (P<0.05).RMS of coma was similar between PRK groups [manual: 0.04±0.29 and Trans-PRK:-0.14±0.37µm] and,also between SMILE (-0.43 ± 0.4 µm) and LASIK (-0.43 ± 0.36 µm) (P>0.05).These results suggest the presence of unique features between flap/cap(SMILE,LASIK) vs no-flap/cap(PRK,Trans-PRK) procedures.Using the AI model,the overall AUC, classification accuracy (CA) and precision was 0.7, 57.9%and 58.2%, respectively, across the 4 groups.

Conclusions:

EZI might aid in early detection of ectasia and routine monitoring of postoperative wound healing after refractive surgery. Further studies with larger sample size and longer follow-up are needed.

Financial Disclosure:

None

Back to previous