Your browser doesn't support javascript.
loading
New simulation software to predict postoperative corneal stiffness before laser vision correction.
Francis, Mathew; Shetty, Rohit; Padmanabhan, Prema; Vinciguerra, Riccardo; Vinciguerra, Paolo; Lippera, Myrta; Matalia, Himanshu; Khamar, Pooja; Chinnappaiah, Nandini; Mukundan, Deepa; Nuijts, Rudy M M A; Sinha Roy, Abhijit.
Afiliação
  • Francis M; From the Imaging, Biomechanics and Mathematical Modelling Solutions, Narayana Nethralaya Foundation, Bangalore, India (Francis, Sinha Roy); Department of Corneal and Refractive surgery, Narayana Nethralaya, Bangalore, India (Shetty, Matalia, Khamar, Chinnappaiah); Medical Research Foundation, Sankara Nethralaya, Chennai, Tamil Nadu, India (Padmanabhan, Mukundan); Humanitas San Pio X Hospital, Milan, Italy (R. Vinciguerra); The School of Engineering, University of Liverpool, Liverpool, United Kin
J Cataract Refract Surg ; 49(6): 620-627, 2023 06 01.
Article em En | MEDLINE | ID: mdl-36791274
ABSTRACT

PURPOSE:

To develop a new virtual surgery simulation platform to predict postoperative corneal stiffness (Kc mean ) after laser vision correction (LVC) surgery.

SETTING:

Narayana Nethralaya Eye Hospital and Sankara Nethralaya, India; Humanitas Clinical and Research Center, Italy.

DESIGN:

Retrospective observational case series.

METHODS:

529 eyes from 529 patients from 3 eye centers and 10 post-small-incision lenticule extraction (SMILE) ectasia eyes were included. The software (called AcuSimX) derived the anisotropic, fibril, and extracellular matrix biomechanical properties (using finite element calculation) of the cornea using the preoperative Corvis-ST, Pentacam measurement, and inverse finite element method assuming published healthy collagen fibril orientations. Then, the software-computed postoperative Kc mean was adjusted with an artificial intelligence (AI) model (Orange AI) for measurement uncertainties. A decision tree was developed to classify ectasia from normal eyes using the software-computed and preoperative parameters.

RESULTS:

In the training cohort (n = 371 eyes from 371 patients), the mean absolute error and intraclass correlation coefficient were 6.24 N/m and 0.84 (95% CI, 0.80-0.87), respectively. Similarly, in the test cohort (n = 158 eyes from 158 patients), these were 6.47 N/m and 0.84 (0.78-0.89), respectively. In the 10 ectasia eyes, the measured in vivo (74.01 [70.01-78.01]) and software-computed (74.1 [69.03-79.17]) Kc mean were not statistically different ( P = .96). Although no statistically significant differences in these values were observed between the stable and ectasia groups ( P ≥ .14), the decision tree classification had an area under the receiver operating characteristic curve of 1.0.

CONCLUSIONS:

The new software provided an easy-to-use virtual surgery simulation platform for post-LVC corneal stiffness prediction by clinicians and was assessed in post-SMILE ectasia eyes. Further assessments with ectasia after surgeries are required.
Assuntos

Texto completo: 1 Base de dados: MEDLINE Assunto principal: Inteligência Artificial / Córnea Idioma: En Ano de publicação: 2023 Tipo de documento: Article

Texto completo: 1 Base de dados: MEDLINE Assunto principal: Inteligência Artificial / Córnea Idioma: En Ano de publicação: 2023 Tipo de documento: Article