AI algorithm based on Corvis ST measurements may predict perimetric progression
Click Here to Manage Email Alerts
Key takeaways:
- An AI algorithm based on the Corvis ST dynamic corneal response parameters predicted glaucomatous visual field progression after prostaglandin analog treatment.
- Early identification of fast progressors may allow ophthalmologists to be more aggressive with treatment.
VILAMOURA, Portugal — An AI algorithm based on the Corvis ST dynamic corneal response parameters showed the ability to predict glaucomatous visual field progression 1 month after treatment with prostaglandin analogs.
“We analyzed 49 eyes with open-angle glaucoma and 16 with ocular hypertension. VF data were collected from 2013 to 2019, and in the AI algorithm, all forms of defining VF progression were considered,” Marta Isabel Martínez-Sánchez, MD, said at the European Society of Cataract and Refractive Surgeons winter meeting.
Different AI models were tested to achieve the best predictions evaluated by their accuracy and area under the curve, and different variables were used, including IOP measured with Goldmann applanation tonometry, the Ocular Response Analyzer (ORA, Reichert) and the Corvis ST (Oculus) dynamic corneal response parameters registered before and 1, 3 and 6 months after initiating therapy. The predictive variables were analyzed to assess which were relevant for the final prediction of visual field (VF) progression.
“The most accurate data predictors were those registered with Corvis 1 month after initiating the treatment, predicting glaucoma VF progression with an accuracy of 86.2%,” Martínez-Sánchez said. “Across different algorithms, the main variables involved in the prediction were the highest concavity Corvis ST parameters such as the highest concavity (HC) dArc length, HC time and HC deflection length.”
Forcing the model by including baseline parameters or additional variables such as IOP values or corneal hysteresis measured with the ORA did not improve the precision of the model.
“The early identification of potential fast progressors could have an evident clinical benefit since an optimization of the resources can be obtained by focusing more on these eyes,” Martínez-Sánchez said. “In addition, the prognosis of our patients can be improved because we can be more aggressive with the treatment and follow-up in eyes with predicted high risk of progression.”
Collapse
Martínez-Sánchez MI, et al. Prediction of perimetric progression in open angle glaucoma patients based on high-speed corneal imaging (Corvis ST) measurements using artificial intelligence. Presented at: European Society of Cataract and Refractive Surgeons winter meeting; March 10-12, 2023; Vilamoura, Portugal.
Read more about
Click Here to Manage Email Alerts