This content is only available for ASCRS Members
This content from the 2020 ASCRS Virtual Annual Meeting is only available to ASCRS members. To log in, click the teal "Login" button in the upper right-hand corner of this page.
Filed Under
Cornea
graft failure
endothelial keratoplasty
artificial intelligence
2020 paper presentation
Purpose
To evaluate the diagnostic performance of an artificial intelligence (AI) algorithm in the diagnosis of corneal graft rejection using anterior segment optical coherence tomography (ASOCT) images.
Methods
Thirty-six eyes of 36 patients with corneal grafts were imaged using ASOCT machine (Envisu R2210, Bioptigen, Buffalo Grove, IL, USA). A deep learning AI algorithm (Bascom Palmer AI, version 1.0, Miami, FL) was used to evaluate the ASOCT scans of these grafts. Results were compared to the clinical diagnosis given by corneal experts at Bascom Palmer Eye Institute. Prediction scores for diagnosis of corneal graft rejection and receiver operating curves (ROC) were generated.
Results
The cornea experts diagnosed 22 grafts as healthy grafts and 14 as rejected grafts. The algorithm correctly diagnosed all healthy grafts and 12 out of the 14 rejected grafts. For the diagnosis of corneal graft rejection, the AI algorithm achieved an area under the curve (AUC) of 0.9231, sensitivity of 84.62 % and specificity of 100 %.
Conclusion
The deep learning AI algorithm is a novel autonomous technique that can accurately diagnose cornea grafts rejection.
To evaluate the diagnostic performance of an artificial intelligence (AI) algorithm in the diagnosis of corneal graft rejection using anterior segment optical coherence tomography (ASOCT) images.
Methods
Thirty-six eyes of 36 patients with corneal grafts were imaged using ASOCT machine (Envisu R2210, Bioptigen, Buffalo Grove, IL, USA). A deep learning AI algorithm (Bascom Palmer AI, version 1.0, Miami, FL) was used to evaluate the ASOCT scans of these grafts. Results were compared to the clinical diagnosis given by corneal experts at Bascom Palmer Eye Institute. Prediction scores for diagnosis of corneal graft rejection and receiver operating curves (ROC) were generated.
Results
The cornea experts diagnosed 22 grafts as healthy grafts and 14 as rejected grafts. The algorithm correctly diagnosed all healthy grafts and 12 out of the 14 rejected grafts. For the diagnosis of corneal graft rejection, the AI algorithm achieved an area under the curve (AUC) of 0.9231, sensitivity of 84.62 % and specificity of 100 %.
Conclusion
The deep learning AI algorithm is a novel autonomous technique that can accurately diagnose cornea grafts rejection.
View More Presentations from this Session
This presentation is from the session "SPS-101 Corneal Procedures & Diagnostics: EK, PK, CXL, Other" from the 2020 ASCRS Virtual Annual Meeting held on May 16-17, 2020.