Automated lesion segmentation and quantification for prediction of paradoxical worsening in patients with Tubercular serpiginous-like choroiditis
Purpose
To develop and evaluate a fully automated pipeline that analyzes color fundus images in patients with TB SLC for prediction of PW.
Methods
Retrospective study, patients with TB SLC with a follow-up of 9 months after initiation of ATT were included. A fully automated custom-designed pipeline developed after confirming reliability using Bland-Altman plots and intraclass correlation coefficient (ICC), the pipeline was deployed for all patients. Two automatic thresholding algorithms were applied and quantitative metrics were compared between PW+ and PW- groups using non-parametric tests and logistic regression model to assess binary classification performance.
Results
The study included 139 patients (139 eyes; 92 M and 47 F; mean age: 44.8 ± 11.3 Yrs) with TB SLC. The PW+ group had significantly higher mean lesion area (p=0.0152), mean pixel intensity (p=0.0181), and integrated pixel intensity (p<0.0001) compared to the PW- group. Using a sensitivity optimized threshold cut-off for mean pixel intensity, an area under the curve of 0.87 was achieved (sensitivity: 96.80% and specificity: 72.09%).
Conclusion
Automated calculation of lesion metrics such as mean pixel intensity and segmented area in TB SLC is a novel approach with good repeatability in predicting PW during the follow-up.
Conflict of interest
No
Authors 1
Last name
DANGI
Initials of first name(s)
M
Department
Advanced Eye Centre, Post Graduate Institute of Medical Education and Research
City
Chandigarh
Country
India
Authors 2
Last name
KALRA
Initials of first name(s)
G
Department
Cole Eye Institute , Cleveland Clinic,
City
Ohio
Country
United States
Authors 3
Last name
AGARWAL
Initials of first name(s)
A
Department
Eye Institute, Cleveland Clinic Abu Dhabi
City
Abu Dhabhi
Country
United Arab Emirates
Authors 4
Last name
MARCHESE
Initials of first name(s)
A
Department
San Raffaele Scientific Institute
City
Milan
Country
Italy
Authors 5
Last name
BANSAL
Initials of first name(s)
R
Department
AEC, PGIMER
City
Chandigarh
Country
India
Authors 6
Last name
GUPTA
Initials of first name(s)
V
Department
AEC, PGIMER
City
Chandigarh
Country
India
This website uses cookies to ensure you get the best experience on our website.
Learn more