Software Can Determine Inflammation in Uveitis, Study Shows

Researchers presenting at the American Academy of Ophthalmology 2020 meeting reviewed the clinical acumen of automated software designed to quantify inflammation in patients with uveitis. The study shows that measures used to quantify the images have high sensitivity and specificity in determining uveitis activity. Investigators also demonstrated that, in this study, the software was superior to a clinical exam alone.

Researchers say the software calculated anterior chamber cell density (cells/mm³) and retinal leakage index using 6-months worth of ultra-wide fluorescein angiography (UWFA) and spectral domain optical coherence tomography (SD-OCT) images. Retinal leakage index was determined based upon the percentage of retinal leakage detectable on UWFA.

The machine reviewed 106 eyes of 53 patients with uveitis for the 6-month study period. Imaging measures were successfully obtained at all time points. The images were reviewed by clinical exam and masked readers as well as by clinical exam with clinician imaging assessment. 

The researchers found that clinical exam alone correlated moderately to the exam with clinical exam plus clinician imaging assessment (k =.56, P <.001) in determining activity, while masked readers grading correlated highly with clinical exam plus clinician imaging (k =.87, P <.001). The sensitivity of clinical exam and masked readers were 50% and 98%, respectively. The specificity of clinical exam and masked readers was 100% and 91%, respectively.

Investigators say this study shows that the software’s imaging quantification measures have high sensitivity and specificity in determining uveitis activity.

Disclosure: Several study authors declared affiliations with the pharmaceutical industry. Please see the original reference for a full list of authors’ disclosures.

Reference

Srivastava S, Deaner J, Baynes K, Venkat A, Sharma S. Prospective use of imaging quantification measures of inflammation in determining activity in uveitis patients. Presented at: American Academy of Ophthalmology 2020 Annual Meeting; November 13-15, 2020. Abstract PO156.