Measuring intraocular inflammation with standard care imaging

Reviewed by Ian Han, MD

Physicians may soon have new tools for quantifying inflammation associated with uveitis. Automated algorithms have been developed that can correlate im­aging-based measures of uveitis activ­ity with visual acuity (VA) and visual function. These algorithms work with commercially available, standard care imaging and their metrics have potential use as surrogate outcome measures in uveitis clinical trials.

Current measures of inflammation are limited because they are ordinal and subjective, according to Cindy Chen, BA, of the Cleveland Clinic’s Cole Eye Institute in Cleveland, Ohio. She and a team of col­leagues have been working on automated and semi­automated imaging-based quantifications of inflam­mation to produce continuous and objective mea­sures of intraocular inflammation. Their goals are to demonstrate that these quantitative metrics are not only reliable and reproducible but relate well to visual acuity and measures of function.

Qualifying inflammation

The investigators performed a prospective obser­vational case series, the Imaging Quantification of Inflammation study, that included patients with active uveitis who received standard-of-care medi­cations and were followed for 6 months. The patients underwent standard clinical examinations and optical coherence tomography (OCT), anterior-segment OCT (AS-OCT), and ultrawidefield fluorescein angiogra­phy (UWFFA) at baseline and 1, 3, and 6 months. At each visit, their VA was measured, and the patients completed the National Eye Institute Visual Func­tion Questionnaire-25 (NEI VFQ-25).

Investigators created customized automated soft­ware that can continuously image and measure intra­ocular inflammation, specifically anterior-chamber cell quantification with AS-OCT, UWFFA leakage quantification, and OCT fluid quantification, with the focus on the latter 2 measures. They then corre­lated the imaging-based measures of ocular inflam­mation with the VA and visual function scores.

Chen was able demonstrate the method of semi­automated quantification of vascular leakage using UWFFA. Using this method, the software developed the total leakage index, defined as the total leakage area divided by the total region of interest analyzed. They also devised a macular-centered retinal leak­age index; the software provides an image with 3 regions of interest circled, with the first represent­ing the central 3 disc diameters, the second repre­senting 6 disc diameters, and the third representing 9 disc diameters. The leakage index was measured in each diameter.

With semiautomated volumetric fluid analysis using spectral-domain OCT, a fluid index was generated that was the sum of the intraretinal and subreti­nal fluid divided by the actual retinal volume from a 6-mm cube scan, Chen explained.

Study results

Eighty six eyes from 43 patients (aver­age age, 46 years) were included in the study. The most frequent diagnoses were idiopathic uveitis, sarcoid uveitis, and birdshot chorioretinopathy. The aver­age logarithm of the minimum angle of resolution (logMAR) VA was 0.22 and 0.25 at baseline and at 6 months, respectively.

Semiautomated quantification of OCT fluid demonstrated good correlation with visual acuity.

“We found that the OCT fluid index was corre­lated moderately with the logMAR VA at all evalu­ations,” Chen said. “A worsening fluid index was correlated positively with worsening VA (r = 0.4, P < 0.001).”

Promising results were also found when relat­ing vascular leakage on UWFFA with visual acuity. Specifically, an overall higher total retinal leakage index was correlated with worse VA (r = 0.4, P < 0.001), and a similar relationship was found when evaluating macular leakage in the central 3 disc diameter zone (r = 0.489, P < 0.001).

“The highest correlation between the macular leakage in the central 3 disc diameters and the wors­ening logMAR VA was seen at baseline in patients with active uveitis (r = 0.617, P < 0.001),” Chen said.

In addition to visual acuity, visual function was also mildly correlated to vascular leakage at all vis­its, with a higher total retinal leakage (r = –0.156, P = 0.015) and macular leakage in the central 3 disc diameter zone (r = –0.237, P = 0.0002) sug­gesting worse visual function.

Between the retinal leakage and OCT fluid mea­sures themselves, the investigators also found a mild correlation. A greater total retinal leakage index was positively correlated with a greater OCT fluid index (r = 0.272, P < 0.001), and macular leakage in the central 3 disc diameters was corre­lated mildly with greater OCT fluid (r = 0.343, P < 0.001), Chen noted.

The take-home messages from this study are as follows:

• Macular leakage at the 3-disc diameter on FA and the retinal fluid index on OCT are correlated moderately with VA and visual function.
• These continuous measures of inflammation are continuous and can quickly be generated from commercial imaging modalities.
• The parameters of inflammation quantified using an automated algorithm can be used as surrogate outcome measures.

CINDY CHEN, BA

e: srivass2@ccf.org

This article is adapted from Chen’s presentation at the American Academy of Ophthalmol­ogy’s 2020 virtual annual meeting. Chen has no financial interest in this subject matter; the study was funded by Santen Inc.