Sensitivity and specificity of swept source optical coherence tomography (SS-OCT) for detecting macular pathologies versus spectral domain (SD-OCT)

Abstract

Purpose: To evaluate the sensitivity and specificity of swept-source optical coherence tomography (SS-OCT) biometer compared with the gold standard spectral-domain (SD-OCT) for detecting macular pathology in cataract patients. Setting: Study design: Prospective, cross-sectional, observational, examiner-masked. Methods: The study included 132 subjects aged ≥50 years, who underwent pre-cataract surgery work-up. All subjects underwent fixation check retinal scans using SS-OCT biometer (IOLMaster 700) as well as full macular scans using Cirrus SD-OCT. Three independent masked examiners evaluated the scans if they were normal or had a suspected pathology. Different measures of diagnostic accuracy were calculated for three examiners. Results: True positive rate (sensitivity) ranged from 71.1 (32/45) to 79.2% (42/53) and false negative rate was between 20.8 (11/53) to 28.9% (13/45) for the three examiners. True negative rate (specificity) ranged from 86.8 (59/68) to 94.1% (64/68) and false positive rate was between 5.9 (4/68) and 13.2% (9/68). The fitted receiver operating characteristic (ROC) area ranged from 0.83 to 0.95. Conclusion: Using retinal SS-OCT biometer scans as a replacement of the dedicated macular SD-OCT for screening or diagnosing macular health would not be appropriate due to its low sensitivity. SS-OCT biometer may potentially fail to identify approximately one-fourth of patients who actually have the disease. Therefore, the final decision on macular health should be based on the gold standard SD-OCT scans. When full macular SD-OCT scans are not accessible, the limited retinal scan information from SS-OCT biometer may still provide useful insights into the macular health.