Automated Optic Nerve Head Hemoglobin Measurements versus General Ophthalmologists Evaluation to Differentiate Glaucomatous from Large Physiological Cupping: A Diagnostic Performance Comparative Study

Janaina Andrade Guimarães Rocha,1,2,* Cláudia Gomide Vilela de Sousa Franco,3 Leopoldo Magacho,3,4 Augusto Paranhos Jr,1 Fábio Nishimura Kanadani,1,2,5 Carolina Pelegrini Barbosa Gracitelli,1,6,* Tiago Santos Prata1,2,5,7,* 1Department of Ophthalmology, Federal University of São Paulo, São Paulo, São Paulo, Brazil; 2Department of Glaucoma, Glaucoma Institute, Belo Horizonte, Minas Gerais, Brazil; 3Department of Ophthalmology, Federal University of Goiás, Goiânia, Goiás, Brazil; 4Department of Glaucoma, Excellence in Ophthalmology, Goiânia, Goiás, Brazil; 5Department of Ophthalmology, Mayo Clinic, Jacksonville, FL, USA; 6Department of Glaucoma, Vera Cruz Oftalmologia, Campinas, São Paulo, Brazil; 7Glaucoma Unit, Opty Group Brazil, São Paulo, São Paulo, Brazil*These authors contributed equally to this workCorrespondence: Janaina Andrade Guimarães Rocha, Department of Ophthalmology and Visual Science, Federal University of São Paulo, Botucatu street, 821 Vila Clementino, São Paulo, São Paulo, 04023-062, Brazil, Tel +551150852010, Fax +551150852000, Email janainaaguimaraes@yahoo.com.brPurpose: The aim of this study is to compare the diagnostic performance in differentiating patients with glaucoma from those with presumed large physiological optic disc cupping (LPC), using optic nerve head hemoglobin levels (ONH Hb), as a screening method, versus the evaluation of general ophthalmologists.Patients and Methods: Twenty general ophthalmologists evaluated PowerPoint images of 40 patients with glaucoma and 40 presenting LPC. Presentation of patient’s exams were distributed as follows: Group 1 (GI): color retinography (CR), Group 2 (GII): CR + visual field (VF), Group 3 (GIII): CR + optical coherence tomography (OCT), Group 4 (GIV): CR + VF + OCT. The Laguna ONhE software was used to estimate ONH Hb based on CR. Main outcomes were the comparison of sensitivity and accuracy between general ophthalmologists’ evaluation and the glaucoma discriminant function (GDF) index from Laguna ONhE and also the agreement between examiners (Kappa statistics).Results: Laguna ONhE GDF index demonstrated higher sensitivity values (GI- 90%; GII-90%; GIII-100%; GIV-100%) comparing to all groups (GI-59%; GII-86.5%; GIII-86.5%; GIV-68.5%). In GI, in which it was observed the worst accuracy result (64.8%), we found 75% for GDF. In GII, the accuracy was 81.3% and we found 55% for GDF. The highest agreement was in GII (Kappa=0.63; 95% CI=0.53– 0.72), and the lowest in GI (Kappa=0.30; 95% CI=0.20– 0.39).Conclusion: Laguna ONhE software, a low-cost and non-invasive method, showed good sensitivity and great utility as a screening method in differentiating patients with glaucoma from those with LPC, compared with evaluation of general ophthalmologists.Keywords: glaucoma, diagnosis, retinography, colorimetric analysis