advertisement
Editors Selection IGR 22-3
Clinical Examination Methods: ONH Hb levels from color fundus images and functional loss in glaucoma
Comment by Alon Harris on:
95993 Optic Nerve Head Hemoglobin Levels in Glaucoma: A Structural and Functional Correlation Study, Rocha JAG; Dias DT; Lemos MBC et al., Journal of Ophthalmology, 2021; 2021: 9916102
Technological innovations over the past several decades have provided various novel methods for assessing aspects of ocular blood flow and metabolism. However, translation of hemodynamic biomarkers into improved clinical care remains challenging due to limited longitudinal data linking them to structural and functional glaucoma progression along with high costs of equipment and required multi-dimensional analysis. In purist of adding clarity, Rocha and colleagues contribute novel data on optic nerve head (ONH) hemoglobin measurements and their relationship to structural and functional glaucoma progression.
Translation of hemodynamic biomarkers into improved clinical care remains challenging due to limited longitudinal data linking them to structural and functional glaucoma progression
The authors found applied automated analysis (Laguna ONhE software) to identify significant non-linear correlation between their estimates of glaucoma discrimination and visual field mean deviations (r2=0.3; p<0.001) and a linear relationship with peripapillary retinal nerve thickness (r2=0.6; p<0.001). The non-linearity of the visual field and (ONH Hb)-derived biomarkers across subgroups (mild, moderate, and advanced) are suggested to point to ONH hemoglobin reductions occurring before glaucomatous visual function changes are detectable. The authors should be congratulated for addressing the unmet need of linking vascular imaging biomarkers with clinical progression outcomes. The study is also well designed with a novel approach of computer-assisted analysis of imagery, a paradigm that will help make transformative leaps in data analysis in the coming years. Ophthalmology is a specialty that involves a variety of imaging devices and corresponding high volume of data. The future of ophthalmic research will undoubtedly utilize machine learning and transfer learning platforms to reveal hidden markers of risk and improve specificity and speed of analysis. The prospective nature and dual monitoring of both structural (OCT) and functional (visual field) are strengths, and the authors also provide a statistical power estimate of their study sample. A weakness of the study is that it lacks any ocular blood flow modalities for its comparisons, including the very relevant biomarkers of ONH capillary densities as assessed by OCT-angiography (OCTA). Knowing ONH hemoglobin outcomes in relation to ONH vascularity, IOP, blood pressure and ocular perfusion pressures would greatly strengthen the data and its potential conclusions. Also, as the authors note, specific care must be made to not apply these groups results across all glaucoma populations as differences in anatomy, pigmentation, and other factors altering optics, physiology, and access may influence outcomes. Looking forward, a comprehensive approach of data analysis via automated machine and transfer learning combined with physician expertise seems inevitable to crack the code of glaucoma.
