Kelly and Wendel’s landmark article describing pars plana vitrectomy, air–fluid exchange, and head-down positioning for full-thickness macular holes began a new era in the surgical treatment of macular diseases. Not surprisingly, their outside-the-box approach to a previously untreatable problem generated considerable skepticism. The academic analysis, put forth in a concurrent editorial by Stuart Fine, pointed out the modest anatomic (58%) and visual acuity (40%) improvement rates in this elderly cohort, most of whom had good vision in the fellow eye. He advised the surgical community to proceed slowly and cautiously and to await the results of randomized trials before proceeding with widespread adoption of the procedure.

Most retina surgeons viewed this as an opportunity to cure a large group of untreatable patients. Perhaps they took Maurice Landers III’s (personal communication, 1989) pithy advice, “When a new surgery is described, perform it frequently, before it stops working.” After all, history is replete with compelling surgical techniques—such as subfoveal removal of choroidal neovascular membranes—that failed to meet lofty original expectations. Fortunately for all, scores of subsequent reports, including a randomized study, validated the success of macular hole surgery. More importantly, improvements in anatomic success rates and visual acuity far eclipsed those of the pilot study.

The predominantly analog technology of the day limited Kelly and Wendel’s preoperative and postoperative anatomic and functional examination techniques (visual acuity, fundus photography, fluorescein angiography, and threshold visual field testing). Subsequent investigators documented the inadequacy of these techniques in differentiating some full-thickness macular holes from lamellar holes and pseudoholes. To address this, early advances in computerized imaging technology, particularly the scanning laser ophthalmoscope, enabled users to perform microperimetry and confocal fundus autofluorescence. These techniques improved diagnostic accuracy, resulting in a greater number of cases correctly selected for macular hole surgery.

Despite the early diagnostic successes with microperimetry and fundus autofluorescence (FAF), the introduction of optical coherence tomography (OCT) pushed these technologies aside. Clinical features—including rapid scanning, ease of use, ability to scan through undilated pupils, high-quality cross-sectional images through the retina, and quantitative measurements of both retinal and glaucoma conditions—soon made OCT indispensable for evaluating macular pathologic features. Thus, OCT quickly became the gold standard for the diagnosis and postoperative evaluation of full-thickness macular holes.

The predictive values of macular hole stage and size, as determined by time-domain OCT, long have been known. The recently introduced high-resolution spectral-domain OCT enables more detailed qualitative and quantitative evaluation of the outer retina, with reproducible imaging of the junction between the inner segment and outer segment (IS/OS) line. This important indicator of outer retinal function correlates with visual acuity in numerous retinal vascular diseases.

Despite its diagnostic accuracy and widespread adoption, OCT may inadequately evaluate some full-thickness holes, lamellar holes, and pseudoholes. Some authors question the ability of OCT to differentiate the residual photoreceptors found in a freshly developed stage 2 or 3 hole from the remaining outer tissue in a pseudohole or lamellar hole. Others have noted similar FAF patterns in pseudoholes and lamellar holes, suggesting similar amounts of residual neurosensory pigment, despite different measured macular thicknesses. These observations suggest that OCT provides us with an incomplete representation of macular tissue.

This brings us back to fundus autofluorescence and its role in evaluating macular disease. Ageing, postmitotic cells throughout the body (e.g., neurons, cardiac muscle cells) accumulate partially degraded materials known as lipofuscin. Lipofuscin accumulation within the retinal pigment epithelium represents partially phagocytosed rod and cone outer segment discs. The myriad molecular species within lipofuscin absorb short-wavelength visible light, exciting electrons to a higher energy orbital; the release of longer-wavelength photons accompanies the spontaneous return of the electrons to the ground state. This behavior results in low-intensity light emission known as fundus autofluorescence . Because the 2 naturally occurring macular carotenoid pigments, lutein and zeaxanthin, block FAF, measured FAF signal strength depends on both the lipofuscin content of the retinal pigment epithelium (RPE) and the carotenoid content of the overlying retina. FAF generally increases with RPE dysfunction because of its inability to process lipofuscin completely and decreases in disorders that lead to photoreceptor loss (no outer segments for the RPE to process).

The first in vivo FAF imaging (in 1989) used the confocal scanning laser ophthalmoscope and recent filter additions to digital photography systems allow the performance of FAF using slightly longer wavelengths. Although the 2 imaging systems excite different macular fluorophores, the overall images generated by the 2 different systems appear quite similar. FAF has become popular for the evaluation of retinal dystrophies and likely will provide critical quantitative data in upcoming dry AMD therapeutic trials.

In this edition of The Journal, Chung and associates report their results in evaluating successfully operated macular holes with FAF, microperimetry, and OCT. As expected, a decrease in FAF accompanied successful surgery and correlated well with improvements in microperimetry. However, the FAF improvement did not correlate with OCT visualized IS/OS lines or retinal thickness.

So why does OCT and FAF imaging suggest different outcomes? Although the carotenoid levels of successfully treated macular holes correlate with visual acuity, they remain less than normal. This suggests that photoreceptor sequestration of carotenoids, something not detected by OCT, may serve as an indicator of macular function. FAF also correlates poorly with macular thickness in lamellar holes and macular pseudoholes, conditions with intact photoreceptors and only moderate visual loss. So it may be that carotenoid content surrounding the photoreceptor axons, as opposed to the inner plexiform layer, makes the greatest contribution to both FAF and visual function.

FAF correlated poorly with Chung and associates’ estimated IS/OS defect, although Oh and associates discovered a better correlation between postoperative visual acuity and the area, rather than the diameter, of the IS/OS defect. Unfortunately, we cannot compare FAF with the area of IS/OS defect in Chung and associates’ study, although in light of Oh and associates’ report, such a correlation would be interesting.

Although Chung and associates noted full RPE recovery in their successfully operated macular holes, chronic holes sometimes lead to RPE atrophy. This results from decreased RPE phagocytosis of outer segments, the inability to sequester fluorophores, and in severe cases, irreversible cellular dysfunction with decreased FAF. Could an unexpectedly low preoperative FAF correlated with unfavorable OCT findings be a better predictor of poor postoperative RPE function and visual acuity in some eyes with chronic holes?

What about the future? Future studies will best incorporate several currently available imaging methods: OCT analysis, particularly of the IS/OS line; FAF; and microperimetry. We know that currently available imaging methods provide overlapping but dissimilar abilities to predict visual outcomes after successful macular hole surgery. Because OCT and FAF acquire different information regarding retinal morphologic and physiologic features, neither will provide a comprehensive analysis of the retina. Multivariate analysis may correlate preoperative findings with postoperative results better, thereby enabling the creation of multivariable probability formulas to help patients and surgeons make better-informed decisions and minimize disappointments. Perhaps future studies with existing but less frequently used methods—focal electroretinograms can identify at-risk fellow eyes in patients with macular holes —may expand further our diagnostic and prognostic capabilities.