RPS: From the Podium to the Practice

This new series recaps the insights on clinical care and research presented at our annual meeting.

JACK PERSICO, EXECUTIVE EDITOR · ANDREW MATHIS, PhD, MEDICAL EDITOR

With the sun-dappled waters of the Bahamas as its backdrop, the 5th annual Retinal Physician Symposium took place in a convivial atmosphere that fostered the free exchange of ideas and inspiration. Hosted by this publication each spring, RPS gives attendees the unique opportunity to interact one-on-one with prominent retinal specialists at the forefront of both research and clinical practice. The faculty of thought leaders from around the world shared their expertise on diabetic macular edema, AMD, vein occlusion, melanoma, laser therapy, surgical techniques, and more.

Highlights of this year's symposium, which took place March 25-28 at the Atlantis resort, will be presented in a series of features beginning this month called "RPS: From the Podium to the Practice." Our first installment covers the clinical utility of spectral-domain optical coherence tomography, laser and pharmacologic treatment options for diabetic macular edema, and a preview of new implantable steroid therapies for DME and other chronic inflammatory conditions. Contributors to this article include:

• Jeffrey Heier, MD, a vitreoretinal surgeon at Ophthalmic Consultants of Boston and president of the Center for Eye Research and Education.

• Diana V. Do, MD, Assistant Professor of Ophthalmology at the Wilmer Eye Institute, Johns Hopkins School of Medicine in Baltimore.

• Pravin Dugel, MD, Managing Partner at Retinal Consultants of Arizona in Phoenix.

• Darius Moshfeghi, MD, Director of Ocular Oncology/Pediatric Vitreoretinal Surgery and Assistant Professor at Stanford University in Menlo Park, CA.

CLINICAL APPLICATION OF DIAGNOSTIC ADVANCES IN RETINAL DISEASE: PART 2

Befitting the conference's goal of bridging the research and clinical worlds, the educational program began with a lecture by Jeffrey Heier, MD called "Clinical Application of Diagnostic Advances in Retinal Disease: Part 2" — intentionally designated as Part 2 because "it takes us past the images and shows us how we're incorporating these into our clinical practices," Dr. Heier explained. "Initially, all of us were really enamored of spectral-domain OCT because the images were so spectacular" compared to earlier OCT technology; the next step was to document their clinical utility."

With that premise in mind, Dr. Heier described several instances in which SD-OCT yielded clinical insights not possible with time-domain OCT, likening SD-OCT's impact to a positive feedback loop of sorts in which therapeutic advances have led to diagnostic advances that have in turn led to additional therapeutic advances, and so on. Macular degeneration diagnosis and management has been the biggest beneficiary of this phenomenon, he said.

When the ANCHOR and MARINA studies of anti-VEGF therapy began, only 10% of investigational sites had OCT of any type; clinical decisions largely relied on fluorescein angiography and visual observation. Just a few years later, OCT has emerged as the essential tool used to inform clinicians of retinal thickness and the presence of fluid. The increasing reliance on that sort of data has accelerated development of ever more sophisticated SD-OCT technology, as time-domain OCT suffers from deficiencies in reproducibility, small sampling size, and low resolution. Any measurement below 50 microns was considered "noise," he said.

SD-OCT reduces or eliminates those problems, while also providing the ability to perform volume scans and measure large amounts of data in a short period of time. "The key, however, is in the reproducibility," Dr. Heier said. Registration based on ocular anatomy instead of patient fixation allows precise mapping of the same structures on repeat testing to document disease progression and response to treatment. Although SD-OCT is not an endpoint that's accepted yet by FDA, ongoing research is seeking to validate SD-OCT findings so that they may be incorporated into future clinical trials.

The higher resolution of SD-OCT gives clinicians the ability to observe changes in the inner segment/outer segment photoreceptor layer that would have been impossible to interpret with time-domain OCT, and multimodality imaging allows clinicians to find subtle defects that would have previously gone unnoticed (Figure 1). Dr. Heier demonstrated this point via several case presentations in which SD-OCT was essential in assessing often-subtle changes, including examples of cystoid macular edema (Figure 2), diabetic macular edema, CRVO (Figure 3), macular hole repair, and others.

Figure 1. High myope with new-onset scotoma and a small hemorrhage, unable to undergo fluorescein angiography due to allergy. Multimodality SD-OCT revealed no defects surrounding the hemorrhage and no RPE deficit.

Figure 2. Cystoid macular edema (top) and response to treatment (bottom).

Figure 3. Serial SD-OCT images of a CRVO patient with complaints of visual fluctuation throughout the day. From left to right, scans at 8am, 10am, 12pm and 2pm document an 81-micron decrease over an eight-hour period. Precise registration allows for repeat imaging at identical locations.

"We're seeing things we never saw before," Dr. Heier said, describing as an example a patient who presented with 20/200 acuity and normal results on funduscopy, fluorescein angiography, and autofluorescence; however, some disruption in the photoreceptor layer was revealed by SD-OCT. Even when a diagnosis is lacking, he explained, SD-OCT may provide evidence of a change in anatomic status that can be described to the patient.

SD-OCT is proving especially valuable in recent dry AMD trials. Visual acuity alone is not a very good endpoint in studies of geographic atrophy but SD-OCT combined with autofluorescence is helping researchers to understand the anatomic changes described as autofluorescence, as well as symmetry of GA lesions and correlation between autofluorescence and SD-OCT scans. Though investigational at the moment, research of this sort by Frank Holz, MD, and others could have tremendous application for future treatments or approaches to dry AMD because certain autofluorescent lesions have almost no growth while others have tremendous growth.

"Technological advances are really only helpful if they allow us to take care of patients better," Dr. Heier concluded. "It's our belief in our clinic that these advances clearly enable us to take better care of patients. They enable us to detect disease better, monitor progression better, and they give us confidence that what we're seeing is real and reproducible."

ANTI-VEGF THERAPY FOR MACULAR EDEMA

Diana Do, MD presented an update on VEGF inhibitors for diabetic macular edema (DME). In her presentation, data from the READ-2 (Ranibizumab for Edema of the mAcula in Diabetes) trial and from the VEGF-Trap-Eye phase 1 trial were reviewed.

The READ-2 Trial was a phase 2 clinical study evaluating ranibizumab vs focal laser photocoagulation vs combination therapy (ranibizumab followed by focal laser). In the READ-2 trial, A total of 126 patients were randomized. Group 1 received intravitreal injections of 0.5 mg of ranibizumab at baseline and at months 1, 3, and 5; group 2 received focal laser treatment at baseline and at month 3 if DME was still present; and group 3 received 0.5 mg of ranibizumab combined with focal laser at baseline and at month 3. The primary outcome of the study was the change in best corrected visual acuity at month 6.

One hundred fifteen of the 126 original subjects were followed up through month 6; the other 11 patients had discontinued from the study. The greatest gain in BCVA was seen in group 1 (Figure 4), which received ranibizumab alone, and the ranibizumab treatment group had a mean gain of 7.62 letters at month 6. Group 2 (focal laser) lost a mean of 1.07 by month 6. Group 3 (combination treatment group) had a 6-month BCVA letter gain of 3.8.

Figure 4. Change in best-corrected visual acuity at the 3- and 6-month intervals of the READ-2 trial.

Changes in excess retinal thickness as measured on OCT showed similar results in that the ranibizumab treatment group had the greatest reduction in retinal thickening (Figure 5). Group 1 (ranibizumab) had a 57% reduction in excess retinal thickness at month 6, the primary endpoint of the study. Groups 2 (focal laser) and 3 (combination treatment) showed reductions of 11% and 42%, respectively.

Figure 5. Retinal thickness change data at month 6.

In addition to bioactivity, ranibizumab was not associated with any ocular or systemic adverse events in the READ-2 Trial. A phase 3 clinical trial with long-term follow-up is necessary to determine whether the benefits of ranibiziumab therapy can be sustained for longer periods of time.

Dr. Do also discussed the VEGF Trap-Eye phase 1 study of 5 eyes with DME. In this open label study, 5 patients each received a single 4 mg injection of VEGF Trap-Eye and were then followed for 6 weeks, the primary endpoint of the study. Visual acuity, retinal thickness, and safety assessments were monitored.

In this pilot study, there were no serious ocular or systemic adverse events related to VEGF Trap-Eye. VEGF Trap-Eye demonstrated bioactivity and a single intravitreal injection was associated with a significant reduction in excess retinal thickness in all 5 patients, as well as a gain in VA. Dr. Do presented an example of 1 study subject who received a single injection and experienced complete resolution of leakage as seen on fluorescein angiogram and complete resolution of retinal thickness on OCT.

A phase 2 randomized clinical trial known as the DA VINCI study is now underway in which 200 patients will be randomized to 5 treatment arms (4 groups will receive different doses of VEGF Trap Eye with variable dosing schedules and 1 group will receive standard of care, which is focal laser photocoagulation). The primary endpoint of the study is the change in best corrected visual acuity at month 6.

DRCR TRIAL OF FOCAL GRID LASER VS. INTRAVITREAL TRIAMCINOLONE IN DME

Another presentation by Diana Do, MD offered an update on a DRCR network phase 3 randomized clinical trial evaluating the safety and efficacy of preservative-free intravitreal triamcinolone (IVT) vs. focal/grid laser photocoagulation for diabetic macular edema. A total of 840 eyes from 693 subjects were enrolled in the study at 88 clinical sites.

Patients were randomized to one of three groups: focal laser photocoagulation, 1 mg triamcinolone and 4mg triamcinolone and followed every 4 months and retreated if DME was still present on OCT. Treatment could be withheld if DME had resolved with a retinal thickness of less than 225μm on OCT or an adverse event occurred. The primary outcome measurement was mean change in best corrected visual acuity (BCVA) through 2 years, with secondary outcomes of change in retinal thickness on OCT and safety measures such as cataract progression and elevation of IOP. Approximately 88% of subjects completed the 2-year follow-up. The mean baseline central retinal thickness was 424um.

Mean change in BCVA during 2-year period was a 1-letter gain for the laser-treated patients, a 2-letter loss for the 1 mg IVT group and a 3-letter loss for the 4 mg group. At the 4-month interval, the 4mg IVT-treated eyes demonstrated greater gain in VA than laser-treated eyes; however, by month 8 the gain narrows and at month 12 the VA gain in the laser-treated eyes exceeds that of either IVT dose, a finding that persisted out to the 24-month mark. Laser-treated eyes were more likely to gain 10 or more letters and less likely to lose 10 or more letters than IVT-treated eyes. OCT retinal thickness changes followed similar curves. Patients in the 4mg IVT group showed a greater reduction of thickness at month 4, but the difference was negligible at month 8. Between months 12 and 16, the laser and IVT groups switched paths and ultimately at 24 months the laser-treated eyes had greater reduction in retinal thickness than either the 1 or 4mg doses of IVT (Figure 6). A subgroup analysis found that results were unaffected by pretreatment phakic status.

Figures 6 (left) and 7 (right) show central subfield thickness changes in a DRCR network phase 3 study of 1 mg and 4 mg doses of preservative-free intravitreal triamcinolone vs. focal/grid laser photocoagulation for diabetic macular edema.

Noting the greater gains in visual acuity and reduction of retinal thickness among IVT patients at the 4-month follow-up, Dr. Do commented that "there is a bioactivity signal that could suggest the possibility that combination therapy of intravitreal triamcinolone with focal laser may be beneficial." The DRCR network is conducting a phase 3 clinical trial where combination therapy is employed, she said.

Regarding safety and adverse effects, there were no cases of endophthalmitis or pseudoendophthalmitis in any of the three groups. However, IOP increase was greater with IVT patients, particularly those in the 4mg group, for which 33% of patients experienced IOP rise greater than 10mm Hg. Glaucoma surgery was needed in 2% of this group. Steroid-treated eyes were more likely to develop cataract during the study. Rates of cataract surgery among patients who were phakic at baseline were 13% for laser-treated patients, 23% for IVT 1mg patients and 51% for those in the IVT 4mg group.

Citing the greater efficacy and lower side effect rates for laser-treated patients in the study, Dr. Do commented that "focal grid laser photocoagulation still is the standard of care for eyes with diabetic macular edema and clinical trials evaluating new treatments should always include focal laser as the benchmark for comparison."

PRP TIPS AND TRICKS

Dr. Moshfeghi's presentation on panretinal photocoagulation (PRP) set out to explain PRP techniques, go over the advantages and disadvantages of the recently conducted PASCAL study, and suggest adjuvant therapy for PRP.

Dr. Moshfeghi began with a description of the process of panretinal photocoagulation. First describing how lasers work and how laser energy is applied to the eye, he then described how tissue reacts to laser application. Higher-powered lasers applied over a short time or lower-powered lasers applied over a longer time can deliver the same amount of energy, but can produce radically different effects on lasered tissue.

Laser-tissue interactions, Dr. Moshfeghi continued, fall into 3 categories: photochemical, thermal, and thermoacoustic. Photochemical interaction occurs when low-power light of long duration is applied to specific chromophores or chemical bonds, affecting the structure of the molecules without hear or explosion. Thermal interaction is marked by low- to medium-power laser and duration, which results in heating and denaturation of proteins, but not acoustic effects or lysis of chemical bonds. Finally, thermoacoustic interaction occurs when short to ultrashort duration and high local temperature causes explosion or vaporization. There is the creation of a gas bubble, as well as lysis of chemical bonds.

Most noninvasive laser therapy is performed with a slit-lamp delivery system, Dr. Moshfeghi explained, and then turned to the issue of how and why delivery systems for PRP need to be redesigned. First, he said, there is a new generation of tools and software that can be implemented in the process. Second, there is a need for increased efficiency in the office while reimbursements are decreasing. And third, a desire exists to improve patient outcomes and satisfaction, including duration of treatment, precision, safety, and comfort.

The semiautomated retinal patterned scanning laser delivery system, called PASCAL, answers these issues, Dr. Moshfeghi said. Until now, he said, single-spot PRP has "ruled the day" since commercial laser technology was introduced over 30 years ago. PASCAL, by contrast, is based on the principle of applying more than 1 spot at a time. Designed by Mark Blumenkrantz, MD, and colleagues, PASCAL is microprocessor-driven and produces a variety of scalable patterns viewable on a computer screen for selection by the retinal physician. Control comes in the form of an electronic joystick. Dr. Moshfeghi presented several slides that showed the graphical user interface that comes with the PASCAL system. Dr. Moshfeghi then discussed the initial use of PASCAL in rabbit eyes and the first human clinical trials, which were conducted in Mexico.

"When you're using a smaller duration," Dr. Moshfeghi said, "you are reducing the duty cycle of the laser, which reduces the total amount of energy and thus favorably influences the impact on the thermal relaxation threshold — basically, the collateral damage caused by conventional PRP." He demonstrated the outcome of this new application of laser with data from the Mexican trial, showing that the pulse duration using PASCAL was 20% of that of standard single-spot laser therapy. In practice, this meant that PRP was 3.8 times quicker with PASCAL and macular grid therapy was 7.8 times quicker. Furthermore, while mean patient discomfort did not differ between singlespot and PASCAL, patient discomfort (ranked on a scale of 0 to 5) was 2.82 for single spot and 0.61 for PRP.

Dr. Moshfeghi continued by comparing the effects of conventional and shorter-duration burns, stating that traditional burns were larger, more heterogeneous, and more likely to cause pain. Showing the relationship between pulse duration and thermal diffusion length, thermal diffusion length increased tenfold as pulse duration increased from 1 to 100 milliseconds. Cooling also occurs in a different manner, related to spot size, power, and pulse duration. Dr. Moshfeghi presented several slides containing data supporting his theses. Of particular interest were the safety thresholds demonstrated on these graphs, as well as a table that demonstrated the size and intensity of laser spots depending on the pulse duration and megawattage of the beam (Figures 8 and 9).

Figures 8 (left) and 9 (right). Altering the power and pulse duration but keeping spot size constant produces significant variation in burn size and lesion characteristics.

Dr. Moshfeghi also demonstrated that with the current standard of care, in which the pulse rate is 100 milliseconds, the burn size is significantly larger than with smaller durations, as would be offered by the PASCAL. Other slides included OCT images of laser burns on retinas, including burns made with the 132-μm laser, to show comparative retinal damage and "collateral damage" caused with longer pulse durations.

He also presented a series of case images where the PASCAL was used instead of traditional panretinal photocoagulation. Conditions presented included retinal tear with subclinical detachment, clustered peripheral tears, neovascularization association with ischemic branch retinal vein occlusion, and lattice degeneration.

Dr. Moshfeghi's presentation concluded with a nod toward what he termed the "concept of conservative PRP (Figure 10), which would follow several guidelines. These would include: the notion that spot size should not be equivalent to final burn size, taking into account the aforementioned cooling time of laser burns; lateral thermal blooming, or enlargement of burns, should be considered indicative of "collateral damage" to the choroid and inner retinal layers; if used, shorter pulses should produce more precise, spatially confined burns with less potential for both pain and field loss; and that new therapeutic opportunities exist for patients and physicians, including better quantitation of ablation and subthreshold effects.

Figure 10. Shorter pulses should produce precise, spatially confined burns with less potential for pain and field loss.

CURRENT STATE OF STEROID IMPLANTS

Pravin Dugel, MD, described the mechanism of action and unique aspects of four different implantable steroids: Posurdex, Retisert, Iluvien and I-Vation.

Posurdex was developed based on idea that dexamethasone is the most potent of the steroids. Although that may be up for debate, as studies are inconclusive, dexamethasone also has the benefit of high solubility and a short half-life, Dr. Dugel noted, and as a result it may lend itself well to an intravitreal implant suited to short-term drug delivery.

Originally developed by Oculex Pharmaceuticals, Posurdex is now part of the Allergan family of products and will mark the company's first entry into posterior segment therapy sometime later this year. Posurdex's advantages include a simple insertion procedure using a 22-gauge applicator through the pars plana, sustained drug release, the biodegradability of the implant, and a favorable long-term safety profile. It is a nonsurgical, office-based procedure that creates a sutureless, self-sealing wound. Posurdex provides an initial pulse of therapy followed by clear zeroorder pharmacokinetics and a steady, low rate of drug release for the remainder of the implant's lifespan.

The first trial of Posurdex investigated its efficacy in all forms of macular edema that last for over 90 days following laser or medical treatment, including uveitis, vein occlusion, pseudophakic macular, diabetic macular edema, and others. This initial study protocol required surgical implantation of the device. The most favorable responses in visual acuity improvement and retinal thickness decrease were seen in the uveitis and pseudophakic macular edema groups. The adverse effects were fairly mild (slight rise in IOP but no increase in the rate of cataract); however, the rate of vitreous hemorrhage — believed to be a result of the surgical procedure itself — was a cause for concern. Consequently, a new implant device and handheld applicator was developed that minimizes risk of vitreous hemorrhage without compromising efficacy.

The Retisert implant (Bausch & Lomb) is unique in its projected lifespan of 1000 days or longer, by far the longest-lasting of the intraocular steroid implants. Surgical implantation of the device is needed. Comprised of 0.59 mg fluocinolone acetonide, Retisert nearly eliminates the need for systemic medications in select patients, has very little systemic toxicity, and patient compliance is not a factor, Dr. Dugel said. Following on the heels of the gancyclovir device, Retisert is the only agent FDA approved for chronic noninfectious uveitis therapy.

The release characteristics are interesting, Dr. Dugel noted. The amount released is nominal, beginning with an initial dose 0.6 μg per day but steadily declines to about 0.3 to 0.4 μg per day. "Never have we seen a dose of steroids this low work so well," Dr. Dugel remarked.

Dr. Dugel reviewed the data from a trial of Retisert conducted in patients with bilateral uveitis and a history of systemic medication use; the worse eye was used in each case. Recurrence rate dropped rapidly and stayed low for up to 3 years, he noted. At 34 weeks 90% of patients noted improvement or stabilization of vision and 19% gained greater than 3 lines of vision. Such results were largely maintained for the duration of the study. At 3 years, 85% experienced improvement or stabilization of vision, with 23% of patients gaining 3 lines or more. A second trial noted similar or better results. Patients' use of immunosuppressive agents declined dramatically at the 34-week mark in both trials.

This modality can reduce the cumulative damage done by episodic recurrence of uveitis. Systemic absorption was negligible and the impact on patient's quality of life was remarkable, said Dr. Dugel. However, he noted, there was a price to be paid: within 3 years, 71% of patients developed an increase in IOP of 10 mm or more, 75% required a topical pressure-lowering medication, almost 40% required a surgical procedure to reduce IOP, and nearly all patients developed cataract. Calling Retisert the "heavy artillery" of the intraocular steroid armamentarium, Dr. Dugel noted it may be prudent to reserve its use for the most severe cases of uveitis.

Figure 11. Posurdex-treated patients with improvements in visual acuity of either 10 or more letters or 15 or more letters at day 90 and day 180.

The Iluvien (formerly Medidur) implant from Alimera Sciences is a newer option in implantable steroid therapy. This fluocinolone acetonide implant is wrapped around a polyimide device to restrict exposure and regulate drug release, which occurs at both ends of the device in the high-dose insert. The shell, which is inert, remains in the eye after therapy is complete. Two doses (0.5μg/day and 0.2μg/day) are being evaluated in the treatment of DME in a phase 3, sham-controlled study of 956 patients known as the FAME study. Results are anticipated in the fourth quarter of this year. A phase 2 pharmacokinetics study involving 20 low-dose and 17 high-dose patients, known as FAMOUS, is ongoing. Interim one-year data are being presented this month at ARVO.

Iluvien's later development relative to other options has allowed researchers to take advantage of some of the lessons learned in the Retisert trials. Two lessons in particular are as follows: (1) the point of release of the implant affects rate of adverse effects, and (2) low doses of steroid can be effective as long as zero-order kinetics are established. On the first point, researchers noted that anterior placement may expose the trabecular meshwork and lens to potential adverse effects (cataract and elevated IOP), while posterior release may mitigate those effects. Regarding the second point, a preliminary report on the FAMOUS data indicated that there were no IOP adverse effects in the low-dose group but 23.5% of high-dose patients had IOPs of 30mm Hg or more. Visual acuity improvement of 15 letters or more occurred in 27.3% of high-dose patients and 23.1% of low-dose patients. "If you can implant the low-dose device and not have the IOP concerns, perhaps you can still have the efficacy you can get from the high dose," Dr. Dugel suggested.

The newest, and perhaps most unusual, implantable steroid is the I-Vation implant from Surmodics, Inc. (Figure 12). This novel device uses a coiled non-ferrous metallic scaffold coated with triamcinolone and a proprietary polymer seal called Bravo that enables drug delivery to be time-released rather than immediate. The helical design of the implant maximizes the amount of implant surface area in the eye but minimizes the invasiveness involved in the process of implantation.

Figure 12. The I-Vation implant.

The implant itself is easily inserted, Dr. Dugel noted, but does require surgery. A 25-gauge needle-stick is made through the pars plana and the screw-shaped device is passed through it.

A phase 1 prospective, randomized, double-masked, multicenter study followed 31 diabetic macular edema patients for 3 years to evaluate the safety and tolerability of a slow-release and a fast-release implant, each comprised of 925 μg triamcinolone. As in the other studies, these patients were difficult to treat. In nearly all, laser photocoagulation had failed to achieve desired therapeutic effect; many had been unsuccessfully treated with intraocular steroids as well.

The efficacy profile shows a sustained clinical effect. At the 24-month follow-up visit, mean change in central retinal thickness was 126 μm. The early reduction in thickness, occurring at about the 2-month mark, was preserved throughout the study. The visual acuity data were more confounding, with improvement noted at 6 months, followed by a decline between months 9 and 18, and then improvement during months 18 to 24. This was concluded to be due to cataract formation during the study; acuity improved following IOL implantation. One serious adverse event (endophthalmitis) was reported in this study, but this was felt to be due to the surgical procedure rather than the implant itself. Progression of cataract development was common; however, mean IOP change was just 1.8 mm Hg.

What are the future roles for these implants? The Posurdex device may be useful in the shorter term to treat conditions such as pseudophakic macular edema, while Iluvien may be best suited for diseases requiring medium-term therapy and Retisert perhaps reserved for long-term use in uveitis. However, if the Iluvien device does show efficacy comparable to Retisert while maintaining its favorable safety profile. It may be an appropriate option for long-term therapy as well.

The I-vation device is intriguing perhaps more for its potential in different applications, as agents other than triamcinolone can be loaded onto it. At the moment, AMD and DME are the primary indications for implantable steroid devices of all stripes. There's a very good physiologic rationale for intraocular steroid release, and the prospects for this method of drug delivery, either as monotherapy or combination therapy, look bright. RP