Phase 3 Failures Stunning But Not Uncommon

Why phase 2 promise is no harbinger of success.

BY PAUL A. YATES, MD, PHD

■ Editor’s note: Paul A. Yates, MD, PhD, is associate professor of ophthalmology and bioengineering and director of retina services at the University of Virginia; CEO of RetiVue LLC; and principal investigator in a phase IIb/III clinical trial evaluating Oracea for the treatment of GA.

In recent months, 2 retina drugs that showed great promise in phase 2 studies disappointed in large-scale pivotal phase 3 clinical trials. Both Ophthotech’s PDGF inhibitor Fovista for wet AMD in combination with anti-VEGF, and Genentech’s complement factor D inhibitor lampalizumab for GA, had stirred high hopes in the retina community. Did something go wrong?

Fovista and lampalizumab are just the most recent retina drugs to disappoint in phase 3 trials. About a decade ago, Opko’s revolutionary RNA-interference drug bevasiranib failed in a phase 3 trial for wet AMD, and Alcon’s anecortave acetate (Retaane) did the same.

History tells us that investigational ophthalmic drugs perform significantly above industry average at a 17% likelihood of going from phase 1 to FDA approval vs 9.6% for all other disease indications.¹ But it is in phases 1 and 2 where ophthalmic drugs gain most of their advantage, performing significantly above the mean, advancing from phase 1 to phase 2 84.8% of the time vs 63.2% for all indications and from phase 2 to phase 3 44.6% of the time vs 30.7% for all indications. In phase 3 studies, ophthalmic therapies actually revert to the mean, with phase 3 results leading to FDA approval 45.2% of the time vs 49.6% for all indications.

Why is the success rate so high in phases 1 and 2? In part, toxicities are likely lower for ophthalmic drugs, due to dosing and local application. Preclinical animal studies also likely weed out drug candidates that cause toxicity. Nevertheless, unanticipated ocular toxicities, such as intraocular inflammation or corneal deposits, have caused some trials to be suspended in phase 1. Also, intravitreal anti-VEGF injections have been demonstrated to affect plasma levels of VEGF, so these generalizations are not absolute.²

In critical phase 3 studies, ophthalmics revert to the mean. This can in part be attributed to a preponderance of novel, first-in-class therapeutics for ophthalmology. While this makes ophthalmic therapeutics an exciting field, it also means greater likelihood of failure. Additionally, phase 2 trials for any indication have a well-known set of confounding variables that complicate predicting successful phase 3 outcomes. These issues include the following:

1. Smaller sample sizes in phase 2 leading to nonuniformity in treated vs control populations even after attempts to assure uniformity.
2. Generalization, by which pivotal phase 3 trials necessarily include a broader population that may not have been fully represented by the phase 2 study group.
3. Proceeding into phase 3 when statistical significance is not achieved in phase 2, leading to high false positive rates, with perhaps fewer than half of positive phase 2 results representing true positives.³
4. Proceeding into phase 3 studies based on post-hoc subgroup analysis of the phase 2 data to parse out a treatment effect. Subgroup analysis to search for a significant P value (P<.05) is commonly referred to as P-hacking.
5. Use of secondary outcome measures to drive phase 3 trials even when primary outcome measures are not met.
6. Primary outcome measures are met but expected secondary outcomes measures are not in phase 2. This should serve as a red flag that the underlying hypothesis is not correct.

This all means that, in this context, phase 3 failures are to be expected — and even expected often. All pharmaceutical companies will certainly perform exacting risk–benefit analyses before spending hundreds of millions of dollars on pivotal trials for their new drug candidate. But there is financial incentive to move drug candidates forward where they deem doing so to be a reasonable risk.

Risk of phase 3 failure can be mitigated by increasing patient population in phase 2 and honing the optics through which positive phase 2 results are viewed before proceeding into phase 3. Thus far, development costs and risk–benefit analyses have supported the current, more limited phase 2 trial models, leading to greater phase 3 failures.

Although disappointing, failure should be reassuring. Drug development provides an almost impossibly large hill to climb. If drugs reach the top of that hill and are proven to work in phase 3, they work because the biology is correct, not because of how much money was spent on development or clinical trials.

References

1. Thomas DW, Burns J, Audette J, Carroll A, Dow-Hygelund C, Hay M. Clinical development success rates 2006-2015. Biotechnology Innovation Organization. Available at https://www.bio.org/sites/default/files/Clinical%20Development%20Success%20Rates%202006-2015%20-%20BIO,%20Biomedtracker,%20Amplion%202016.pdf .
2. Carneiro AM, Costa R, Falcao MS, et al. Vascular endothelial growth factor plasma levels before and after treatment of neovascular age-related macular degeneration with bevacizumab or ranibizumab. Acta Ophthalmol. 2012;90(1):e25-e30.
3. Liu PY, LeBlanc M, Desai M. False positive rates of randomized phase II designs. Control Clin Trials. 1999;20(4):343-352.

Allergan Builds Strong Retina Pipeline

A mix of late- and early-stage initiatives.

BY JERRY HELZNER, CONTRIBUTING EDITOR

■ Through a combination of internal research, partnerships, and acquisitions, Allergan has been expanding its presence in developing medical therapies for retinal disease, and today has one of the strongest retina-related pipelines in ophthalmology. In addition to its FDA-approved sustained-release implant, Ozurdex, the company now has late- and early-stage investigational initiatives for both the most prevalent retinal diseases and for less common inherited retinal dystrophies.

“Our core focus in retina is on unmet medical needs,” says Jag Dosanjh, Allergan senior vice president of Eye Care. “Currently, we have a phase 3 trial ongoing for a more durable therapy for wet AMD, a phase 2 trial for the unmet need of GA, and longer-range early phase programs in gene editing through our recent collaboration with Editas, and via optogenetics — providing light perception to damaged or naïve photoreceptor cells through our 2016 acquisition of RetroSense.”

The gene-editing collaboration with Editas provides Allergan with an entry into the revolutionary CRISPR-Cas9 technology that appears to have wide-ranging applications. The companies have selected Leber congenital amaurosis as their first therapeutic target.

Dosanjh estimates that the Allergan pipeline today encompasses the majority of the medical management of retinal diseases for the key patient types.

“We don’t have any ‘me-too’ therapies in the pipeline,” notes Dosanjh. “Through our open science model, we are committed to pursuing the best concepts, whether they are developed internally, through acquisition, or by collaborating with outside partners.”

One productive partnership is with Switzerland-based Molecular Partners, which developed the science for the designed ankyrin repeat protein (DARPin) abicipar pegol, a small molecule that binds to soluble isoforms of VEGF-A. Currently the DARPin is in 2 large-scale, phase 3 clinical trials for wet AMD (CEDAR and SEQUOIA).

A novel internally developed initiative is the study of brimonidine, which has neuroprotective qualities that might slow the progression of GA in a sustained-release format. After promising phase 1 and phase 2a studies, the concept is currently fully recruited in the phase 2b BEACON study utilizing a sustained-release delivery system similar to Ozurdex.

Early AMD Diagnosis From Blood Samples

Metabolites can be a biomarker of the disease.

■ Patients with any stage of AMD carry signs of the disease in their blood that may be found through special laboratory tests, according to a new study led by AMD researchers at Massachusetts Eye and Ear in Boston. The study, published in Ophthalmology, describes a new technique known as metabolomics.

“With metabolomics, we can identify blood profiles associated with AMD and its severity through laboratory testing,” said co-senior author Joan W. Miller, MD, chief of ophthalmology at Mass. Eye and Ear and Massachusetts General Hospital. “Because the signs and symptoms of early stage AMD are very subtle, with visual symptoms only becoming apparent at more advanced stages of the disease, identification of biomarkers in human blood plasma may allow us to better understand the early to intermediate stages of AMD so we may intervene sooner, and ultimately provide better care.”

The researchers’ approach revealed 87 metabolites, or small molecules in the blood, that were significantly different between subjects with AMD and those without. Furthermore, the team noted varying characteristics between the blood profiles of each stage of disease.

Six of the 7 most significant metabolites identified in the study were lipids. Previous research has suggested that lipids may be involved in the development of AMD, although the exact role of lipids in the disease process remains unclear. The results from this study support this suggestion, as well as indicate that metabolomics profiling may provide novel insights into relationship between lipids and AMD.

“Our work gives us a novel biomarker for early diagnosis, and it gives us clues to differentiate the progressors from nonprogressors,” said co-senior author Deeba Husain, MD, a retina specialist at Mass. Eye and Ear and associate professor of ophthalmology at Harvard Medical School.

Advisory Panel Backs Spark Gene Therapy

On track for landmark FDA approval.

BY JERRY HELZNER, CONTRIBUTING EDITOR

■ Spark Therapeutics cleared a major hurdle October 12 when an FDA advisory panel voted 16-0 to recommend FDA approval of the company’s gene therapy procedure for treating progressively blinding inherited retinal dystrophies (IRDs), initially Leber congenital amaurosis. If the FDA accepts the recommendation, which usually carries significant weight, the Spark procedure would be the first-ever gene therapy for a genetic disease to be approved in the United States.

The vote came several months after Spark’s submission of a biologics license application (BLA) with the FDA for voretigene neparvovec (Luxturna). Voretigene neparvovec is an investigational, single-treatment gene therapy using an adeno-associated virus (AAV) as a delivery platform for patients with vision loss from confirmed biallelic RPE65 mutation-associated IRD. The BLA included data from 3 clinical trials that enrolled participants with RPE65-mediated IRD, including the first randomized, controlled phase 3 trial for a gene therapy for a genetic disease.

Key to the advisory panel’s favorable recommendation was the pivotal phase 3 trial that met its primary endpoint of improvement of functional vision in a 20-patient intervention group compared to a 9-patient control group, as measured by the change in bilateral mobility testing between baseline and 1 year. All patients in the control group then opted for intervention and 8 of those individuals also responded positively to treatment. Patients who received the therapy were successful in navigating a marked path in dim light, something they could not do at baseline testing. There were no serious adverse events reported.

Subjects who received gene therapy outperformed control subjects in full-field light sensitivity threshold testing and the mobility test change score for the first injected eye. Visual acuity did not show statistically significant evidence of benefit.

“This is the first of hopefully many gene therapy approaches to treat both IRDs and other retinal diseases. We look forward to helping our patients with these previously untreatable conditions.” said Peter Kaiser, MD, Chaney Family Endowed Chair for Ophthalmology Research with the Cole Eye Institute at Cleveland Clinic.

Ophthotech Will Study Zimura in Wet AMD

Phase 2a trial in combination with Lucentis.

BY JERRY HELZNER, CONTRIBUTING EDITOR

■ As part of its refocused research effort following the failure of Fovista in wet AMD, Ophthotech Corporation has initiated an open-label phase 2a clinical trial of Zimura (avacincaptad pegol), the company’s complement factor C5 inhibitor, in patients with wet AMD. Zimura will be administered in combination with Lucentis (ranibizumab) in patients with wet AMD who have not been previously treated with anti-VEGF drugs. A range of Zimura dosing regimens will be assessed.

“We believe that supplementing anti-VEGF therapy with a complement inhibitor such as Zimura may have the potential to further enhance the efficacy of anti-VEGF monotherapy in wet AMD,” stated Kourous A. Rezaei, MD, senior vice president of Medical Strategy. “Our earlier phase 1/2a clinical trial assessing Zimura in combination with Lucentis, while small and uncontrolled, showed intriguing results.” Dr. Rezaei added that another recent study further supported that anti-VEGF therapy upregulates complement activation and therefore complement inhibition during anti-VEGF therapy may have therapeutic merit.

The company also announced it is on track to initiate 2 additional Zimura clinical trials before year end. The company’s strategy in orphan indications will be led by a randomized, controlled clinical trial to assess Zimura monotherapy in Stargardt disease. The other trial will be an open-label phase 2a clinical trial evaluating Zimura in combination with anti-VEGF therapy for idiopathic polypoidal choroidal vasculopathy, an age-related eye disease. Ophthotech is also planning a phase 2a clinical trial of Zimura monotherapy for intermediate/posterior noninfectious uveitis beginning next year.

Also, after the recent announcements of conflicting topline results of clinical trials from Genentech and Apellis Pharmaceuticals, assessing the role of 2 different types of complement inhibitors to treat GA, Ophthotech has decided to modify its ongoing phase 2/3 clinical trial of Zimura monotherapy in GA. The trial will be modified to accelerate the anticipated timeline to obtain topline data by reducing the number of patients, shortening the time point for attaining the primary efficacy endpoint and thereby reducing the cost to complete the study. The modified study design will incorporate patients already enrolled in the study.

Retina Scans May Help Identify Alzheimer Proteins

Imaging detects beta-amyloid deposits.

■ A 16-patient study led by researchers at Cedars-Sinai and NeuroVision Imaging offers a scientific basis for using noninvasive eye imaging to detect the pathologic hallmarks of Alzheimer disease. The experimental technology, developed by Cedars-Sinai and NeuroVision, scans the retina using techniques that can identify beta-amyloid protein deposits that replicate those in the brain.

The researchers report the feasibility of identifying beta-amyloid in the eye using autofluorescence imaging. They also provide detailed analyses and several new findings on Alzheimer pathology in the retina, results of additional research with donated eyes and brains of 37 deceased patients, 23 with confirmed Alzheimer disease and 14 controls.

Among key findings, the researchers report a 4.7-fold increase in retinal plaque burden in patients with Alzheimer’s, compared to controls, and observations regarding geometric distribution and layer location of amyloid pathology in the retina. With the imaging technology’s ability to detect autofluorescence signals related to retinal beta-amyloid, these findings may lead to a practical approach for large-scale identification of the at-risk population and monitoring of Alzheimer’s, the researchers say.

“This clinical trial is reinforced by an in-depth exploration of the accumulation of beta-amyloid in the retina of Alzheimer’s patients vs matched controls, and a comparison analysis between retina and brain pathologies. Findings from this study strongly suggest that retinal imaging can serve as a surrogate biomarker to investigate and monitor Alzheimer’s disease,” stated Maya Koronyo-Hamsoui, PhD, associate professor of Neurosurgery and Biomedical Sciences at Cedars-Sinai, and a cofounder, inventor and scientist at NeuroVision, in a news release. She is the senior leading author of an article in JCI Insight recently published online.