Reply

Abstract

To the Editor: Thank you for allowing us to respond to the Feltgen et al. letter to the editor in response to our published article titled “LIGHTSITE III: 13-Month Efficacy and Safety Evaluation of Multiwavelength Photobiomodulation in Nonexudative (Dry) Age-Related Macular Degeneration Using the LumiThera Valeda Light Delivery System.” We have reviewed the letter written regarding the LIGHTSITE III (LTIII) 13-month interim trial publication on the use of photobiomodulation (PBM) in subjects with intermediate dry age-related macular degeneration (AMD). Upon receipt of this letter from the journal, none of the authors had reached out to the LTIII principal investigators or LumiThera for clarification or discussion. Several points allude to discussion of end points that will be more appropriately addressed at the completion of the study and were not included in the limited 13-month interim analysis. Therefore, some of the comments by the authors are premature—the 13-month report is an interim analysis; final details on the completed 24-month study have not yet been published. This response will aim to address the comments, including those on the results of the 24-month data presented at conferences, without impacting future study publications in preparation. Along these lines, it is noteworthy to highlight the length limitations installed by the journal condensing the information on trial design, methodology, and study results that can be provided. Inquiries directed toward the corresponding author, or any author on the publication, provides a direct avenue for additional information circumventing this limitation. The LTIII study was an independent US medical device trial initiated following months of discussions with the FDA, which included formal review and approval of the final clinical protocol. The LTIII study employed a prospective, randomized, controlled trial design including double masking. Randomization of subjects to either the PBM or sham treatment group was conducted with a third-party vendor interactive web response system. Subjects enrolled into the study met restrictive inclusion/exclusion criterion that allowed a controlled population of subjects with anatomical signs of dry AMD and loss of vision to participate. The study enrolled patients that were diagnosed with a mean time since diagnosis of 5 years with dry AMD. Feltgen et al. comment on masking of the study and speculate whether the study was compromised by enrolling one eye in the PBM versus sham treatment group, making the patient aware of the two different light intensities. This is an incorrect statement because patients with either one or both eyes eligible were enrolled into the same treatment group to not compromise the masking. This is inherent to the study's fundamental design and clearly self-evident as the study employed a double-masked trial design consistent with the prior LIGHTSITE trials. The analysis by covariates such as by eye and other factors were conducted in the 24-month completed data set. Benefits of PBM on BCVA were observed in both subgroups, regardless of enrollment of one versus two eyes treated throughout the study. A further comment points to the significance of the best-corrected visual acuity (BCVA) effect and whether there was a placebo effect and any treatment effect of the sham treatment. An overall group mean of 5.4 letters was gained at 13 months following PBM treatment. In addition, subjects showed a mean letter improvement of > 5 letters, a mean gain of 9.7 letters, or the equivalent of two lines of vision. Natural progression in BCVA loss for intermediate dry AMD patients is estimated at 2.5 letters each year or a loss of one line of vision (5 letters) over 2 years. Furthermore, the adjusted annual probability of visual acuity (VA) progression for patients with mild-to-intermediate dry AMD (>20/63 or 60 letters) is 91% to 95%, highlighting a high estimate of VA loss progression among eyes in better-vision VA categories.1 In this study, the PBM group maintained and showed further improvement in BCVA benefit (mean of 5.9 letters) at 24 months.2 The significance of BCVA gain in this population with early signs of vision loss is impactful. The sham treatment delivers a lower dose of PBM and therefore serves as an active control. A 10- to 100-fold reduction in the 590 and 660 nm wavelengths, and removal of the nonvisible 850 nm NIR wavelength was performed. A lower dose of PBM still activates photoacceptor molecules and thus is assumed to have some effect on bioenergetic output of the cell. Some patients may benefit from the sham treatment but with a noted waning of this effect observed over time that paralleled disease progression in this group. The statistical robustness of PBM benefit in BCVA was shown over multiple end points with a mean 5.9 letter gain at 24 months compared with the 1.0 letter gain in the sham group (decreased from a 3.0 letter gain at 13 months). These findings will be discussed in more detail in the final study manuscripts but do not support any placebo effect. The study population included a majority of eyes with baseline BCVA >70 letters, which also brings into discussion the potential for a BCVA ceiling effect. Indeed, trials in AMD and diabetic macular edema with anti-vascular endothelial growth factor (VEGF) treatments demonstrate that a worse acuity at baseline predicts more gain in vision and that patients presenting with good acuity had any gains afforded by treatment impacted by a BCVA ceiling effect where the mean absolute BCVA achieved at 12 months, irrespective of trial and across patients, was approximately 70 letters.3,4 Analysis of baseline BCVA stratification in this study demonstrated a more robust response to PBM in eyes with a lower starting baseline (e.g., 61–70 vs. >70 letters) across 13-, 21-, and 24-month time points. These findings will be discussed in detail in the final manuscripts. This study used a 2:1 randomization ratio of PBM:sham. The patient population enrolled already shows anatomical and VA deficits due to the progressive and degenerative nature of AMD. Those randomized to the sham group were considered at risk for permanent vision loss over the course of a 24-month study. The results of the current trial confirm this assumption. After 24 months, more patients in the sham group progressed to geographic atrophy (GA) and showed a BCVA decline of >5 letters. Thus, the 2:1 randomization ratio was ethically instated to prevent needlessly putting more patients at risk for permanent vision loss. Although the overarching goals of clinical research are to evaluate safety and efficacy data, the patient risk of permanent vision loss should be considered in the trial design for lengthy trials conducted in progressive and degenerative ocular disease states. As always, we thank the authors for their diligence in review of the article details. To clarify any confusion, as stated, a total of 100 subjects and 148 eyes were enrolled into the study. A total of three eyes (from two subjects) were removed from the analysis in the modified intent-to-treat (mITT) group. The mITT group included 145 eyes and is correctly stated in the study diagram and results sections. The differing numbers for GA data are correct and represent different imaging modality analyses. A limited number of eyes that were enrolled into the study with noncentral GA used FAF imaging only as the primary diagnosis for enrollment and to determine lesion size and subsequent growth (Supplemental Table 2). The development of new-onset GA was further evaluated by the masked reading center (Duke Imaging Center) using all imaging modalities (i.e., OCT, FAF, Fundus photos) to confirm new occurrence and analyze all subjects for baseline categorization (Table 2). Regarding Table 2 where macular drusen volume numbers are identical for baseline and Month 13, we thank the authors for catching a publishing error during the proof generation of the article in which it appears a duplication of the 13-month line values occurred (Figure 1). This should be corrected to a baseline value of 0.941 (0.02) [0.22], which corresponds to the data presented in Figure 5A.Fig. 1.: Macular drusen volume following photobiomodulation treatment. A. Submitted table information (correct). B. Published table information.The study used an independent, leading global imaging reading center well experienced in clinical trials of AMD. The population enrolled had baseline drusen criteria set to at least three medium drusen, and thus, it is expected that drusen volume would be variable in the study population. Drusen was not the primary end point and only a surrogate for retinal health, and thus, it was not stratified for; however, a number and minimum size was required. The analysis further looks at change from baseline to minimize the difference in individual drusen baseline size levels. The imaging center was masked to treatment and qualified each eye before enrollment. Equipment was standardized across centers for all clinical and imaging outcome measures. The study was powered for the primary end point of BCVA. The imaging used standard segmentation parameters for the Heidelberg Spectralis as outlined in the imaging protocol and summarized in additional detail elsewhere.5 Feltgen et al. raise concern on the conversion of dry AMD eyes to nAMD. This is of great importance as increased nAMD conversion has been reported with recently approved complement inhibitors.6,7 Although established treatments for nAMD exist, this is an important safety end point that requires more discussion. This study independently evaluated each eye, and allowed subjects with nAMD in one eye to be enrolled if the fellow eye had dry AMD and met eligibility criterion. As the conversion to nAMD is only approximately 10% to 15% in the overall AMD population, groups were not balanced a priori—enrolled eyes were not stratified by fellow eye risk factors for neovascularization conversion. A review of the literature demonstrates that the presence of nAMD in one eye is a major risk factor for the development of disease in the fellow eye. Studies have reported conversion rates of 12% to 22% after 2 years and 22% to 38.7% after 5 years. The rate of progression is also affected by the visual status of the fellow eye, with eyes that have early or intermediate AMD and a diagnosis of CNV in the fellow eye progressing the fastest.8–10 Thus, when analyzing this subgroup of eyes that converted to nAMD during the study, it was revealed that of the 16 patients that had a high-risk study eye (i.e., study eyes whose fellow eye had nAMD), 12 were randomized into the PBM group and only four into the sham group. Thus, the percent of high-risk fellow eyes were 12.9% versus 7.3% in the PBM versus sham groups, respectively, indicating an imbalance between the groups. We also plotted the time to conversion versus the treatment protocol. Overall, conversion to nAMD occurred 78.9 (SE 10.0) days following the last PBM treatment, and thus, no correlation to treatment was observed. Normalization of the data to account for this imbalance provides similar conversion rates between both groups and will be discussed in further detail in the final publications. We reported the absolute percent for conversion to nAMD and only identified the overall nAMD conversion rate in the PBM group versus the sham group that may be influenced by the imbalance in the distribution of patients that had a fellow eye with nAMD. However, the absolute percent of conversion to nAMD was less in the PBM group than in the nonstudy (untreated) eye group and below the overall 10% to 15% reported in the AMD population. We find the comments on the article transparency of the trial curious from a group that has not once reached out to the investigators or sponsor to inquire on any further details regarding clinical trial design or published results. The literature cited in this letter is minimal and predominately falls to the 2021 Cochrane report, which aimed to review trials for the effectiveness and safety of PBM compared with standard care, no treatment, or sham treatment for patients with dry AMD. This report is outdated (current only to May 2020) and includes only two trials; one being the initial LIGHTSITE I pilot study using multiwavelength PBM delivered through Valeda, and the other being the ALight study using a mask to deliver 505 nm (green light). These studies differ immensely in regard to wavelengths used, cellular targets, and treatment protocols. The ALight study employed a light-emitting mask to emit a dim, green light (505 nm) worn each night for 12 months. The LIGHTSITE I trial treated patients using multiwavelength PBM through Valeda (yellow, 590 nm; red, 660 nm; Near infrared, 850 nm) using an intermittent protocol where patients received treatment 3× per week over 3 to 4 weeks (total of nine treatments at <5 minutes per treatment/per eye). In our opinion, these studies are not poolable. The majority of the literature supports PBM using wavelengths of 600 to 1,100 nm; wavelengths <600 nm are not as frequently researched and have rarely been used in ophthalmology. In addition, the use of 400 to 500 nm doses are surrounded by significant controversy related to the potential deleterious effects and safety hazard due to the ill-defined safety margin surrounding ultraviolet light. The Cochrane report is not robust and does not accurately reflect the current status of PBM utility in ophthalmology and more specifically AMD. The last few years has seen a surge of PBM trials in ophthalmology. Dozens of peer-reviewed clinical trials detailing PBM use in ocular conditions exist, 11 of which are published and conducted in AMD. An extensive review on PBM use in ophthalmology was recently published that includes both preclinical and clinical studies and details these reports, which may be of interest to the authors and this audience.11 LumiThera employees and consultants who have significantly contributed to the design and implementation of these studies are included as contributing authors on study articles. It is reflective of the sponsor both being transparent and giving credit to those who worked hard on these important clinical studies. We would request the journal to include the disclosures for the authors of this letter to the editor as none are currently reported, contradicting statements from prior work published from authors in this group. It is noteworthy that these disclosures may include companies with technologies and treatment options within the AMD field that may be considered in competition to Valeda. The authors comment on their underlying concern about the commercial cost of PBM use in Europe—very odd and inappropriate to see in a response to the published outcomes of a randomized controlled trial in the United States, where the product in not approved. LumiThera does not and cannot by law dictate the price a doctor charges for treating patients with PBM in Europe. Thus, we have not heard of $3,000 cost for a series of nine treatments being charged to patients in the Europe. Separately, the general cost for ocular injections used to treat later stage AMD on a monthly basis can be estimated around $2,000 USD or more per treatment (e.g., aflibercept [2 mg] estimated at $2,056 monthly cost or $26,718 projected cost per year).12 This includes the direct cost of procedures, medications, and health visit. Accordingly, the cost for an office visit and PBM treatment using Valeda would be considered a much more conservative health care cost and may be more practical as an earlier stage treatment for AMD patients. Any comment of the study not being sufficiently powered to draw valid conclusions about any effect on GA lesion growth or other anatomical end points is accurate and agreed upon. The results are included as additional data detailing findings from the study and provide foundational information for future studies to assess in a more robust format. Given the mechanistic underpinnings of PBM effect at the cellular level, these outcomes are of interest and are valuable to clinicians and researchers alike who may be interested in other evaluations. Finally, the growing volume of PBM data being generated by many investigators is not trivial; it is a dynamic and growing field of research in ophthalmology. The science has been shown to work both in vitro and in vivo with a multitude of animal models of ocular disease and injury and has repeatedly shown benefits in clinical populations with PBM under the right conditions, appropriate wavelengths, dosing, and safeguards on the devices. Healthy skepticism for new medical interventions is welcomed and appreciated; however, the timing, similarity, intensity of the criticisms, and lack of any engagement from the authors to better educate themselves on the core technology or trial design from the sponsor or colleagues involved in the trial are at a minimum curious and disappointing. We believe a repeated pattern of misleading questions appears to be an agenda-driven motive to slow the use and access to PBM treatment to which millions of patients could potentially benefit. Photobiomodulation stands on decades of preclinical and clinical research that demonstrates efficacy at the cellular and clinical level with a favorable safety profile. There is clear scientific evidence drawing experts to explore this modality with advances in research, providing more robust evaluations and thus better understanding of the technology and how to optimize its parameters for patient benefit.

Topics & Concepts