Yellow-orange elevated areas subfoveally (A) and along arcades (A and D). Elevated areas are better visualized on infrared imaging (B and E), which highlights small pockets of SRF temporally (B). Subretinal fluid is seen subfoveally (C) and along temporal arcades (C and F). Green lines (B and middle line in E) represent the location of the optical coherence tomographic images (C and F).
A, Pockets of subfoveal SRF are present under the interdigitation zone. B, Large pockets of SRF are seen along the arcade vessels. C, A subtle diffuse layer of SRF is present under the interdigitation zone with a localized pocket of SRF temporally. Green lines in the infrared images (left, middle line of B) represent the location of the optical coherence tomographic images (right).
Serial OCT of a single study participant acquired prior to dosing (A and B) and hourly thereafter (C-F). Subtle subretinal fluid (SRF) is seen at 1 hour, reaches its peak at 3 hours, and is resolved by 4 hours after the dose.
Cumulative probability of having SRF detected during a study visit over time in 2 subgroups.
eFigure. Normal Optical Coherence Tomography
Customize your JAMA Network experience by selecting one or more topics from the list below.
Weber ML, Liang MC, Flaherty KT, Heier JS. Subretinal Fluid Associated With MEK Inhibitor Use in the Treatment of Systemic Cancer. JAMA Ophthalmol. 2016;134(8):855–862. doi:10.1001/jamaophthalmol.2016.0090
The use of mitogen-activated protein kinase (MAPK) kinase (MEK) inhibitors has become more common in the treatment of systemic cancer. These agents have been associated with a central serous-like retinopathy in some patients. Recognition of such retinal findings and the relatively benign nature of these events is important to avoid unnecessary intervention, including the cessation of a potentially life-prolonging medication.
To evaluate the presence and characteristics of subretinal fluid (SRF) associated with the use of MEK inhibitors in the treatment of systemic cancer and to correlate the presence of SRF with visual acuity and symptoms over time.
Design, Setting, and Participants
Post hoc analysis was conducted of prospectively collected data from 51 patients with locally advanced or metastatic cancer undergoing treatment with the MEK inhibitor binimetinib in 1 of 4 clinical trials. All clinical trial participants underwent complete ophthalmic examination by retina specialists at a private practice in Boston, Massachusetts, and were monitored between February 29, 2012, and January 8, 2014. The examination included Snellen-measured visual acuity, dilated fundus examination, and spectral-domain optical coherence tomography at baseline, biweekly for 2 months, then monthly for the remainder of their trial participation. Post hoc design and data analysis were performed between December 1, 2013, and June 20, 2014.
Main Outcomes and Measures
Visual symptoms, visual acuity, fundus appearance, and the presence and characteristics of SRF noted on optical coherence tomography. The characteristics of angiograms performed at the discretion of the treating physician were reviewed.
Of the 51 participants, 18 (35%) were men; the mean (SD) age was 60 (13) years (range, 32-87 years). Forty-six (90%) study participants developed SRF during the study period, with 9 (20%) experiencing symptoms at any point. The mean (SD) central retinal thickness of 39 study participants who developed SRF at the first visit increased from 280 (26) µm at baseline to 316 (43) µm at the first visit after starting binimetinib treatment (paired t test, P < .001). On examination, SRF appeared as elevated, yellow-orange pockets in the fovea and/or along the arcades. Corresponding optical coherence tomographic imaging revealed SRF beneath the interdigitation zone. The fovea was affected in 37 of 46 (80%) individuals; the location of SRF accumulation varied. Visual symptoms were mild and mainly transient, occurring in 9 participants with SRF (20%; 95% CI, 10%-33%). Only 2 participants (4%) were found to have SRF at the last study visit after discontinuation of treatment with binimetinib. Both had Snellen-measured visual acuity of 20/25 or better.
Conclusions and Relevance
The presence of SRF was common in study participants undergoing treatment with the MEK inhibitor binimetinib. Visual symptoms were mild and mainly transient. The presence of SRF did not lead to permanent ocular sequelae. Cessation of life-extending treatment with MEK inhibitors is not indicated when SRF is present.
Recent advances in the treatment of melanoma, multiple myeloma, and advanced solid tumors, including colorectal, non–small cell lung, pancreatic, breast, and endometrial cancers, have included the use of mitogen-activated protein kinase (MAPK) kinase (MEK) inhibitors.1 MEK inhibitors target MEK1/2 in the MAPK/extracellular signal–regulated kinase (ERK) pathway that controls cellular proliferation, differentiation, and survival and is activated in many of the cancers listed above.1,2 When the MAPK pathway is inhibited with monotherapy or combination therapy, tumor burden has been shown3,4 to decrease and progression-free survival to increase. The lack of a complete response, however, is believed to be influenced by both incomplete inhibition of the MAPK pathway and cross talk between the MAPK pathway and the phosphatidylinositol 3′-kinase/protein kinase B/mammalian target of rapamycin (PI3K/AKT/mTOR) pathway whose activation has been shown2 to suppress apoptosis and cause cancer. Combination therapies that target both pathways are under investigation.
As the use of MEK inhibitors has increased, physicians and patients are faced with numerous adverse effects, with ocular adverse events among the most common of these. Phase 1 and 2 clinical trials3-5 have noted the presence of subretinal fluid (SRF), described as a serous retinal detachment or central serous-like retinopathy, in 2% to 65% of study participants. This effect has been hypothesized to be secondary to acute toxic effects on the retinal pigment epithelium (RPE) and resultant dysfunction by inhibition of the MAPK pathway that lies downstream of the fibroblast growth factor receptor, which binds a neurotrophic factor essential for maintenance, repair, and survival of the RPE.6 Other evidence7,8 suggests that MEK inhibition causes upregulation of aquaporin 1, a membrane water channel essential in the permeability of the RPE, leading to increased permeability across the RPE and, thus, accumulation of SRF. A recent case series9 suggested the role of antiretinal and/or anti-RPE antibodies as the cause of RPE dysfunction.
MEK inhibitors have been reported10 to cause blurred vision, halos around lights, and colorful spots in the vision. A review of the oncology literature5,6 suggests that the development of SRF is typically asymptomatic, self-limited, and responds to dose reduction or cessation of the drug regimen. Some case reports and series in the ophthalmology literature have shown similar findings.9,11-13 Given that MEK inhibitors, alone or in combination with other targeted drug therapies, have been demonstrated4 to improve survival in study participants with metastatic cancer, further investigation of the natural course of SRF secondary to MEK inhibitor use is needed to prevent unnecessary interruption of therapy. To our knowledge, this is the largest and most comprehensive review to date reporting on the prevalence and patterns of SRF, symptoms associated with SRF, time course to development of SRF, and the response to changes in MEK dose or cessation of therapy.
Question Does the presence of subretinal fluid (SRF) associated with the use of MEK inhibitors correlate with visual acuity and symptoms?
Findings Post hoc analysis of prospectively collected data from 51 participants in 4 oncology clinical trials using the MEK inhibitor binimetinib showed that 46 individuals developed subretinal fluid; only 9 participants had symptoms, which were transient; the majority had no vision change.
Meaning This study suggests that SRF in association with MEK inhibitors is a common finding that does not require cessation of this potentially life-prolonging cancer treatment.
We conducted a post hoc analysis of prospectively collected data on 51 study participants in 4 phase 1b/2, open-label, dose escalation, and expansion clinical trials between February 29, 2012, and January 8, 2014.14-17 Design and post hoc analysis took place between December 1, 2013, and June 20, 2014. The clinical trials investigated combination therapy with binimetinib, a second-generation, selective MEK1/2 inhibitor, plus one of the following agents: encorafenib, a highly selective RAF kinase inhibitor; buparlisib, a pan-class I PI3K inhibitor; alpelisib, class I α–specific PI3K inhibitor; or dactolisib, a pan-class I PI3K and mTOR inhibitor. All agents were investigational, and the clinical trials were sponsored by Novartis Pharmaceuticals. The institutional review board at each participating center reviewed and approved all clinical trial protocols. All aspects of the post hoc analysis were designed and implemented in accordance with the Declaration of Helsinki18 and Good Clinical Practice guidelines. The participants provided written informed consent; they did not receive financial compensation. The retrospective review of data was conducted and analyzed by us in collaboration with the sponsor (Novartis Pharmaceuticals). However, no funding was received for the post hoc analysis.
Participants with protocol-specific mutations causing locally advanced or metastatic cancer, without a history or evidence of central serous chorioretinopathy or retinal vein occlusion at baseline ophthalmic examination, were included in the clinical trials. Individuals were excluded from the clinical trials if symptomatic brain metastases were present or if brain metastases had been treated with radiotherapy less than 4 weeks before study enrollment. Participants who had disease progression with prior treatment using MEK inhibitors in combination with RAF or PI3K inhibitors were also excluded. Adverse effects observed with the use of RAF and PI3K inhibitors include rash, elevated results of liver function tests, anorexia, nausea, vomiting, peripheral neuropathy, hyperesthesia, diarrhea, hyperglycemia, and thrombocytopenia. To our knowledge, no cases of SRF associated with RAF or PI3K inhibitors have been reported; therefore, any accumulation of SRF in this study was determined to be an adverse effect of the MEK inhibitor.
Comprehensive ophthalmic evaluation, including Snellen-measured visual acuity, slitlamp examination, dilated fundus examination, color vision testing (Farnsworth D15; Colblindor), visual fields examination (Humphrey 24-2; Carl Zeiss Meditec), fundus photography (Carl Zeiss Meditec), and spectral-domain optical coherence tomography (OCT; Heidelberg Spectralis) was completed at baseline prior to study enrollment, every 2 weeks for 2 months after starting the study medication, and monthly thereafter for the remainder of study participation. Fluorescein angiography was performed at the discretion of the examining ophthalmologist.
All study participants were enrolled in 1 of the 4 previously mentioned clinical trials14-17at Massachusetts General Hospital or the Dana-Farber Cancer Institute, and all ophthalmic examinations were performed by a retina specialist at Ophthalmic Consultants of Boston. A complete review of the electronic medical records and associated ophthalmic imaging obtained between February 29, 2012, and January 8, 2014, was performed. Drug dose and dates the study drug was withheld or ceased, the dose was modified, and treatment was restarted were obtained from the oncology records. Dose modifications after clinical trial enrollment were performed based on the severity of adverse effects of binimetinib (ie, rash, peripheral edema, blurred vision, central serous-like retinopathy, nausea, diarrhea, vomiting, fatigue, anorexia, stomatitis, anemia, and elevated levels of creatine kinase). For statistical analysis, the date of the first study visit was the date of the first visit during the clinical trial when the participant was receiving binimetinib (ie, the drug was not being withheld). The dose of binimetinib was between 30 and 45 mg twice daily in all participants. Ophthalmology records were reviewed to determine visual acuity and associated visual symptoms. Ancillary ophthalmic testing, specifically OCT, was used to evaluate the presence and location of SRF.
We performed statistical analysis using Microsoft Excel, version 2010 (Microsoft Corp) and R (https://www.R-project.org/). Statistical significance was calculated using the paired t test. The modified Wald method (GraphPad Software Corp) was used to calculate 95% CIs for proportions.19
The medical records of 52 clinical trial participants were reviewed; all participants met the inclusion criteria of the clinical trial and included those diagnosed with locally advanced or metastatic cutaneous melanoma, colorectal, non–small cell lung, pancreatic, or breast cancer, among others (Table). One individual was excluded from post hoc data analysis because the only study visit after the baseline evaluation occurred after treatment with the medication was stopped. Participants were monitored for a median of 56 days (range, 6-364 days). No central or branch retinal vein occlusion or episode of uveitis occurred during this study period.
Dilated fundus examination on many study participants receiving binimetinib revealed discrete, yellow-orange, elevated areas subfoveally and/or along the arcades. Optical coherence tomographic imaging confirmed these areas to coincide with the presence of fluid under the interdigitation zone (Figure 1 and the eFigure in the Supplement). Optical coherence tomographic scans from all study visits were reviewed.
Forty-six of the 51 participants (90%) had SRF present during the study period and 40 individuals (78%) had SRF present on OCT at their first visit after starting binimetinib therapy (Table). The location of SRF was variable (Table and Figure 2). Subretinal fluid was present both subfoveally and along the arcades in 29 of the 46 study participants (63%). Eight participants (17%) had SRF alone and 8 others (17%) had SRF only along the arcades. Close evaluation of OCT scans also revealed the presence of a thin, diffuse layer of fluid under the interdigitation zone in 14 of the 46 individuals (30%), which was usually associated with larger localized pockets of SRF. This diffuse fluid was not evident on clinical examination. Subretinal fluid was present in a different location compared with the fellow eye in 6 study participants (12%); typically, SRF was present subfoveally or along the arcades in one eye and in both locations in the fellow eye. Only 2 participants (SRF along the arcades in one and fluid subfoveally in the other) had no SRF in the fellow eye. These individuals were considered to have SRF present for the statistical analyses. Five study participants (10%) were noted to have variability in the presence or absence of SRF from one visit to the next without a change or interruption in study medication. Fluorescein angiography was performed on 7 participants with SRF; there was no evidence of leakage or pooling.
The mean (SD) central retinal thickness (CRT) of 39 study participants who developed SRF at the first visit increased from 280 (26) µm at baseline to 316 (43) µm at the first visit after starting binimetinib therapy (paired t test, P < .001). The scans from one study participant with SRF at the first visit were excluded from CRT analysis owing to poor scan quality. The baseline CRT between study participants with SRF and those without SRF were statistically similar (280 vs 269 µm, P = .17). There was no significant difference in CRT between the baseline examination and first study visit in participants without SRF (269 vs 268 μm; P = .72).
Of the 46 study participants who developed SRF over the course of the study period, 9 individuals (20%; 95% CI, 10%-33%) were symptomatic, with 8 noting only transient symptoms. Three were receiving encorafenib; 4, buparlisib; 1, alpelisib; and 1, dactolisib in combination with binimetinib. Eight of these participants reported blurry vision or central distortion; the ninth individual noted a central halo or circle in her vision. The time to developing symptoms after a drug dose was reported to be 45 minutes to 2 hours. Two participants noted resolution of the symptoms within 2 to 3 hours after onset. Optical coherence tomography confirmed a subfoveal pocket of fluid in all participants who were symptomatic at that study visit. No individuals with delayed presentation of SRF (after the first visit while receiving study medication) had symptoms.
Serial OCT scans were taken of 1 study participant starting before the drug was administered and every hour after administration for 4 hours. A subtle diffuse layer of SRF was seen at 1 hour after the dose. The amount of SRF increased through 3 hours after the dose; by 4 hours after the dose, the SRF had resolved (Figure 3).
Visual acuity decreased by more than 1 line in 6 eyes of 5 study participants. Two of those study participants were symptomatic in both eyes: one had intermittent blurry vision and the other had intermittent circles in the central vision. In one participant, Snellen-measured visual acuity changed from 20/25 at baseline to 20/50 at the last study visit. The patient was not receiving the study medication at that time.
The cumulative probability of developing SRF over the course of the study was calculated. Two data sets—intention-to-treat and per-protocol subsets—were evaluated. The per-protocol group contained only participants who were receiving the study medication for an uninterrupted period, which culminated in either the last study visit or the presence of SRF. The intention-to-treat group included participants who underwent a single or multiple episodes of drug treatment being withheld and restarted before the observation of SRF. These 2 data sets resulted in very similar survival curves, with medians of 11 days (95% CI, 8-15 days) to SRF observation in the intention-to-treat group and 8 days (95% CI, 7-13 days) in the per-protocol group (Figure 4).
Twenty-one study participants (41%) had no SRF detected at their last visit, a median of 60 days (range, 19-280 days) after initiating treatment with the medication. Of those individuals, 10 (48%) had stopped the drug treatment more than 24 hours before their visit and 6 (29%) had discontinued the drug treatment within 24 hours of the last visit. Five (24%) study participants without SRF were still taking binimetinib at the last follow-up visit. Twenty-five (48%) study participants had SRF present on OCT at their last visit with a median of 56 days (range, 6-364 days) of follow-up. Of those patients, 14 (56%) were receiving binimetinib. Nine (36%) study participants stopped treatment with the medication within 24 hours prior to the last visit and 2 (0.8%) patients discontinued treatment with the drug more than 24 hours previously owing to systemic adverse effects. Snellen-measured visual acuity was 20/25 or better in both of these study participants at the last visit. Five (10%) individuals did not develop SRF at any point in the study (Table). Two patients (4%) stopped binimetinib treatment within 24 hours of the last visit. All 5 study participants had 1 or 2 follow-up visits with a median of 7 days (6-21 days) of follow-up.
There were 10 study participants who underwent dose reduction of binimetinib from 45 mg twice daily to 30 mg twice daily owing to systemic adverse effects during a median of 138 days (range, 27-280 days) of follow-up. Eight study participants had SRF present in study visits before the dose reduction. Nine participants had recurrence of SRF a median of 9 days (range, 0-72 days) after dose reduction. The remaining participant did not have SRF at the study visit 3 days after starting the reduced dose and withdrew from the study 5 days later.
Advances in the genetic and molecular understanding of cancer have provided a basis for developing new therapies that antagonize pathways contributing to cancer pathophysiology. Such treatments may decrease tumor burden and increase progression-free survival. As with any treatment, beneficial effects can often be accompanied by adverse events that are mediated by on-target effects in normal tissues. One specific class, the MEK inhibitors, has been shown to cause the accumulation of SRF, which has been termed central serous-like retinopathy in the oncology literature. Although its exact pathophysiology is unknown, central serous chorioretinopathy is typically seen in men between the third and sixth decade of life with an increased risk in those using glucocorticoids.20 In the acute or classic form of central serous chorioretinopathy, ocular examination reveals SRF in the macula, sometimes in conjunction with an RPE detachment. A focal leak is frequently present on fluorescein angiography.20 Although similar in appearance, the presentation and location of SRF accumulation associated with the use of MEK inhibitors is unique compared with that of central serous chorioretinopathy. Further study may reveal additional insight into the cause of SRF associated with both central serous chorioretinopathy and MEK inhibitors.
The most common symptoms reported in this study were similar to those associated with central serous chorioretinopathy: central distortion or a circle in the central vision. Visual symptoms were noted in 9 study participants, all of whom had subfoveal SRF. Complete resolution of symptoms in those with MEK inhibitor–associated SRF was seen by 1 month after patients began receiving the study medication, despite the continued presence of SRF in all but 2 participants. The single participant who had intermittent symptoms throughout the majority of his follow-up period (364 days) had a Snellen-measured visual acuity of 20/10 in both eyes at the final study visit.
Subretinal fluid associated with binimetinib use was shown to be present as soon as 1 hour after the drug was administered. The rapid onset of symptoms, within 45 minutes to 2 hours after the dose, correlates with the mean time of maximum plasma drug concentration, which is 1.2 to 1.5 hours. Although the time to resolution of the SRF was not measured among all of our participants, the symptoms were reported to resolve in 2 to 3 hours in 2 patients and SRF was observed to be resolved 4 hours after the dose in one participant with serial imaging. The absence of SRF at a particular visit may represent a difference in the number of hours between drug dosing and acquisition of the OCT compared with other visits.
If SRF accumulates within hours before rapidly disappearing, the survival analysis performed in this study may represent the detection, rather than the incidence, of SRF. Similarly, the cumulative probability of developing SRF during this study represents the likelihood of observing the presence of SRF over the course of the study instead of the actual onset of SRF. Further research will be needed to better describe the association between the time of drug administration and the presence of SRF.
Only 2 study participants had residual SRF after discontinuation of binimetinib therapy. The amount of residual SRF was small and located along the arcades, a location unlikely to affect central vision, and both patients were asymptomatic with a final Snellen-measured visual acuity of 20/25 or better. However, the possibility of permanent damage to the RPE causing the extended presence of SRF cannot be excluded.
This study has limitations, including selection bias, information bias secondary to incomplete data points, and inability to control times between drug dosing and acquisition of OCT but, to our knowledge, represents the largest series of MEK inhibitor–associated SRF. Although SRF may persist for an unknown time after treatment with the drug is discontinued in less than 5% of study participants, these individuals were asymptomatic and experienced no change in visual acuity. In addition, visual acuity remained within 1 line of baseline vision in more than 90% of the participants with the remaining patients retaining Snellen-measured visual acuity of 20/50 or better in the setting of transposition error, concurrent ocular disease unrelated to the retina, or severe systemic disease. No participants in these trials required discontinuation of the drug regimen owing to ocular adverse effects, thus avoiding cessation of a potentially life-prolonging treatment in the management of advanced cancer.
The prevalence of binimetinib-associated SRF was 90% in our study and is significantly more than would be anticipated according to reports in the current oncologic or ophthalmologic literature.5,9-13 Our analysis demonstrates that binimetinib-associated SRF is unlikely to affect vision long-term and that SRF may resolve without intervention. Although these data are encouraging, further study needs to be undertaken to examine the long-term effects of MEK inhibitors as this treatment becomes more widely available.
Corresponding Author: Jeffrey S. Heier, MD, Ophthalmic Consultants of Boston, 50 Staniford St, Ste 600, Boston, MA 02114 (firstname.lastname@example.org).
Submitted for Publication: September 24, 2015; final revision received January 9, 2016; accepted January 11, 2016.
Published Online: June 16, 2016. doi:10.1001/jamaophthalmol.2016.0090.
Author Contributions: Dr Weber had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.
Study concept and design: Weber, Liang.
Acquisition, analysis, or interpretation of data: All authors.
Drafting of the manuscript: Weber, Flaherty.
Critical revision of the manuscript for important intellectual content: All authors.
Statistical analysis: Weber, Liang.
Administrative, technical, or material support: Flaherty.
Study supervision: Liang, Heier.
Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest. Dr Heier is a paid consultant for Aerpio, Alcon/LPath, Allergan, Avalanche, Bayer, EyeGate, Foresight Biotherapeutics, Forsight Vision4, Genentech, Icon Therapeutics, Janssen R&D, Kala Pharmaceuticals, Kanghong, Kato Pharmaceuticals, Novartis Pharmaceuticals, Ohr Pharmaceuticals, QLT, Regeneron, RetroSense, Santen, Shire, Stealth Biotherapeutics, Thrombogenics, Vision Medicines, and Xcovery. He receives research grants from Acucela, Alcon/LPath, Allergan, Astellas, Corcept, Genentech, Kala Pharmaceuticals, Kato Pharmaceuticals, Novartis Pharmaceuticals, Ohr Pharmaceuticals, Ophthotech, QLT, Regeneron, Sanofi/Genzyme, Stealth Biotherapeutics, and Thrombogenics. He owns stock in and is on the Board of Directors for Ocular Therapeutix. Dr Flaherty has received an honorarium as a consultant for Novartis. No other disclosures were reported.
Disclaimer: The views expressed by Dr Weber do not reflect those of the US Army, Defense Health Agency, Department of Defense.
Additional Contributions: Daniel I. Brooks, PhD (Walter Reed National Military Medical Center), contributed to data analysis. No financial compensation was provided.
Create a personal account or sign in to: