Is it melanoma-associated retinopathy or drug toxicity? Bilateral cystoid macular edema posing a diagnostic and therapeutic dilemma

Abstract

Purpose

To report the clinical presentation, multimodal imaging, and management of a patient with metastatic melanoma who presented with cystoid macular edema (CME).

Observations

We report a case of a 71-year-old Caucasian male with metastatic melanoma who presented with bilateral cystoid macular edema after being on treatment with a programmed T cell death ligand 1 inhibitor, MPDL3280, for 1 year. Multimodal imaging techniques, including color fundus photographs, autofluorescence, spectral domain optical coherence tomography (OCT), fluorescein angiography (Spectralis, Heidelberg, Germany), and spectral-domain OCT angiography (Zeiss; California, USA) were performed to evaluate the etiology of his CME and to monitor his response to treatment. Clinical examination and multimodal imaging revealed 1 + chronic vitreous cells, an epiretinal membrane, and mild macular edema in both eyes. Fundus autofluorescence showed paravenous hypoautofluorescence in the right eye and scattered hypoautofluorescent spots in the left eye. Optical coherence tomography angiography (OCTA) revealed mild drop out of superficial vessels in the peri-foveal region bilaterally. These findings were concerning for melanoma-associated retinopathy, drug-related uveitis, or activation of a previous chronic autoimmune process. The patient was started on prednisone 30 mg oral daily and ketorolac tromethamine 0.5% 1 drop four times daily. He was then treated with bilateral sustained-release dexamethasone intravitreal implants (Ozurdex). He had complete resolution of CME, and was tapered off of oral steroids within 6 weeks.

Conclusions and Importance

Melanoma-associated retinopathy can be accompanied by CME, which presents a diagnostic and therapeutic dilemma in cases where a new drug has been recently initiated. By treating the condition locally, the ophthalmologist may be able to taper systemic immunosuppression more quickly.

1. Introduction

The use of monoclonal antibodies for the treatment of melanoma has resulted in higher survival rates but also substantial side effects and visual disturbances.^1,2 Here, we report a case of cystoid macular edema (CME) in a patient after initiation of treatment with the biological modifier MPDL3280A for which it was unclear if the CME was due to MAR or to the medication itself, resulting in a diagnostic and therapeutic dilemma.

2. Case report

A 71-year-old Caucasian male, with a history of metastatic melanoma that was in clinical remission with treatment under a clinical trial with a programmed T cell death ligand 1 (PD-L1) inhibitor MPDL3280A and vemurafenib, and with ocular history significant only for bilateral cataracts recently having undergone surgery, was referred with the complaint of increased blurry vision bilaterally. He had been on these medications for 1 year with gradual bilateral vision changes and floaters. Thirty days after cataract extraction with intraocular lens insertion in the right eye, the patient was found to have anterior chamber cells and flare in both eyes for which he was treated with oral prednisone 40mg daily, tapered over 1 week to 30mg four times daily, as well as topical prednisolone acetate 1% four times daily in the right eye and twice daily in the left eye, and cycloplegic drops. He had been off his anti-neoplastic agents for 6 days prior to referral due to the onset of these symptoms and because his clinical trial prevented use of antineoplastic agents while on prednisone.

On presentation best-corrected visual acuity (BCVA) was 20/125 in the right eye and 20/50 in the left eye. Intraocular pressures, visual fields, and motility were within normal limits bilaterally. Slit lamp exam of the right eye demonstrated a well-centered posterior chamber intraocular lens implant in the right eye, while the left eye had rare pigmented cells in the anterior chamber with inferior posterior synechiae and a mild nuclear sclerotic cataract. Fundus examination revealed mild epiretinal membranes (ERM) and peripheral pigmentary changes bilaterally (Fig. 1 A, E).

Fig. 1.

A. Fundus photograph of the right optic nerve showing peripapillary atrophy and pigmentary changes nasal to the disc. B. Fundus autofluorescence of the right eye showing peripapillary hypoautofluorescence as well as paravenous hypoautofluorescence. C. Optos ultrawidefield autofluorescence of the right eye showing multiple round areas of well-circumscribed hypoautofluorescence, most predominantly in the paravenous region and also in between vessels in the periphery. D. Fundus photograph of the left optic nerve showing peripapillary atrophy and round pigmentary changes near the optic disc. E. Fundus autofluorescence of the left eye showing peripapillary hypoautofluoresence and multiple round foci of hypoautofluoresence tracking along the vessels, with fewer areas than the right eye. F. Optos ultra-widefield autofluorescence of the left eye showing multiple round areas of well-circumscribed hypoautofluorescence, most predominantly paravenous and also in between vessels in the periphery, less prominent than the right eye.

Optical coherence tomography (OCT) confirmed the presence of the ERMs with mild cystic intraretinal fluid, an intact ellipsoid zone (EZ) and retinal pigment epithelium (RPE) in both eyes, distortion of the foveal contour in the left eye, and relative preservation of the foveal contour in the right eye (Fig. 2A–D). Autofluorescence showed paravenous hypoautofluorescence in the right eye (Fig. 1B and C) and scattered hypoautofluorescence in the left eye (Fig. 1E and F). Fluorescein angiography showed bilateral window defects without vasculitis and weak signal likely secondary to media opacity from the cataract in the left eye. Optical coherence tomography angiography (OCTA) revealed mild drop out of superficial retinal vessels perifoveally (Fig. 3A–F). Based on the clinical exam and multimodal imaging, the patient's differential diagnoses included panuveitis with bilateral CME concerning for melanoma-associated retinopathy, drug-related uveitis, pseudophakic macular edema, or activation of a previous chronic autoimmune process.

Fig. 2.

A. Infrared photo of the macula of the right eye, denoting the slice through which spectral domain optical coherence tomography (SD-OCT) imaging was performed. B. SD-OCT through the macula of the right eye, showing epiretinal membrane with relative preservation of the foveal contour and cystic intraretinal fluid. C. Infrared photo of the macula of the left eye, denoting the slice through which SD-OCT imaging was performed. D. SD-OCT through the macula of the left eye, showing epiretinal membrane with distortion and loss of the foveal contour and a solitary area of cystic intraretinal fluid subfoveally.

Fig. 3.

A. Spectral domain optical coherence tomography angiography (SD-OCTA) of the whole retinal thickness of the right eye, with possible perifoveal small vessel dropout. B. SD-OCTA though the superficial retinal layer of the right eye, from the internal limiting membrane to the superficial inner plexiform layer, showing a small amount of arteriolar and capillary dropout inferior and superior to the fovea. C. SD-OCTA through the deeper retinal layer of the right eye, from the deep inner plexiform layer to the superficial outer nuclear layer, demonstrating capillary dropout, mostly perifoveally. D. SD-OCTA of the whole retinal thickness of the left eye, with possible perifoveal and temporal small vessel dropout. E. SD-OCTA though the superficial retinal layer of the left eye, from the internal limiting membrane to the superficial inner plexiform layer, conveying a small amount of arteriolar and capillary dropout, again most prominent temporal to the fovea. F. SD-OCTA through the deeper retinal layer of the left eye, from the deep inner plexiform layer to the superficial outer nuclear layer, demonstrating more prominent capillary dropout, most prominent temporal to the fovea.

As the panuveitis improved, prednisone was tapered to 25 mg PO daily for 1 week and then 20 mg PO daily. He was also started on topical ketorolac tromethamine 0.5% 1 drop four times daily in the right eye. Fourteen days later BCVA improved to 20/30 in the right eye and 20/70 in the left eye, without change to the bilateral CME. To treat the CME and allow for local immunosuppression and to reduce time spent off of his chemotherapy regimen, an intravitreal dexamethasone implant (Ozurdex, Allergan Inc., Irvine, CA) (IVO) 0.7 mg was injected into the right eye and one week later into the left eye. One week after IVO, the BCVA had remained stable at 20/30 in the right eye and improved to 20/50 in the left eye with improvement in CME bilaterally. His prednisone was lowered to 5 mg a day and tapered off within 5 weeks, and the ketorolac was also tapered and stopped over a period of two months.

He underwent additional IVO ten weeks later bilaterally and again seven weeks after that in the right eye with near resolution of his CME within two weeks, and the decision was made with the patient's oncologist to restart his anti-neoplastic medications (Fig. 4A–D). After restarting the medications, he did not have a recurrence of CME, and blood testing revealed positive anti-retinal antibodies against 30-kDa (carbonic anhydrase II), 33-kDa (at very high titer), and 35-kDa (GAPDH) proteins. Prednisolone acetate 1% drops twice daily bilaterally were continued as a maintenance treatment. He then underwent cataract extraction with intraocular lens implantation in his left eye, and had a BCVA of 20/40 in the right eye and 20/30 in the left eye on last follow-up. Unfortunately, during this period of time, there was noted to be a relapse of his melanoma, for which he was treated with dabrafenib and trametinib. Systemic immunosuppression thus remained contraindicated, necessitating continued local intravitreal injections to control his uveitic process.

Fig. 4.

A. Infrared photo through the macula in the right eye, denoting the slice through which spectral domain optical coherence tomography (SD-OCT) imaging was performed. B. SD-OCT through the macula of the right eye after treatment with intravitreal dexamethasone implant. C. Infrared photo through the macula in the left eye, denoting the slice through which SD-OCT imaging was performed. D. SD-OCT through the macula of the left eye status after treatment with intravitreal dexamethasone implant.

3. Discussion

CME is defined as macular thickening as a result of a dysfunctional blood-retinal barrier that allows for fluid accumulation within the retina.³ CME is most commonly associated with cataract surgery, diabetes, retinal vein occlusion, and uveitis, but it can also be found as a reaction to biologic modifiers such as fingolimod,⁴ Of note, paravenous hypoautofluorescence, which was seen in our patient, has been found to be correlated with cancer associated retinopathy (CAR), MAR (often lumped together as paraneoplastic autoimmune retinopathy (pAIR)), and non-paraneoplastic autoimmune retinopathy (npAIR).⁵

In patients with multiple risk factors for CME, such as the patient described in this report, it can be difficult to determine the cause of CME. In these circumstances, a thorough history and exam are needed to elucidate its etiology. In stable patients, such as our patient, stopping possible instigators of the CME is a reasonable approach, and subsequent re-initiation without worsening of the CME can intimate at other secondary causes of the CME. Certain laboratory findings, such as the presence of anti-retinal antibodies, are also highly suggestive of CAR or MAR causing the CME. Particular imaging findings, though not highly specific, may also be more suggestive of MAR versus a drug side effect. For example, findings of paravenous hypoautofluorescence, may point more to npAIR or pAIR.⁶ In the case of our patient, CME did not recur after restarting his anti-neoplastic medications. Given his response to treatment and resolution of CME without worsening on re-initiation of treatment, the presence of anti-retinal antibodies, and the imaging findings in our patient it is clear that his CME was due to MAR, rather than a drug effect.

The treatment of CME is generally determined by the underlying etiology. Historically, systemic steroids were used for many causes of CME, but they have unwanted side effects and must be used with caution in certain patients, including patients with history of malignancy, in whom non-targeted systemic anti-inflammatory medications may be relatively contraindicated.^7,8 This provides a treatment dilemma since the benefit of vision improvement and preventing lasting ocular damage must be weighed against the risks of relapse of cancer.

The development of newer intraocular steroids, such as the intravitreal sustained-release dexamethasone implant, has helped address this dilemma, and local treatment has been found to cause relief of CME in some cases, particularly in CME caused by diabetes.⁹ To the best of our knowledge, there are very few cases suggesting that local treatment can successfully treat CME from CAR or MAR. One case demonstrated successful local treatment of bilateral CME in a patient with CAR from squamous cell carcinoma (SCC) of the lungs.¹⁰ After worsening visual acuity failure of response to systemic prednisone, mycophenolate mofetil, and four doses of IVIG, intravitreal triamcinolone was attempted in both eyes. BCVA and OCT findings subsequently improved. Moyer et al. also reported a similar case study in a patient with CME in the setting of CAR from SCC who was treated with subtenon's triamcinolone injection and diclofenac with improvements in visual acuity.¹¹ These studies, along with our patient who demonstrated significant improvement and stabilization of CME with local therapy, suggest that local therapy may have some advantages over systemic therapy, particularly for patients with cancer.

4. Conclusions

In the case of our patient, CME did not recur after restarting his anti-neoplastic medications. Given his response to treatment and resolution of CME without worsening on re-initiation of treatment, it is clear that his CME was due to MAR, rather than a drug effect. MAR and biologic modifiers to treat metastatic melanoma can both lead to CME, which poses a diagnostic and therapeutic dilemma. By treating the condition locally, the ophthalmologist may be able to quickly taper or avoid systemic immunosuppression altogether. In concert with the oncologist, the ophthalmologist can monitor the patient's response as the biologic modifiers are reinitiated, and thus better understand the etiology of CME.

Patient consent

No personal information or identifiable images of the patient were used in this report. Consent to publish this report was not obtained.

Acknowledgements and disclosures

Funding

Supported by an unrestricted grant from Research to Prevent Blindness, New York, NY 10022.

Conflict of interest

The following authors have no financial disclosures: SP, SM, NK, PS, AK, NR.

Authorship

All authors attest that they meet the current ICMJE criteria for Authorship.

Disclosures

Supported by an unrestricted departmental grant from Research to Prevent Blindness, New York, NY 10022. Dr. Rodger is a paid consultant for Allergan, Inc. All authors attest that they meet the current ICMJE criteria for Authorship.