Outcomes of systemic bevacizumab in radiation-induced optic neuropathy, case series

Optic neuropathy is a rare, delayed complication after radiation with no universally accepted treatment modality. We report the outcomes of 6 patients with radiation-induced optic neuropathy (RION) who were treated with systemic bevacizumab. This is a retrospective series of 6 cases of RION, treated with intravenous (IV) bevacizumab. “Improved” or “worse” visual outcomes were defined as a change in best corrected visual acuity of ≥ 3 Snellen lines. Otherwise, the visual outcome was noted as “stable”. In our series, RION was diagnosed 8 to 36 months after radiotherapy. IV bevacizumab was initiated as treatment within 6 weeks of the onset of visual symptoms in 3 cases and after 3 months in the other cases. Although no improvement in visual function was observed, stabilization of vision was noted in 4 of the 6 cases. In the other 2 cases, the level of vision declined from counting fingers to no light perception. In 2 cases, bevacizumab treatment was discontinued prior to completion of the planned course due to renal stone formation or worsening of renal disease. One patient developed ischemic stroke 4 months after bevacizumab completion. Systemic bevacizumab may stabilize vision in some patients with RION, though the limitations of our study do not allow us to draw this conclusion definitively. Therefore, the risks and potential benefits of using IV bevacizumab should be considered in each individual case.

and ocular complications, and lack of demonstrated benefit if initiated more than 72 h after vision loss [10,11].
Several case reports have shown promising results with using systemic bevacizumab in the treatment of RION [12,13]. However, another report showed progression of RION to no light perception (NLP) vision despite treatment with high-dose corticosteroids, hyperbaric oxygen, and intravenous (IV) bevacizumab [14]. To date, there is no universally accepted treatment modality for RION. In this paper, we report the outcomes of 6 cases in which patients with RION were treated with systemic IV bevacizumab.

Methods
This is a retrospective case series of all patients seen at a single academic tertiary referral center from August 2016 to August 2021 who were diagnosed with RION and treated with IV bevacizumab. In the following 6 cases, RION Background Radiation-induced optic neuropathy (RION) is an uncommon and late ischemic complication after radiation therapy in proximity to the optic nerve, and it typically leads to rapidly devastating vision loss [1,2]. Some treatment options that have been suggested in the literature for RION are corticosteroids, heparin, and pentoxifylline, but these have not shown consistent benefit [3][4][5]. Hyperbaric oxygen initially was suggested as a potentially beneficial treatment strategy [6][7][8][9]; however, its use is limited by cost, risk of systemic Ahmad Mohammed AlAmeer dr.ahmadalameer@gmail.com 1 was diagnosed by an expert neuro-ophthalmologist (ADH or ARC) based on clinical suspicion, magnetic resonance imaging (MRI) findings of optic nerve enhancement, and exclusion of other etiologies of optic neuropathy, such as recurrence of the primary tumor (Fig. 1). IV bevacizumab doses ranged from 5 to 10 mg/kg every 2-3 weeks for 3-4 cycles. The estimated maximum dose of radiation to the optic nerve was obtained via direct communication with each patient's radiation oncologist.
The best-corrected visual acuity (BCVA) on the Snellen chart was obtained with corrective lenses or a pinhole occluder. If BCVA was worse than 20/400, off chart visual acuity was recorded as 6/200, 5/200, 4/200, 3/200, 2/200, counting finger (CF), hand motion (HM), light perception, and NLP, with each step treated as representative of a single line of vision. Visual field (VF) testing was performed with an automated Humphrey Field Analyzer using a size III stimulus (or size V if poor vision), 24 − 2 test using SITA-Fast strategy. BCVA and VF results were recorded at the time of diagnosis and again at the last follow-up visit. Visual outcomes were noted as "improved" or "worse" if there were ≥ 3 lines of improvement or worsening of BCVA. Otherwise, the visual outcome was noted as "stable."

Results
A total of 6 cases (7 eyes) of RION treated with IV bevacizumab were identified, with patients ranging from 51 to 77 years old. RION was diagnosed between 8 and 36 months after radiotherapy. No improvement in visual function after IV bevacizumab was observed in any of the cases; however, stabilization of vision was noted in 4 of the 6 cases. In the other 2 cases, vision declined despite bevacizumab treatment. Table 1 summarizes the patients' demographics, tumor diagnosis, radiotherapy protocols, and visual outcomes of RION cases treated with systemic bevacizumab.

Case 1
A 75-year-old woman underwent whole-brain radiotherapy (30 Gy (Gy) in 10 fractions) along with oral osimertinib (a tyrosine kinase inhibitor) for treatment of metastatic non-small lung carcinoma. She presented 16 months after radiotherapy with blurred vision in the left eye beginning 4 weeks prior to her visit. Her BCVA in the left eye was 20/60, color vision on Ishihara plates was 7/13, and the VF showed a superior arcuate defect. The optic nerve was pale on examination, and pre-chiasmatic optic nerve enhancement was noted on the MRI (Fig. 1). The diagnosis of RION was made, and within 2 weeks, she began the first of 3 cycles of IV bevacizumab (7.5 mg/kg every 3 weeks). The optic nerve enhancement on MRI resolved after the first cycle of bevacizumab (Fig. 2), and her visual function remained stable for at least 3 months after initiating IV bevacizumab. At her last follow-up visit, the BCVA was 20/80, color vision on Ishihara plates was 7/11, and her VF showed generalized depression (though VF testing was unreliable due to positioning difficulties).    weeks earlier. Her vision in the left eye was 20/200, and her VF testing showed a dense cecocentral scotoma. Her optic nerve was pale on examination, and her MRI showed an equivocal area of left pre-chiasmatic optic nerve enhancement. She was diagnosed with RION, and 3 months later, started on 4 cycles of IV bevacizumab (7.5 mg/kg every 2 weeks). The patient reported no changes in the vision after completing 4 cycles of bevacizumab, and her vision in the left eye remained stable for 18 months.
However, 4 months after the last bevacizumab infusion, the vision in the right eye dropped to HM (from 20/50) following a seizure episode and several instances of loss of consciousness. RION, along with traumatic optic neuropathy, were considered in the differential diagnosis. MRI showed restricted diffusion in the right temporal lobe consistent with ischemic stroke, adding to the complexity of this case and preventing further bevacizumab infusions.

Case 5
A 73-year-old man received intensity-modulated proton therapy (47.6 Gy in 29 fractions) for treatment of a left spheno-orbital meningioma. His past medical history included hypertension, hyperlipidemia, obstructive sleep apnea, chronic kidney disease, and renal cell carcinoma. He was referred to the neuro-ophthalmology clinic for a 1-week history of sudden vision loss in the left eye, which occurred 11 months after radiation therapy. His BCVA in the left eye was CF, and VF showed superior altitudinal and inferior arcuate defects. His left optic nerve was pale temporally on examination, and MRI revealed increased DWI signal and focal enhancement of the pre-chiasmatic left optic nerve. He was diagnosed with RION and started on IV bevacizumab (5 mg/kg every 2 weeks) 1 week later. Treatment with IV bevacizumab was discontinued after the third cycle due to worsening of renal disease, and his vision subsequently declined to NLP 1 month later. The MRI continued to show increased diffusion restriction 1 month after cessation of the IV bevacizumab (Fig. 3), but the contrast enhancement was difficult to compare with the previous MRI due to differences in the imaging techniques.

Case 6
A 51-year-old man underwent whole brain radiotherapy (30 Gy in 10 fractions) for metastatic bladder cancer. One year after radiation, he noted blurred vision in the nasal corner of his right eye. Neuro-ophthalmic evaluation 1 week later showed BCVA of 20/20 in the right eye and 20/30 in the left eye, color vision (Ishihara color plates) was 9/11 with the right eye and 7/11 with the left. VF testing showed an inferior altitudinal defect in the right eye and a temporal

Case 2
A 65-year-old man received intensity modulated radiotherapy (60 Gy in 30 fractions) for left temporal lobe glioblastoma multiforme. His medical history included type 2 diabetes and hypertension. Twelve months after radiotherapy, he was referred to neuro-ophthalmology for evaluation of decreased central vision in the left eye over the prior 14 weeks. His BCVA in the left eye was CF, his VF showed a superior arcuate defect, and he had a pale optic nerve on examination. MRI demonstrated left pre-chiasmatic optic nerve enhancement. He was diagnosed with RION and began the first of 4 cycles of intravenous bevacizumab (10 mg/kg every 2 weeks) within 2 weeks of diagnosis. The optic nerve enhancement on MRI resolved after his second dose of bevacizumab, and his BCVA and VF testing were unchanged for at least 12 months of follow up.

Case 3
A 77-year-old man underwent proton beam therapy (74 cobalt Gray equivalent given twice daily in 61 fractions) for recurrent esthesioneuroblastoma of the olfactory bulb. Three years later, he reported a 1-month history of inferior visual field loss in the right eye. His BCVA in the right eye was 20/30, and his VF testing showed an inferior altitudinal defect with superior constriction. The right optic nerve was swollen inferiorly on examination, and his MRI showed posterior intra-orbital right optic nerve enhancement. He was diagnosed with RION, and treatment with IV bevacizumab (10 mg/kg every 2 weeks) was recommended; however, due to initial concerns from the patient's care team regarding potential for neoplastic disease recurrence, as well as limited literature data on bevacizumab, therapy was not started until 8 weeks after the diagnosis was made. The patient's vision in the right eye declined from 20/30 to CF in the interim (prior to starting bevacizumab). Additionally, bevacizumab treatment was ceased early after the third cycle due to renal stones. Unfortunately, the patient's vision further declined to NLP 9 months after IV bevacizumab was halted, and his right optic nerve continued to show MRI enhancement for at least 3 months after stopping IV bevacizumab.

Case 4
A 60-year-old woman with past medical history of right spheno-orbital meningioma, hypertension, hyperlipidemia, and obstructive sleep apnea underwent subtotal resection of the meningioma followed by intensity modulated radiotherapy (52 Gy in 29 fractions). Eight months after radiation therapy, she presented to the neuro-ophthalmology clinic with sudden vision loss in the left eye, which occurred 4 had a maximum exposure of more than 50 Gy to the optic nerve and/or chiasm, and the other 2 cases included patients who received whole brain radiotherapy and were exposed to approximately 30 Gy delivered in 10 fractions (3 Gy in each fraction).
The pathophysiology of RION is not yet fully understood, but several mechanisms have been proposed. One proposed mechanism suggests that direct radiation causes the release of free radicals that damage the vascular endothelium or neuroglial progenitor cells [15,19]. In addition to direct damage, it is also thought that somatic mutation in glial cells from radiation exposure may result in metabolically incompetent cells, leading to demyelination and subsequent neuronal degeneration [2]. Regardless of the mechanism, an ischemic component is a major factor in the pathogenesis of RION. Pathological studies have shown loss of pericyte and vascular smooth muscle cells in radiation brain necrosis [20]. Furthermore, vascular changes have been observed in radiation retinopathy such as capillary ischemia and reduced retinal blood flow, similar to those vascular changes seen in diabetic retinopathy [21,22]. This ischemic vascular injury leads to increased production of vascular endothelial growth factor (VEGF) and subsequent vascular hyperpermeability and tissue edema, which may ultimately exacerbate hypoxia and tissue necrosis [23].
Several reported models of radiation necrosis have also demonstrated elevated levels of VEGF [24]. Thus, medications that block VEGF, such as bevacizumab, are mechanistically reasonable for treating radiation-induced neural scotoma in the left eye. He also had moderate optic nerve swelling and intra-retinal hemorrhages in both eyes, and MRI showed diffusion restriction and focal intra-orbital optic nerve enhancement bilaterally. He was diagnosed with RION and started the first of 4 cycles of IV bevacizumab (5 mg/kg every 2 weeks) 2 weeks after diagnosis. Three months after IV bevacizumab was initiated, his BCVA remained stable at 20/25 in the right eye and 20/30 in the left eye. His VF also remained unchanged, and his color vision testing improved to 11/11 in both eyes. On fundoscopic examination, optic nerve edema improved, with resolution of the retinal hemorrhages. Optic nerve enhancement on the MRI resolved after bevacizumab treatment.

Discussion
RION is a rare, delayed complication of radiotherapy to the head, orbits, and paranasal sinuses, with incidence ranging from 0.53 to 9% [15]. RION classically presents with acute irreversible severe painless loss of vision in one or both eyes that can occur months to years after exposure to radiotherapy [7,16,17].
The development of RION is dose dependent, with higher risk if the total cumulative radiation dose exceeds 50 Gy or the fractional doses exceeds 2 Gy [15,16,18]. However, cases of RION have been reported in patients who received fractional doses between 1.8 and 2 Gy and total dose to the optic nerve < 50 Gy [3]. In our case series, 4 of the patients though the role of bevacizumab (if any) in this complication, which occurred in an area already affected by radiation vasculopathy, was unclear.
RION carries a poor visual prognosis with nearly 85% of cases presenting with a BCVA of worse than 20/200, and approximately half of the cases further declining to NLP vision [2,15]. In our series of 6 cases, IV bevacizumab was not shown to improve visual function, but stabilization of vision was achieved in 4 of the 6 cases, 2 of which had BCVA > 20/200. The limited number of cases in this retrospective review precludes analysis to determine the optimal IV bevacizumab dosage, duration of treatment and time period to initiate the treatment in RION. In patients with better BCVA, bevacizumab was administered within 6 weeks of the onset of visual symptoms, but it was delayed in the second, third, and fourth cases with a final BCVA < 20/200. This suggests that earlier initiation of bevacizumab may be beneficial in cases of RION. However, despite early bevacizumab administration in the fifth case, the patient progressed to NLP vision. This may be due to other factors, such as severity of disease at the initial presentation, the early discontinuation of bevacizumab, or the use of proton therapy. Evidence about factors contributing to radiation necrosis following proton therapy is sparse in the literature, and clinicians may need to have a higher index of suspicion for RION when evaluating patients who have undergone treatment with this newer modality of radiotherapy [29].
The variable natural course of the disease adds to difficulties in assessing the efficacy of available treatment options in RION. Larger prospective studies would be beneficial to better determine its efficacy, though these studies may not be feasible due to the rarity of RION. That being said, toxicity. Bevacizumab has indeed been proven in randomized controlled trials to be more effective than corticosteroids in treating radiation-induced central nervous system toxicity [25,26]; however, there are no prospective studies evaluating IV bevacizumab in treating RION. Farooq et al. reported a case of bilateral RION that was initially unresponsive to corticosteroids, but when IV bevacizumab was initiated 4 weeks after the onset of vision loss, the visual acuity improved from NLP to 20/25 in the right eye and from 20/100 to 20/40 in the left eye [13]. Dutta et al. later reported 2 other cases of RION in which patients treated with IV bevacizumab within 4 weeks of the vision loss recovered vision [12]. They also reported an additional case of a patient with gradual improvement of vision after starting systemic bevacizumab treatment within 7 weeks of the vision loss [12]. The dose of bevacizumab in these reports ranged from 5 to 10 mg/kg every 2-3 weeks for 3-6 cycles. A summary of the visual outcome of these case reports is shown in Table 2.
Previous studies have suggested that bevacizumab has good efficacy and is well-tolerated in the treatment of advanced cancers. Based on data from 2018, the cost of bevacizumab in the U.S. was $910.25 for 100 mg or $3,034.16 for 400 mg. Patients treated with IV bevacizumab receive 5-10 mg/kg per treatment, which would cost between $3,944.41 and $7,888.82 for a 100-kg patient per infusion [27]. Adverse effects of bevacizumab include hypertension, proteinuria, thrombocytopenia, hemorrhages, and thromboembolic events, so careful consideration is needed before initiating therapy [28]. In our series, bevacizumab therapy had to be stopped in 2 of the cases due to renal stone formation or worsening renal disease. An ischemic stroke developed in one patient 4 months after completion of bevacizumab, collaborative multi-center case-control studies are warranted to investigate the treatment effect of bevacizumab in RION.
In conclusion, IV bevacizumab may aid in stabilizing the vision of some patients with RION, though the limitations of our study do not allow us to draw this conclusion definitively. Therefore, risks and benefits of bevacizumab in each individual case should be carefully considered and discussed with patients before starting bevacizumab in the context of RION, and close monitoring after initiating therapy is required.
Author Contributions Ahmad Mohammad AlAmeer, Andrew Rising Carey, and Amanda Dean Henderson were responsible for the conception and design of this study. Data acquisition was performed by Ahmad Mohammed AlAmeer, who also wrote the first draft of the manuscript. Ahmad Mohammed AlAmeer, Andrew Rising Carey, and Amanda Dean Henderson contributed to the analysis and interpretation of the data. Revisions to the manuscript were performed by James Brian Davis, Ahmad Mohammed AlAmeer, Andrew Rising Carey, and Amanda Dean Henderson. All authors read and approved the final manuscript.
Funding The authors declare that no funds, grants, or other support were received during this manuscript's preparation.

Declarations
Ethics approval This is an observational study that was performed in line with the principles of the Declaration of Helsinki. The study was approved by the Johns Hopkins University School of Medicine Institutional Review Board (IRB00114028).