Article Text

Download PDFPDF
Trans-scleral local resection of toxic choroidal melanoma after proton beam radiotherapy
  1. Lazaros Konstantinidis1,
  2. Carl Groenewald1,
  3. Sarah E Coupland2,
  4. Bertil Damato3
  1. 1Vitreoretinal and Ocular Oncology Service, Royal Liverpool University Hospital, Liverpool, UK
  2. 2Department of Molecular and Clinical Cancer Medicine, University of Liverpool, Liverpool, UK
  3. 3Ocular Oncology Service, University of California, San Francisco, California, USA
  1. Correspondence to Dr Bertil Damato, Ocular Oncology Service, University of California, San Francisco, 10 Koret Way, K304, San Francisco, CA 94143-0730, USA; DamatoB{at}


Aim To report on trans-scleral local resection of choroidal melanoma for exudative retinal detachment and neovascular glaucoma (toxic tumour syndrome) after proton beam radiotherapy (PBR).

Methods A non-randomised, prospective study of secondary trans-scleral local resection of choroidal melanoma for exudative retinal detachment with or without neovascular glaucoma after PBR. The patients were treated at the Liverpool Ocular Oncology Centre between February 2000 and April 2008. The trans-scleral local resection was performed with a lamellar-scleral flap, using systemic hypotension to reduce haemorrhage.

Results 12 patients (six women, six men) with a mean age of 51 years (range 20–75) were included in this study. The tumour margins extended anterior to ora serrata in six patients. On ultrasonography, the largest basal tumour dimension averaged 12.4 mm (range 6.8–18.1) and the tumour height averaged 7.1 mm (range 4.2–10.7). The retinal detachment was total in seven patients. Neovascular glaucoma was present in four patients. The time between PBR and local resection had a mean of 17.4 months (range 1–84). The ophthalmic follow-up time after the local resection had a mean of 46.2 months (range 14–99). At the latest known status, the eye was conserved in 10 patients, with a flat retina in all these patients and visual acuity equal or better than 6/30 in four patients. The reasons for enucleation were: patient request for enucleation when rhegmatogenous retinal detachment complicated the resection (one patient) and phthisis (one patient).

Conclusions Exudative retinal detachment, rubeosis and neovascular glaucoma after PBR of a choroidal melanoma can resolve after trans-scleral local resection of the tumour. Our findings suggest that these complications are caused by the persistence of the irradiated tumour within the eye (‘toxic tumour syndrome’).

  • Choroid
  • Treatment Surgery
  • Neoplasia

Statistics from


Proton beam radiotherapy (PBR) is an effective treatment for uveal melanoma, achieving high rates of local tumour control. Proton beams are highly collimated and their destructive ionising radiation falls abruptly distal to the target tissue (ie, the ‘Bragg peak’).1 It is therefore possible to target the radiotherapy very precisely at the tumour while minimising collateral damage to uninvolved adjacent tissues.2–6 Many eyes develop exudative retinal detachment (RD), rubeosis iridis and neovascular glaucoma (NVG), particularly with large tumours.7 It has been suggested that these problems occur because of irradiation of ciliary body and iris.8 However, we have not seen such complications in any of our patients undergoing PBR for iris melanoma, some of whom had irradiation of the entire anterior segment.9 ,10 In 1989, the senior author (BD) had a diabetic patient who, after ruthenium plaque radiotherapy of a cilio-choroidal melanoma, developed severe exudation, which regressed completely after trans-scleral resection of the irradiated tumour. He hypothesised that it was the tumour itself that caused the exudation and neovascularisation and coined the term ‘toxic tumour syndrome’. This concept has been accepted by other authors who resect irradiated tumours as a means of preventing NVG.11 ,12

The aim of this study was to report a series of patients for toxic tumour syndrome after PBR of a choroidal melanoma who were treated by trans-scleral local resection.

Patients and methods

Patient selection

Patients were included in the study if: (1) they had undergone trans-scleral local resection of choroidal melanoma at the Liverpool Ocular Oncology Centre between February 2000 and April 2008 as treatment for severe exudative RD (>50% of the retina) after PBR. The Liverpool approach to PBR has been described previously.13 Two patients from abroad were excluded from this study because we were unable to collect any follow-up information.

Data collected

Baseline and follow-up examinations included: measurement of visual acuity (VA) using a Snellen or LogMAR chart, slit-lamp examination, binocular indirect ophthalmoscopy and B-scan echography.

Baseline data included: age at diagnosis, gender, VA, intraocular pressure, quadrant location of tumour epicentre, largest basal tumour dimension and thickness on ultrasonography, presence of iris neovascularisation and percentage of retina detached.

Patients were reviewed approximately 1 month after the trans-scleral tumour resection, then 6-monthly for 6 years and then annually. Outcome measures recorded included: best corrected visual acuity (BCVA); presence of exudative retinal detachment (ERD), rubeosis and NVG; and ocular complications, such as rhegmatogenous RD and radiation optic neuropathy (ie, optic disc haemorrhage, oedema and/or peripapillary exudates).

Follow-up was measured from the date of local resection to the date of the latest ophthalmic evaluation. Follow-up concerning occurrence of metastatic disease or death of the patients was longer, that is, approximately 5 years until the study closed in June 2013.

Surgical technique

Trans-scleral local resection was selected if the patient was reluctant to undergo enucleation and when other forms of treatment were considered unlikely to conserve the eye. All operations were performed by the last author (BD). Informed consent was obtained from the patients who were advised on the risks and benefits of trans-scleral local resection and other treatments.

Signed consent was obtained from all patients for the use of data, images and tissues for research, teaching and audit. The tenets of the Helsinki Declaration were followed. Institutional ethical committee approval for outcomes analysis was not required because this was a service evaluation.

Trans-scleral local resection was performed as described previously,14 with modifications. Briefly, a fornix-based conjunctival flap was prepared and any extraocular muscles overlying the tumour were disinserted. The tumour was localised by transpupillary and trans-scleral transillumination and its margins were marked on the sclera with a pen. A polygonal, lamellar scleral flap was dissected, hinged posteriorly. The eye was de-compressed by core vitrectomy, which was performed through a sclerotomy 4 mm posterior to limbus, without infusion, using a contact lens and the illumination from the operating microscope to visualise the posterior segment. The deep sclera was incised so as to create a stepped wound edge to facilitate closure. The systolic blood pressure was lowered to approximately 40–50 mm Hg to reduce haemorrhage. In addition, short posterior ciliary vessels in the quadrant of the tumour were closed with bipolar cautery near the optic nerve. A small opening was made in the choroid by stretching the uveal tissues apart with two pairs of notched micro-forceps, and this was done just posterior to the ora serrata to preserve the ciliary epithelium. The tumour was resected with blunt-tipped corneo-scleral scissors, without safety margins. If the apex of the tumour was adherent to retina, this was sliced off with a Bard-Parker knife and left in situ. The sclera was closed with interrupted, 8-O nylon sutures. During scleral suturing, the globe was indented immediately posterior to the scleral flap, with two sponge-cells, to prevent subretinal haematoma. Once the scleral flap was closed, balanced salt solution was injected into the vitreous cavity until the globe was reformed. The muscles were re-inserted and the conjunctiva was closed. Two modifications were required in special circumstances. In the case of a total RD pushing against the posterior lens surface, the retinal funnel was opened either: (1) by injecting visco-elastic with a blunt cannula passed through a second lamellar scleral flap, created over the pars plana, or, more recently, (2) by injecting balanced salt solution into the anterior chamber so that it flowed through the pupil into the posterior chamber. If, after excising the tumour, the vitreous cavity was almost completely collapsed, when viewed through the scleral opening, digital pressure was gently exerted on the cornea to bulge the retina slightly so that balanced salt solution could safely be injected into the vitreous cavity through the anterior sclerotomy without damaging the retina, using a blunt cannula.


During the period between 2000 and 2008, 12 patients (six women, six men) with a mean age of 51 years (range 20–75) had undergone a secondary trans-scleral local resection for exudative RD after PBR.

Systemic conditions included a previous history of breast carcinoma in one patient.

Concurrent ocular conditions included amblyopia of the fellow eye (two patients), previous central retinal vein occlusion of the fellow eye (two patients) and blind fellow eye with RD since birth in one case. Four patients had previously been treated unsuccessfully with transpapillary thermotherapy and/or antivascular endothelial growth factor (anti-VEGF) intravitreal injections for exudation from the irradiated tumour.

The tumour was located in the right eye in five patients and in the left eye in seven patients.

The tumour margins extended anterior to ora serrata in six patients and to within 3 mm of optic disc or fovea in three patients.

On ultrasonography, the largest basal tumour dimension averaged 12.4 mm (range 6.8–18.1) and the tumour height averaged 7.1 mm (range 4.2–10.7). The exudative RD was total in seven cases, with a funnel shape and the retina in contact with the lens in two of these patients. Four patients had NVG.

The time between PBR and local resection had a mean of 17.4 months (range 1–84).

In four cases, the tumour was partially removed while eight cases had complete tumour resection. Intraoperative complications included the creation of retinal breaks in three patients. The exudative detachment resolved in all cases (figure 1). Rubeosis disappeared in all four patients in whom it was present. Three patients developed rhegmatogenous RD postoperatively. In two of these three cases, RD surgery was successfully performed. In the third case, the patient declined retinal surgery and underwent enucleation.

Figure 1

A patient was referred to our hospital presenting an exudative retinal detachment 8 months after proton beam radiotherapy at another hospital for a choroidal melanoma. (A) Visual acuity was counting fingers. Ophthalmoscopy showed a large, superior choroidal melanoma measuring on echography 18.1×18.1 mm in its basal dimensions with a thickness of 8.6 mm (B) associated with an extensive, bullous retinal detachment. Uneventful trans-scleral local resection was performed and at the last follow-up (C), the visual acuity was 6/30. The retina was flat with no residual tumour.

The mean ophthalmological follow-up time after the local resection was 46.2 months (range 14–99). At the last known status, the eye was conserved in 10 patients, with a flat retina in all and VA equal or better than 6/30 in four eyes.

Three patients developed optic neuropathy (table 1).

Table 1

Characteristics at baseline and at the last follow-up

The reasons for enucleation were: patient request for enucleation when rhegmatogenous RD complicated the resection (one patient) and phthisis (one patient). One case developed local tumour recurrence 3 years after PBR and 2.5 years after trans-scleral local resection and this tumour was treated successfully with Ru106 plaque radiotherapy.

Histological examination of the local resection specimens showed that most tumours were mixed, with varying proportions of epithelioid melanoma cells where these were viable. A number of degenerative alterations were present within the tumours (figure 2). Commonly seen were foci of tumour necrosis, which varied in size, and were often surrounded by dense macrophage infiltrates (figure 2). The blood vessels within the tumours showed thrombus formation, contributing to the ischaemia within the tumour, as well as thickened hyalinised walls following tissue repair. Other changes included melanoma cell vacuolisation and ‘balloonisation’, as well as oedema, fibrosis and sclerosis of the intratumoural connective tissue (figure 2).

Figure 2

(A) Low-power magnification of the local resection specimen of Case 6 demonstrating an extensive central area of necrosis (Nec) within the remaining tumour (Tu), with a dense infiltrate of melanomacrophages. Thrombosed blood vessels are indicated with arrows (H&E stain, objective ×2). (B) The same local resection specimen with the two demarcated areas of Nec with extensive macrophage infiltrate and the viable Tu. On the border between the two areas is a thrombosed vessel (arrow) (H&E stain, objective ×20). (C) Higher power magnification of the same tumour sample (Tu) with a freshly-thrombosed blood vessel (arrow) and the surrounding necrosis with some adjacent viable melanoma cells (H&E stain, objective ×40). A variety of other degenerative changes were seen within the local resection specimens, including resolving oedema with fibrosis, which progressed to sclerosis in some cases (D) (H&E stain, objective ×20); extensive balloonisation of the melanoma cells (E) (H&E stain, objective ×40); and dense lymphocytic infiltration (arrow) (F) (H&E stain, objective ×40).

Cytogenetic data were available for eight patients based on fluorescence in situ hybridisation (FISH) performed elsewhere.15 Monosomy 3 was identified in three of the tumours with five showing only disomy 3; the remaining four patients declined cytogenetics.

At study closure in June 2013, five patients had died, the cause of death being metastatic disease in two patients (table 1). In both these patients, histology showed the primary tumour to be composed predominantly of epithelioid cells with closed connective tissue loops. Interestingly, disomy 3 was reported in one of these cases using FISH (the other case was not evaluated cytogenetically), illustrating the need for molecular genetics tests of greater sensitivity, for example, multiplex ligation probe amplification16 and the incorporation of these results together with clinical and histomorphological data.1

selected case reports

Case 1

In 2002, a patient presented with a choroidal melanoma in the left eye associated with reduction in vision to 6/30 (Patient 6 in table 1). The choroidal tumour was located in the nasal, equatorial region. On ultrasonography, the tumour had a collar-stud shape, measuring 13.2×10.5 mm basally with a thickness of 9.3 mm. The right eye was amblyopic with VA of 6/38 mm. Because the patient was on warfarin therapy, with a previous history of stroke, we elected to administer PBR instead of local resection. Fifteen months after this treatment, however, the eye developed exudative RD and NVG with an intraocular pressure of 46 mm Hg after PBR. The tumour had a thickness of 7.9 mm. Because of the amblyopia in the fellow eye, we performed exoresection, with minimal hypotensive anaesthesia. The operation was completed without complications. The RD resolved within a day. The iris neovascularisation regressed within a month, and 24 months after the resection the intraocular pressure was normal without topical medications.17 The last known VA in October 2010 was 6/30 with the treated eye.

Case 2

In 2005, a patient was referred with an infero-nasal ciliochoroidal melanoma in her left eye (Patient 9 in table 1). The VA was hand movements with the affected eye and no light perception with the right eye, which had been blind with RD since birth. The tumour measured 9.5×9.2 mm with a thickness of 9.9 mm. Because this was the only ‘seeing eye’, the patient was treated with PBR. However, 4 weeks after completion of the radiotherapy, the patient presented with total funnel-shaped RD touching the lens. Trans-scleral local resection was performed because it was considered the only means of conserving the eye and restoring vision without delay. Silicone oil tamponade was used as a precaution in case a retinal tear had been created. After removal of the oil about 4 weeks later, the VA improved to 6/24. Six months later, however, the eye developed radiation-induced optic neuropathy and the VA gradually deteriorated to counting fingers at her last appointment in 2009. In the meantime, the patient developed diabetes mellitus. She died of metastases in August 2009.


The main finding of this study is that exudative RD after PBR of choroidal melanoma immediately resolved after trans-scleral resection of the irradiated tumour, with more gradual disappearance of any rubeosis and normalisation of any raised intraocular pressure. These results provide further evidence in support of the concept of ‘toxic tumour syndrome’ whereby the persistence of the irradiated tumour in the eye causes exudative RD, rubeosis iridis and NVG. The most likely explanation for such morbidity is that radiation-induced vasculopathy within the tumour itself causes: (1) ischaemia of the irradiated tumour tissue and (2) exudation from the damaged and incompetent tumour vasculature, resulting in macular oedema, hard exudates and exudative RD. Angiogenic factors arising from the ischaemic tumour and the ischaemic detached retina are likely to be inducing iris neovascularisation, resulting in NVG. Our histological examinations show thickening and hyalinisation of the vessel walls, oedema and necrosis; however, there would seem to be scope for further studies to compare irradiated with non-irradiated tumours.

Resolution of exudative RD and iris rubeosis was achieved in all our cases. Three eyes, developed rhegmatogenous RD; these patients underwent local resection in the early years of this study when it was considered necessary to remove the entire tumour. Later, it was found that any tumour that was adherent to the overlying retina could be left in situ without compromising the success of the operation. Two of the patients were treated successfully for this complication, the third patient opting for enucleation. None of these patients developed any complications as a result of the hypotensive anaesthesia. Some might consider such systemic hypotension to be excessively risky; however, experience with more than 700 primary trans-scleral local resections indicates that the risks of this procedure are low if the anaesthesia is administered by an experienced anaesthetist with adequate monitoring.18

Various treatments modalities have been used in the treatment of ‘toxic tumour syndrome’, and these include intraocular steroids and/or antiangiogenic agents, transpupillary thermotherapy, photodynamic therapy and endoresection.1 ,14 ,19

While anti-VEGF agents may have a beneficial effect in cases of NVG and/or uveitis related to ‘toxic tumour syndrome’, the significance on the effect of anti-VEGF and steroids on exudative RD is inconsistent in the literature. Newman et al20 reported complete resolution of exudative RD secondary to choroidal melanoma in two patients with systemic bevacizumab after plaque radiotherapy. Parrozzani et al evaluated the efficacy of prompt intravitreal triamcinolone acetonide injection versus intravitreal bevacizumab injection versus observation in the management of extensive exudative RD secondary to posterior uveal melanoma treated with brachytherapy. The exudative RD regressed in 22 (69%) of eyes treated with intravitreal triamcinolone, 11 (34%) treated with intravitreal bevacizumab and 9 (28%) untreated eyes.21 Vásquez et al22 reported resolution of ERD after intracameral bevacizumab in a case of NVG after brachytherapy for choroidal melanoma. Dunavoelgyi et al23 described a case of successful management of radiation-induced ERD and NVG with intravitreal ranibizumab.

It is not our intention to claim that trans-scleral local resection is the best treatment for the toxic tumour syndrome; this suggestion would require randomised trials to determine which treatment is superior. Should any such trials ever be considered, the results of the present study may help plan these investigations. In a small number of patients, we have successfully performed trans-retinal endoresection toxic tumour syndrome after PBR and we plan to report this case series in a future article. We deploy this procedure for tumours that are too small and posterior for the trans-scleral approach.

In conclusion, exudative RD, rubeosis and NVG after PBR of a choroidal melanoma can resolve after trans-scleral local resection of the tumour. These findings indicate that such complications are caused by the persistence of the irradiated tumour within the eye (‘toxic tumour syndrome’), at least in some patients.


View Abstract


  • Contributors LK reviewed the literature, participated in the analysis and interpretation of data and prepared the manuscript. All authors were involved in the study concept and design. CG, SEC and BD revised the manuscript for important intellectual content. BD conceived and supervised the study. All authors read and approved the final manuscript.

  • Competing interests None.

  • Patient consent Obtained.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.