Auditing outcomes after treatment of Scottish patients with uveal melanoma in Liverpool

P. K. WRIGHT* and B. E. DAMATO#
*Aberdeen Medical School, Foresterhill, Aberdeen, U.K. and #Ocular Oncology Service, Royal Liverpool University Hospital, Liverpool, U.K.

Uveal melanoma is the most common primary intraocular malignant tumour, although it is a rare condition. Between January 1993 and December 1997 a total of 120 patients were referred from Scotland to the Ocular Oncology Service at St. Paul's Eye Unit in Liverpool. Of these, 98 had uveal melanoma, which was treated with primary enucleation (24), or by conservative methods, consisting of plaque radiotherapy (30), proton beam radiotherapy (19), trans-scleral local resection (19), or endoresection (6). After conservative treatment, 92% of eyes were retained, with 65% of eyes maintaining the same level of vision as at presentation or better. The main complications were retinal detachment (8), death from metastatic disease (7) and local tumour recurrence (4). This study gives an overview of the treatment of uveal melanoma and demonstrates that most patients with uveal melanoma can be treated conservatively, with the large majority retaining the eye and good vision.

Uveal melanomas are rare, having an incidence of approximately six per million per year.¹ Their treatment, therefore, requires specialised skills and techniques that are not widely available.

The diagnosis of uveal melanoma is based mostly on binocular ophthalmoscopy and colour photography, supported by ancillary investigations such as biopsy.² Ultrasonography is most useful for treatment selection and planning. The treat-ment of uveal melanoma is selected according to size and location, as well as the patient’s own priorities.³ With regards to choroidal melanoma, for example, the first line of treat-ment in most centres is plaque radiotherapy, administered using either ruthenium⁴ or iodine isotopes.⁵ In a small number of centres, proton beam radiotherapy is the preferred treatment⁶, whereas in others this treatment is reserved for small posterior tumours that are not easily treated with a plaque. Trans-scleral local resection is a difficult operation requiring skilled hypotensive anaesthesia and is possible, therefore, in only a few centres, where it is reserved for large tumours not easily treatable with radiotherapy.⁷

Radiotherapy of juxtapapillary tumours tends to be followed by radiational optic neuropathy with severe visual loss, so some centres prefer transpupillary thermotherapy^8,9 or trans-retinal resection (endoresection).^10,11 Iris and ciliary body melanomas are usually treated by surgical excision, with plaque or proton beam radiotherapy being reserved for unresectable tumours. In some instances, proton beam radiotherapy is selected as the first line treatment for inferior iris melanomas as a means of avoiding the problems of an iris coloboma (Damato, unpublished data). Enucleation, which was once the standard form of treatment, is now performed only if conservation of the eye and vision is unsafe, or if the patient is not highly motivated.

The Ocular Oncology Service in Liverpool was established in January 1993, and is one of the few centres world-wide where all treatment modalities are available. By December 1997 it had received 1 764 patients with a variety of ocular tumours and pseudotumours. A total of 120 of these patients were referred from Scotland.

The aim of this study was to evaluate the care of patients with ocular tumours referred from Scotland to the Ocular Oncology Service in Liverpool.

The sample included all patients with uveal melanoma referred from Scotland to the Ocular Oncology Service between January 1993 and December 1997. Patients with non-melanomatous conditions are mentioned but are not included in the statistics.

Patients were usually referred by telephone or fax communication, usually backed up by a mailed letter. They were usually seen within two weeks, with treatment being administered one day after the initial consultation at the oncology service.

Ultrasonography was performed by the consultant ophthalmologist, using a variety of scanners (Sonocare, New York; I³, California). Fluorescein angiography was rarely undertaken. Systemic examinations were performed for anaesthetic reasons, with liver ultrasound being reserved for patients with large tumours, who were considered to have a high risk of metastatic disease. Chest radiography was performed only if the patient had respiratory symptoms, or if there was an increased probability of extraocular malignancy.

All the treatment options were discussed with the patient and any relatives present, with the help of a model eye, ultrasound images and fundus drawings. At the end of the consultation, an audiocassette tape-recording of the actual discussion was given to the patient to facilitate recall of all relevant information. All clinical data were documented in the case notes using special proformas and, after August 1994, the data were entered into a specially designed ocular oncology database at the time of the consultation, using a computer terminal in the clinic.

Plaque radiotherapy was administered with 15mm and 20mm ruthenium applicators delivering a minimum dose of 400 Gy to the sclera and a minimum of 100 Gy to tumour apex. Since 1996 dosimetry has been performed using Astrahan software, which enables three dimensional modelling of the eye.12 Proton beam radiotherapy involved the insertion of four tantalum markers under general anaesthesia, simulation approximately two weeks later and, after a further two weeks, the delivery of 53 Gy in four daily fractions.

The techniques of trans-scleral local resection have previously been described.7 Briefly, the tumour was localised by transillumination. A lamellar-scleral flap was prepared applying diathermy to any intervening vortex veins. The eye was decompressed by limited pars plana vitrectomy, performed without infusion through a single port. The tumour was resected, together with the deep scleral lamella and a frill of normal uvea, where possible without damage to the normal retina. Haemorrhage was controlled by systemic hypotensive anaesthesia lowering the blood pressure to about 40mm Hg. The scleral flap was closed with interrupted sutures, and the eye reformed with balanced salt solution. Since 1994, all patients have received adjuvant plaque radiotherapy, delivering 100 Gy to a depth of 2-3mm.

Transretinal local resection (endoresection) involved three port pars planar vitrectomy with posterior vitreous detachment and clearance of entry sites.10 The tumour was removed with a standard vitrector, either through a large retinotomy, or after raising a retinal flap. Endodiathermy was applied to the margins of the base of the coloboma and to the scleral bed. After flattening the retina by fluid-air exchange, endolaser was administered, creating retinal attachment and destroying any residual tumour. Air-silicone oil exchange was performed, and cryotherapy applied to the sclerotomies in case of tumour seeding. The oil was removed after about 12 weeks.

Enucleation was performed in the standard fashion using a silicone implant or, since 1994, a coral implant. A retrobulbar injection of bupivacaine and adrenaline was administered at the start of the operation to prevent haemorrhage, the oculo-cardiac reflex and post-operative pain. A temporary prosthesis was fitted at the end of the operation.

Follow-up examinations were alternated between the referring hospital and the oncology centre until the referring ophthalmologist felt confident about continuing follow-up alone.

All post-operative data of patients attending follow-up clinic at the oncology service were computerised by the ophthalmologist at the time of the clinic visit. Information on discharged patients was obtained from referring ophthalmologists and general practitioners. The data were analysed using SPSS (Statistical Product and Service Solutions) version 8.0 software. Tumours were categorised as ‘large’ if they had a thickness of more than 5mm and/or a basal diameter of more than 15mm. The posterior tumour margin was defined as ‘posterior’ or ‘anterior’ according to whether it extended to within two disc diameters of the optic disc margin or fovea. Circumferential spread was according the extent of ciliary body, iris, and angle involvement, whichever was greatest.

Between January 1993 and December 1997 a total of 120 patients were referred to the oncology service from Scotland. 98 of these patients were confirmed as having uveal melanoma, and 22 patients were found to have other conditions (see Table 1).

suspicious choroidal naevus	5
benign naevus	4
congenital hyperplasia of the pigment epithelium	1
haematoma	1
metastasis	2
age-related macular degeneration	2
choroidal detachment	1
granuloma	1
retinal detachment	1
subretinal haemorrhage	1
other	3

The 98 patients with melanoma comprised 51 males and 47 females with a mean age of 58 years (SD 13.95; range 22-86). The tumour involved the left eye in 45 cases and the right eye in 53 cases. The pre-operative visual acuity in the affected eye was 6/6-6/12 (53), 6/18-6/60 (30), counting fingers (CF) (11) and hand movements - no perception of light (HM-NPL) (4). The location of the tumours was choroidal (71), cyclochoroidal (20) and irido-ciliary (7).

Of the 71 patients with choroidal melanoma, 15 were treated by primary enucleation. The reasons for this included: patient’s choice (4); large tumour (>16mm ultrasound diameter) (3); optic disc involvement (5); extraocular tumour extension (1); subretinal haematoma (1); and hypotensive anaesthesia not possible (1). The remaining 56 patients underwent conservative treatment, which consisted of: plaque radiotherapy (25); trans-scleral local resection (8); proton beam radiotherapy (17); and endoresection (6). In these 56 eyes, the initial vision was 6/6-6/12 (35), 6/18-6/60 (16) and CF (5). The anterior tumour margin was post-equatorial in 22 patients and pre-equatorial in 34 patients. The posterior tumour margin extended within two disc diameters of the optic disc margin, or fovea, in 31 eyes, with 4 involving disc, 2 involving fovea and 1 involving both of these. In 55 out of the 56 patients, ultrasonographic data was available. The tumours had a mean diameter of 10.5mm (SD 2.59) and a mean thickness of 4.6mm (SD 2.36). The tumours were categorised according to size and location as small anterior (13), small posterior (26), large anterior (12), and large posterior (5). Table 2 shows the final visual results according to this categorisation. Of the 13 patients with small anterior tumours, 11 (85%) retained vision of 6/12 or better, with two patients having visual reduction to 6/36 and 6/60 as a result of cystoid macular oedema and radiation retinopathy respectively. Of the 26 patients with small posterior tumours, three had severe visual loss to HM caused by: 1) retinal detachment after endoresection; 2) neovascular glaucoma; and 3) maculopathy after proton beam radiotherapy of a tumour close to the macula. Another patient had enucleation due to an abandoned endoresection of a juxtapapillary tumour. Of the 17 patients with large tumours (posterior and anterior), one had visual loss from CF to HM following proton beam radiotherapy of a macular melanoma and two patients lost their eye as a result of retinal detachment after endoresection and recurrent tumour after trans-scleral local resection.

Tumour size and location	6/6-6/12	6/18-6/60	CF	HM-NLP	Enuc	Total
Small & Anterior	11 85%	2 15%	0 0%	0 0%	0 0%	13
Posterior	15 58%	4 15%	3 12%	3 12%	1 4%	26
Large	3 25%	5 42%	3 25%	0 0%	1 8%	12
Large & Posterior	0 0%	2 40%	1 20%	1 20%	1 20%	5

(CF = Counting Fingers, HM = Hand Movements, NLP = No Light Perception, Enuc = Enucleation)

The 20 patients with cyclochoroidal tumours were treated by primary enucleation (9), plaque radiotherapy (4), trans-scleral local resection (6) and proton beam radiotherapy (1). The reasons for primary enucleation included patient’s choice (5) and large tumour (ultrasound diameter >16mm) (4). In the 11 eyes treated conservatively, the initial vision was 6/6-6/12 (7), 6/18-6/60 (3), and CF (1). The anterior tumour margin was in pars plana (4), pars plicata (5), and angle (2). The posterior tumour margin extended to anterior choroid (1), within two disc diameters of the fovea (1), and optic disc (1). The tumours had a mean diameter of 12.0mm (SD 3.28) and a mean thickness of 6.2mm (SD 3.16). Of the two patients with three or more clock hours of ciliary body involvement, both of whom underwent trans-scleral local resection, one had a final vision of 6/18 due to maculopathy and cataract, and the other had enucleation because of phthisis after treatment of recurrent tumour (see Table 3). One of the patients with less than three clock hours of ciliary body involvement had severe visual loss to HM as a result of chronic iritis and secondary glaucoma.

Table 3: Final vision after conservative treatment of cyclochoroidal tumours

Ciliary body involvement	6/6-6/12	6/18-6/60	CF	HM-NLP	Enuc	Total
< 3 Clock Hours	4 44%	4 44%	0 0%	1 11%	0 0%	9
=/> 3 Clock Hours	0 0%	1 50%	0 0%	0 0%	1 50%	2

(CF = Counting Fingers, HM = Hand Movements, NLP = No Light Perception, Enuc = Enucleation)

The 7 patients with iridociliary melanoma were treated with plaque radiotherapy (1), local resection (5) and proton beam radiotherapy (1). The initial visual acuity in these eyes was 6/6-6/12 (1), 6/18-6/60 (4), and CF (2). The anterior tumour margin was in iris (3) and angle (4). The posterior tumour margin was in iris (4), pars plicata (1) and pars plana (2). Most of these patients were treated before high frequency ultra-sonography became available, thus measurements of tumour dimensions are not available. Of the three patients with circumferential tumour extension less than three clock hours, one had a final vision of 6/24 due to pre-existing cataract, and another other had only HM because of bullous keratopathy. Of the four patients with circumferential spread of three clock hours or more, one retained a pre-operative vision of CF after proton beam radiotherapy as a result of age-related macular degeneration, and two patients had enucleation performed because of tumour recurrence and glaucoma due to loss of nucleus in phacoemulsification (see Table 4).

Circumferential spread	6/6-6/12	6/18-6/60	CF	HM-NLP	Enuc	Total
< 3 Clock Hours	1 33.3%	1 33.3%	0 0%	1 33.3%	0 0%	3
=/> 3 Clock Hours	1 25%	0 0%	1 25%	0 0%	2 50%	4

(CF = Counting Fingers, HM = Hand Movements, NLP = No Light Perception, Enuc = Enucleation)

Table 5 shows the final vision according to the initial vision in the 74 patients treated conservatively. Seventy percent of patients presenting with vision of 6/12 or better retained this level of acuity. Patients with reduced vision at presentation had more scope for visual improvement, but were also more likely to lose their eye because of complications.

Four patients developed local tumour recurrence, all of whom lost the eye. Four patients had possible tumour recurrence, which proved to be sub-choroidal silicone oil in two patients, and choroidal pigmentation in another two patients, which was treated prophylactically by laser photocoagulation. Eight patients had rhegmatogenous retinal detachment which was treated in three patients, but did not lead to an improvement of visual acuity.

By the close of the study nine patients had died. The causes of death included metastatic uveal melanoma (7) and myocardial infarction (2). The patients who died of metastatic disease included four patients with choroidal melanoma and three patients with cyclochoroidal melanoma.

Table 5: Visual outcome in relation to initial visual acuity in 74 patients treated conservatively

Initial visual acuity	Final vision
	6/6-6/12	6/18-6/60	CF	HM-NLP	Enuc	Total
6/6-6/12	30 70%	8 19%	1 2%	3 7%	1 2%	43
6/18-6/60	4 17%	8 35%	5 22%	2 9%	4 17%	23
CF	1 12.5%	3 37.5%	2 25%	1 12.5%	1 12.5%	8

(CF = Counting Fingers, HM = Hand Movements, NLP = No Light Perception, Enuc = Enucleation)

This audit of the treatment of uveal melanoma includes a significant proportion of patients diagnosed in Scotland over a five year period and demonstrates what can be achieved in a specialised ocular oncology centre when there is a wide choice of treatments administered routinely.

The main findings of the study are that retention of the eye is nearly always possible if the tumour is less than 16mm in diameter, not more than 5mm thick, involves less than one third of ciliary body or angle, and if the posterior margin is more than 3mm from the optic disc. Retention of vision is most likely if the tumour is more than 3mm from the disc and fovea, and if the vision is good on presentation. The main complications were recurrent tumour and retinal detachment, which tended to occur when conservative treatment of difficult tumours was attempted. It is interesting that in 65% of patients the last recorded vision was the same as at presentation or better.

The main weaknesses of the study are the small number of patients and the short follow up. These limitations preclude the determination of predictive factors for visual outcome and survival by multivariate analysis. Several hundred patients need to be followed up for at least seven years for complete statistical studies to be possible. Such studies are already under way with a large cohort of patients from Scotland and elsewhere. Problems encountered with long-term studies of large numbers of patients are, firstly, the logistics of obtaining follow up information from referring hospitals, and, secondly, the inevitable fact that there is constant development of treat-ment methods with inclusion of obsolete techniques in the analyses. Neither of these two problems were significant in the present study.

Retention of the eye and vision do not give a complete impression of quality of life in these patients. A questionnaire is being developed specifically for patients with uveal melanoma to measure visual handicap, ocular discomfort and other patient centred outcomes.

It must be difficult for an ophthalmologist in Scotland to decide whether to treat a patient locally or to refer to a distant oncology centre. According to current guidelines on cancer care, patients with malignancy should be treated at centres where multidisciplinary teams manage large numbers of cases.¹³ Studies with breast cancer in Scotland have shown that results are better when treatment is performed at specialist centres.¹⁴ In addition, such centres are perhaps more likely to employ trained counsellors so as to address psychological problems, which tend to be common in patients suffering from cancer.

In conclusion, retention of the eye is possible if the tumour is less than 16mm in diameter, less than 6mm thick, more than 3mm from the disc and fovea and involving less than three clock hours of ciliary body, iris, or angle. The main complications were retinal detachment and recurrent tumour, which were most likely to develop when the tumour was more than 15mm in diameter or when it involved optic disc or three or more hours of clock hours of ciliary body.

Correspondence: Mr Bertil E Damato, Ocular Oncology Service, St Paul’s Eye Unit, Royal Liverpool University Hospital, Liverpool, L7 8XP, UK

�1999 The Royal College of Surgeons of Edinburgh, J.R.Coll.Surg.Edinb., 44; 4: 260-4

AUDIT

Auditing outcomes after treatment of Scottish patients with uveal melanoma in Liverpool