PROTON BEAM RADIOTHERAPY OF IRIS MELANOMA
BERTIL DAMATO, M.D., PH.D., F.R.C.OPHTH.,* ANDRZEJ KACPEREK, PH.D.,†MONA CHOPRA, M.D.,†
MARTIN A. SHEEN, M.A.,†IAN R. CAMPBELL, B.SC.,‡AND R. DOUGLAS ERRINGTON, F.R.C.R.†
*St. Paul’s Eye Unit, Royal Liverpool University Hospital, Liverpool, UK;†Douglas Cyclotron, Clatterbridge Centre for Oncology,
Bebington, Wirral, UK; and‡IC Statistical Services, Wirral, UK
Purpose: To report on outcomes after proton beam radiotherapy of iris melanoma.
Methods and Materials: Between 1993 and 2004, 88 patients with iris melanoma received proton beam
radiotherapy, with 53.1 Gy in 4 fractions.
Results: The patients had a mean age of 52 years and a median follow-up of 2.7 years. The tumors had a median
diameter of 4.3 mm, involving more than 2 clock hours of iris in 32% of patients and more than 2 hours of angle
in 27%. The ciliary body was involved in 20%. Cataract was present in 13 patients before treatment and
subsequently developed in another 18. Cataract had a 4-year rate of 63% and by Cox analysis was related to age
(p ? 0.05), initial visual loss (p < 0.0001), iris involvement (p < 0.0001), and tumor thickness (p < 0.0001).
Glaucoma was present before treatment in 13 patients and developed after treatment in another 3. Three eyes
were enucleated, all because of recurrence, which had an actuarial 4-year rate of 3.3% (95% CI 0–8.0%).
Conclusions: Proton beam radiotherapy of iris melanoma is well tolerated, the main problems being radiation-
cataract, which was treatable, and preexisting glaucoma, which in several patients was difficult to control.
© 2005 Elsevier Inc.
Iris melanoma, Proton beam, Outcomes, Therapy.
Iris melanomas account for approximately 3–5% of all
uveal melanomas (1). Without treatment, they can grow,
seed throughout the anterior chamber, invade the drainage
angle to cause secondary glaucoma, and spread extraocu-
larly. Approximately 5–10% of patients die of metastatic
disease within 10 years of treatment (2, 3).
The standard form of treatment for iris melanoma is
iridectomy, with iridocyclectomy being performed if the
tumor extends to angle or ciliary body (4). The surgical iris
defect tends to cause photophobia, and cyclectomy can
cause lens subluxation, hypotony, and phthisis. To prevent
local tumor recurrence, iridocyclectomy can be performed
with adjunctive plaque radiotherapy or a full-thickness cor-
neo-scleral excision followed by a tectonic graft (5). Pri-
mary brachytherapy has been advocated for patients with
extensive iris melanoma, and good results have been re-
ported (6, 7).
Proton beam radiotherapy has been used for the treatment
of ciliary body and choroidal melanomas for several de-
cades, with excellent rates of local tumor control (8, 9).
However, to our knowledge this modality has not previ-
ously been advocated for the treatment of iris melanomas. In
1994, it occurred to one of us (B.D.) that in view of the
limitations of surgical resection and plaque radiotherapy of
iris melanomas there was scope for treating iris melanomas
with proton beam radiotherapy.
The aim of the present study was to evaluate the early
results of proton beam radiotherapy of iris melanoma, mea-
suring outcomes in terms of visual acuity and ocular com-
PATIENTS AND METHODS
Patients were included in the study if (1) they were managed at
the Liverpool Ocular Oncology Centre and Clatterbridge Centre
for Oncology between 1994, when the first patient was treated, and
September 15, 2004, when the study was closed; (2) the tumor was
diagnosed as a melanoma, either clinically or histologically; (3) the
primary treatment consisted of proton beam radiotherapy; and (4)
at the time of our initial examination the tumor was considered to
originate in iris. Patients were excluded from the study if they had
previously received other treatment or if they were followed-up for
less than 6 months.
Pretreatment assessment included: (1) measurement of Snellen
visual acuity by an ophthalmic nurse, using spectacle or pinhole
Reprint requests to: Bertil Damato, Ph.D. F.R.C.Ophth., St.
Paul’s Eye Unit, Royal Liverpool University Hospital, Prescot St.,
Liverpool L7 8XP, UK. Tel: (?44) 151-706-3973; Fax: (?44)
151-706-5436; E-mail: Bertil@damato.co.uk
database support, Jane Kongerud, Martin Sheen, and Kathy Sztanko for
radiotherapy planning and administration, and the National Specialist
Commissioning Advisory Group (NSCAG) for funding of our service.
Received Dec 5, 2004, and in revised form Jan 10, 2005.
Accepted for publication Jan 11, 2005.
Int. J. Radiation Oncology Biol. Phys., Vol. 63, No. 1, pp. 109–115, 2005
Copyright © 2005 Elsevier Inc.
Printed in the USA. All rights reserved
0360-3016/05/$–see front matter
correction if necessary; (2) slit-lamp examination, with measure-
ment of longitudinal and transverse tumor diameters, performed
before mydriasis; (3) gonioscopy, measuring the circumferential
extent of increased pigmentation in the angle, in clock hours; (4)
binocular indirect ophthalmoscopy; (5) high-frequency B-scan
echography, performed with the 20-MHz probe (Innovative Imag-
ing Inc., Sacramento, CA), using a waterbath filled with 2%
hypromellose; and (6) systematic inquiry. Informed consent was
obtained from the patients, who were informed of the risks and
benefits of proton beam radiotherapy and other forms of treatment.
All patients were provided with an audiocassette tape recording of
their consultation to help them remember what they were told.
Ethics committee approval was not required for this study.
Proton treatment plans were prepared according to information
on a special form, which included: (1) a drawing of the shape and
extent of the tumor; (2) the target area comprising the tumor and
safety margins; (3) echographic measurements of distances from
cornea to back of lens, cornea to retina, retina to outer sclera, and
transverse ocular diameter (Fig. 1). Tantalum markers were not
necessary. The treatment was delivered with the pupil dilated to
reduce the tumor area. The eyelids were fully retracted, improving
accuracy of radiation targeting. The circumferential safety margins
were empirically adjusted to treat 1 clock hour (i.e., 3 mm) of iris
or angle beyond any visible tumor or suspicious pigmentation.
Wherever the angle was directly involved by the main body of the
tumor, the safety margin extended 4 mm posterior to limbus, in
case there was any ciliary body invasion. If the angle was involved
only by lateral extensions from the tumor, then a radial safety
margin of only 2 mm was used. The EYEPLAN eye therapy
program (V.1 to 1.6b, Douglas Cyclotron, Clatterbridge Centre for
Oncology) was used to contour the tumor and treatment areas, to
estimate radiation doses delivered to adjacent healthy structures
such as ciliary body and lens, and to provide data to a precision
milling machine for a brass collimator to be made, customized for
each tumor. The proton beam depth dose was measured before
treatment to confirm that the radiation beam had the required
accuracy in terms of dose and depth. The prescribed dose of
radiation was 53.1 Gy (58.4 Gy
delivered in 4 fractions over 4 consecutive days. Dosimetry checks
were performed before each treatment. A thin-walled MARKUS
60Co-equivalent), which was
Fig. 1. Infero-temporal iris melanoma in the left eye of a 52-year-old man, referred after growth of the lesion. (top left)
Multinodular tumor in December 1998, when the eye had a visual acuity of 20/13 and glaucoma was controlled with
Alphagan and Timolol drops; (top right) beam’s eye view, with line showing collimator edge; (bottom left) lateral view,
showing 90%, 50%, and 20% isodose curves; and (bottom right) appearance 5.5 years after treatment, when the eye was
comfortable with visual acuity of 20/20, an atrophic tumor, no cataract, minimal conjunctival telangiectasia, and normal
intraocular pressure after trabeculectomy in 2000.
110 I. J. Radiation Oncology ● Biology ● PhysicsVolume 63, Number 1, 2005
2543D ionization chamber (NE Technology Ltd, Berkshire, UK)
was used because of the shallow dose (i.e., ?6 mm) deposition
required for iris melanomas. The beamline and patient chair have
been described previously (10). The patient was seated with the
head constrained by an individually prepared ORFIT face mask
(Vertec Scientific Ltd, Reading, UK) and dental mouthpiece,
which restricted movement to ? 0.2 mm. The patient was asked to
gaze at a red light, which was positioned so as to minimize
radiation dose to normal tissues, such as eyelids. The required
targeting of the beam was confirmed using the radiation “light-
field,” projected onto the target area through the collimator.
Patients were reviewed approximately 6 months after the proton
beam radiotherapy, then 6 monthly for 6 years, and then annually.
At each follow-up visit, examinations included the following: (1)
measurement of the best corrected visual acuity with a Snellen
Chart, using spectacles or pinhole if necessary, with this measure-
ment performed by an ophthalmic nurse; (2) slit-lamp examina-
tion; (3) indirect ophthalmoscopy; and latterly (4) B-scan ocular
ultrasonography as described above. For patients living a great
distance from our center, arrangements were made for these as-
sessments to alternate between the referring hospital and our clinic
until the ocular status appeared stable, when patients were dis-
charged from our center for life-long monitoring at their own
hospital. Follow-up information on these patients was retrieved
from referring ophthalmologists by requesting photocopies of their
Outcome measures recorded were best corrected visual acuity;
presence of recurrent or persistent tumor growth; and ocular com-
plications, such as cataract and glaucoma. Follow-up times were
measured from the date of decision to perform proton beam
radiotherapy to the date of the event or the last known status.
Data regarding patient demographics, ocular and tumor charac-
teristics at presentation, treatment, outcomes, and symptoms were
collected prospectively. Each time a patient was examined, the
clinical findings were recorded in a customized database by the
first author (B.D.), using a computer terminal in the outpatient
clinic. Entries were subsequently checked by a full-time, ocular
oncology data manager, who also computerized treatment notes,
and data in correspondence from referring ophthalmologists and
the British National Cancer Registry.
All outcomes analyses were based on data computerized by
September 15, 2004. The follow-up period was measured to the
latest date at which ocular status was known. The results were
analyzed with SPSS software (Version 11.0, SPSS Inc., Chicago,
IL). Cox’s univariate proportional hazards model was used to
identify associations between baseline variables and cataract.
Kaplan-Meier estimates were used to draw survival curves for time
to this outcome. Statistical significance was taken to mean
p ? 0.05.
The sample comprised 88 patients (56% female, 44%
male). The age at treatment averaged 52.0 years (range,
21–76 years). The follow-up had a median of 2.7 years,
exceeding 1 year in 72 patients, 2 years in 54 patients,
and 4 years in 32 patients. The tumor was located in the
right eye in 51% patients and the left eye in 49% patients.
The visual acuity before treatment was 20/17 (40%),
20/20 (36%), 20/30 (15%), 20/40 (3%), 20/60 (1%),
20/80 (1%), 20/200 (1%), Counting Fingers (1%), and
Light Perception (1%). Concurrent ocular abnormalities
included cataract (15%), glaucoma (15%), uveitis (3%),
amblyopia (1%), herpetic keratitis (1%), band keratopa-
thy (1%), hyphema (1%), cellophane maculopathy (1%),
age-related macular degeneration (1%), and keratocon-
junctivitis sicca (1%). Important systemic disease in-
cluded diabetes mellitus (1%), Behçet’s syndrome (1%),
ovarian cancer (1%), thyroid cancer (1%), systemic hy-
pertension (1%), multiple sclerosis (1%), and polymyal-
gia rheumatica (1%).
The posterior tumor margin was in iris (80%), pars plicata
(14%), and pars plana (7%). The quadrant where the tumor
was centered was superior (1%), superonasal (5%), nasal
(3%), inferonasal (25%), inferior (13%), inferotemporal
(44%), temporal (8%), and superotemporal (1%). The tumor
involved 1 (26%), 2 (42%), 3 (26%), 4 (2%), 6 (2%), and 7
(1%) clock hours of iris. Angle spread was present in 59%
of patients and included 1 (15%), 2 (17%), 3 (11%), 4 (6%),
5 (1%), 6 (5%), 7 (2%), 8 (1%), and 12 (1%) clock hours.
The ciliary body involvement by tumor was 1 (6%), 2 (8%),
3 (7%), and 4 (1%) hours. Extraocular extension was
present in 3 patients. The tumor diameter ranged from 1.8
mm to 11.7 mm with a median of 4.3 mm. The tumor
thickness ranged from 0.5 mm to 5.0 mm, with a median of
1.4 mm. Trans-scleral biopsy before radiotherapy was per-
formed in 2 patients and showed spindle-cell melanoma in
In all patients, a dose of 90% or more of the radiation
delivered was received by a median of 0% of retina, 24% of
the ciliary body, 20% of the lens, and 37% of the lens
The latest recorded visual acuity was 20/17 (28%), 20/20
(30%), 20/30 (22%), 20/40 (7%), 20/60 (1%), 20/100 (2%),
20/200 (1%), Counting Fingers (2%), Hand Movements
(2%), Light Perception (1%). Three eyes were enucleated.
Figure 2 shows the last known visual acuity plotted against
Fig. 2. Visual acuity before proton beam radiotherapy of iris
melanoma and at the latest known status. Each circle represents
one case, and each line passing through this circle indicates an
additional case. A ? amblyopia; B ? bullous keratopathy; C ?
cataract; G ? glaucoma; H ? hyphema; R ? recurrence.
111Proton beam radiotherapy of iris melanoma ● B. DAMATO et al.
the pretreatment acuity, with labels for patients with loss of
20/40. The causes of visual loss were cataract, with the
patients waiting for surgery (4), glaucoma and cataract (2),
local tumor recurrence (3), bullous keratopathy, in a patient
who presented with band keratopathy caused by the tumor
(1), amblyopia (1), and hyphema, which was drained with
improvement of vision to 20/30 (1) (Table 1).
Cataract was first recorded after the radiotherapy in 18
patients. After excluding patients with pretreatment cata-
ract, the cumulative incidence of cataract at 4 years was
63% (95% confidence interval [CI], 48–78%). Cox analysis
showed that cataract was related to age at treatment (p ?
0.048; risk ratio, 1.04 per year; 95% CI, 1.00–1.08), initial
visual loss (p ? 0.0001; risk ratio, 3.10 per category (i.e.,
Hand Movements to Light Perception), 95% CI 1.65–5.80),
tumor thickness (p ? 0.0001; risk ratio, 2.77 per mm; 95%
CI, 1.59–4.83), iris involvement (p ? 0.0001; risk ratio,
1.91; 95% CI, 1.39–2.63), angle involvement (p ? 0.008;
risk ratio, 1.22; 95% CI, 1.05–1.41), and ciliary body in-
volvement (p ? 0.001; risk ratio, 1.80; 95% CI, 1.28–2.53).
Figure 3 shows the actuarial rates of cataract according to
tumor size. After cataract surgery, vision improved to 20/30
or better in all 4 patients with follow-up information and no
other causes of visual loss.
Glaucoma was present in 13 patients before treatment and
developed after the radiotherapy in another 3 patients. In 1
patient with secondary glaucoma before treatment, the in-
traocular pressure was controlled serendipitously when a
drainage bleb developed after biopsy, which was performed
before the radiotherapy. In the remaining patients the glau-
coma was treated with a variety of medications and surgical
procedures, with varying degrees of success. The glaucoma
contributed to visual loss in 2 patients.
Other complications were mild and transient conjuncti-
vitis (2 patients), corneal edema (1 patient, who had band
keratopathy before treatment), mild superficial keratitis (2
patients), recurrent hyphema (3 patients, 1 of whom re-
quired evacuation of the blood), and loss of lashes (1
patient). Several patients had temporary conjunctival hyper-
emia soon after treatment, delayed conjunctival telangiec-
tasia, pigment scatter on the iris surface or trabecular mesh-
work, and pupil distortion. No patients developed definite
rubeosis or neovascular glaucoma, although in 3 patients a
dilated iris vessel was recorded. Other than the patient with
Behçet’s syndrome, no patients developed frank uveitis.
The three enucleations were all performed because of
local tumor recurrence, which occurred only in these 3
patients. The cumulative incidence of local tumor recur-
rence and enucleation at 5 years was 6.3% (95% CI,
0–14%). Figure 4 shows ocular conservation according to
tumor size. In 2 patients, the recurrent tumor arose from
untreated diffuse, circumferential spread around the anterior
chamber angle. The third patient, a 68-year-old woman,
presented with an advanced tumor measuring 8.7 mm ? 8.6
mm ? 3.9 mm and received proton beam radiotherapy only
because she was extremely reluctant to undergo primary
enucleation. She developed recurrence 2 years after her
radiotherapy and died of metastasis 4 months later.
One patient (mentioned above) died of metastatic disease.
Two patients died of other causes, which were myocardial
infarction and head injury respectively.
This prospective, noncomparative, interventional study
found that proton beam radiotherapy of iris melanoma was
well tolerated, the main complication being cataract. No
Table 1. Patients with visual acuity worse than 20/40 at last known status
NumberAge (yr) Sex
up (yr)Latest vision Cause of visual lossIrisAngle
1 67M 20/200333 amblyopia,
2 64F20/803633.3 20/200bullous keratopathy
Abbreviations: CF ? Counting Fingers; HM ? Hand Motions; LP ? Light Perception; EN ? Enucleated.
112I. J. Radiation Oncology ● Biology ● PhysicsVolume 63, Number 1, 2005
patients developed iris neovascularization, despite irradia-
tion of iris and ciliary body. To our knowledge, there have
not been any published studies on proton beam radiotherapy
of iris melanoma. The main strengths of this study are the
large number of cases, the length of follow-up, and the
treatment at a single center.
We did not routinely confirm malignancy histologically
or observe patients for tumor growth before treatment so
that some of the tumors in our sample may have been
benign (11). The tumor inactivity after treatment may there-
fore not have been the result of the radiotherapy; however,
proton beam radiotherapy of ciliary body and choroidal
melanomas has been shown to have extremely high rates of
local tumor control, approaching 100%. There is no con-
sensus about the clinical features distinguishing iris nevus
from melanoma, with some authors diagnosing malignancy
if the tumor diameter exceeds 3 mm and others believing
that the tumor diameter and thickness should be at least
5 mm and 2 mm respectively. We have therefore catego-
rized our tumors according to these dimensions, when re-
porting outcomes. We do not believe, however, that tumor
size is necessarily the best predictor of its behavior, because
small tumors can behave more aggressively than large le-
sions. More ominous features, in our opinion, are diffuse
tumor growth and seeding, as shown by our patient with
recurrence after treatment of a 3.3-mm tumor. It was not the
purpose of this study to determine whether melanocytic iris
tumors should be observed or treated. Such a question can
be addressed only by a randomized, prospective study in-
volving several hundred patients and decades of follow-up.
In a series of 27 observed iris tumors, 70% subsequently
showed tumor growth, 11% developed raised intraocular
pressure, 1 patient had recurrent hyphema, and 1 patient was
lost to follow-up (4).
Local tumor control relies on the clinical ability to mea-
sure circumferential spread around the angle. This can be
difficult, as shown by the fact that 2 of the 3 recurrent
tumors arose from angle infiltration that was clinically un-
derestimated. There is scope for further studies aimed at
improving the differentiation of melanoma cells from mela-
nomacrophages, when assessing pigment in the angle or on
the iris surface.
Approximately 20% of the tumors in this study involved
both iris and ciliary body. Which of these tissues was
primarily involved was decided at the initial examination
according to where the tumor was more bulky or extensive;
however, it is possible that some of the tumors originated in
ciliary body. Excluding tumors involving ciliary body
would have avoided this potential error but would have
biased our results by omitting more aggressive iris melano-
One of the most remarkable findings of this study was the
minimal nature of the complications after treatment. The
most frequent complication in our study was cataract, which
also accounted for 4 of the 9 patients with visual loss caused
by the tumor or its treatment. Because this complication is
eminently treatable, with only rare complications and good
improvement of vision, the rates of loss of 20/40 and 20/200
visual acuity would have been appreciably lower if it had
been possible for our referring ophthalmologists to perform
cataract surgery without delay. Indeed, our data show good
improvement in vision after cataract surgery in patients with
no other causes of visual loss. Gragoudas and associates
have reported improved vision with cataract extraction after
proton beam radiotherapy of uveal melanoma, with no in-
crease in mortality (12). Although results are likely to be
similar in patients developing cataract after proton beam
radiotherapy for iris melanoma, there would be scope for
further studies once more such patients have undergone
cataract surgery. Cataract can also occur after iridocyclec-
tomy, in which case phacoemulsification is likely to be more
difficult because of instability of the lens, unless a tension
Fig. 3. Kaplan-Meier survival curves showing time to cataract
according to tumor diameter. A: Diameter ? 3.0 mm (n ? 16); B:
Tumor diameter 3.1–5.0 mm (n ? 45); C: Tumor diameter ? 5.0
mm and thickness ? 2.0 mm (n ? 13); D: Tumor diameter ? 5.0
mm and thickness ? 2.0 mm (n ? 14). Log rank: p ? 0.0001.
Cataract at Day 1 indicates preexisting disease.
Fig. 4. Kaplan-Meier survival curves showing ocular conservation
according to tumor size (Subgroups A–D are as in Fig. 2).
113 Proton beam radiotherapy of iris melanoma ● B. DAMATO et al.