Proton Beam Therapy for Iris Melanomas in

107 Patients
Juliette Thariat, MD, PhD,1,2 Ahmed Rahmi, MD,3 Julia Salleron, PhD,4 Carlo Mosci, MD,5 Benjamin Butet, MD,6
Celia Maschi, MD,6 Francesco Lanza, MD,5 Sara Lanteri, MD,6 Stephanie Baillif, MD, PhD,6 Joel Herault, PhD,2
Thibaud Mathis, MD,3 Jean Pierre Caujolle, MD6

Purpose: To report on the clinical characteristics and outcomes for patients with iris melanoma using
proton therapy.
Design: Retrospective study.
Participants: One hundred seven patients with iris melanoma from 3 regional ophthalmologic centers.
Methods: A retrospective study was conducted for iris melanoma patients from 3 regional ophthalmologic
centers referred to and treated at a single proton therapy facility between 1996 and 2015.
Main Outcome Measures: At each follow-up visit, examinations included measurement of best-corrected
VA, slit-lamp, examination, indirect ophthalmoscopy, and ultrasound biomicroscopy.
Results: With a median follow-up of 49.5 months, 5 of 107 patients experienced a local relapse within a
median of 36.3 months. The cumulative incidence of relapse was 7.5% at 5 years. All 5 patients showed
involvement of the iridocorneal angle (P ¼ 0.056). Diffuse iris melanoma showed a higher risk of relapse
(P ¼ 0.044). Four patients showed out-of-field relapse and 1 showed angular relapse. Three patients were
retreated with proton therapy, whereas 2 other patients, one with T1b disease and another with diffuse T3 dis-
ease, underwent secondary enucleation. None of the patients experienced metastases nor died of iris melanoma.
Vision improved in 59.4% of patients (n ¼ 60/101). However, cataracts occurred in 57.4% of the 54 patients
(n ¼ 31) without cataract or implant at diagnosis. Secondary glaucoma was reported in 7.6% of the patients
(n ¼ 8), uveitis in 4.7% (n ¼ 5), and hyphema in 3.7% (n ¼ 4). All but 5 cases of complications were mild, transient,
and not sight limiting after treatment. Five cases of glaucoma, including 1 with uveitis, were severe and
associated with visual deterioration.
Conclusions: Proton therapy showed efficacy and limited morbidity in iris melanomas. Ophthalmology 2018;125:
606-614 ª 2017 by the American Academy of Ophthalmology

Iris melanomas are rare uveal tumors that develop from
neuroectodermal cells. They represent approximately two
thirds of all primary iris tumors, but only approximately 5%
of uveal melanomas.1,2 Their annual incidence varies be-
tween 2 to 6 cases per 10 million persons.3,4 Their prognosis
seems more favorable than that of choroidal melanomas,
with local control rates of approximately 95% and much
lower metastatic rates on the order of 5% or less, in contrast
to 30% or more for choroidal melanomas. Iridectomy or
sectorial iridocyclectomy has been the mainstay of treatment
of small iris or iridociliary melanomas, whereas enucleation
was the rule for larger tumors.5e8 However, iridectomy is
associated with substantial morbidity and complications,
such as cataracts, hypotonia, corneal damage, retinal
detachment, phthisis,9 and debilitating photophobia, as
reported by patients. Moreover, iridectomy is not exempt
from tumor seeding.10 In view of the excellent oncologic
outcomes of iris melanomas, more conservative treatments
with less morbidity have been investigated. Conservative
ocular treatments were based on similar survival outcomes
with brachytherapy or enucleation with choroidal
melanomas from the 1970s.11,12 Brachytherapy also has
been applied to iris melanomas requiring enucleation.13,14
Complications after brachytherapy include cataracts and
significant inflammation of the anterior segment as well as
corneal damage. Proton therapy is a type of radiotherapy
characterized by a very sharp dose deposition and is a
standard treatment for uveal melanomas. It initially was
performed by Damato et al15 for the treatment of iris
melanoma in 1994. Since then, only a limited number of
studies have reported the long-term outcomes of such an
approach for the management of iris melanomas.15e18
The goals of this study were to evaluate these outcomes,
such as local control of the treatment site, and to report the side
effects and ocular complications of proton therapy for iris
melanomas. The study’s treatment facility is 1 of 12 proton
therapy centers performing ocular treatments worldwide.19
Methods
Patients
All consecutive iris melanoma patients underwent proton therapy
between 1996 and 2015. Patients were referred from 3 oncology

606 ª 2017 by the American Academy of Ophthalmology https://doi.org/10.1016/j.ophtha.2017.10.009

Published by Elsevier Inc. ISSN 0161-6420/17

Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
 Thariat et al 
Proton Beam Therapy for Iris Melanoma
reference centers (Nice and Lyon, France, and Genoa, Italy). The
patients underwent physical examination, ultrasound eye exami-
nation, and ultrasound biomicroscopy before treatment. The tumor
was diagnosed as a melanoma, either clinically or histologically
(either with fine-needle aspiration biopsy or iridectomy). Identifi-
cation of the clinical features suspicious for melanoma remains
challenging. However, positive clinical criteria, as defined by
Shields et al,20 are pigmented iris lesions of 3 mm or more in
diameter, 1 mm in thickness, or both by ultrasound
biomicroscopy. Any of the following signs could be associated
with confirmed growth: a pre-existing nevus, pupillary deforma-
tion, uveal ectropion, tumoral neovessels, sectorial cataract, and
elevated intraocular pressure (Fig 1). Achromic lesions could be
diagnosed as melanoma in the presence of 1 of these associated
signs. Diffuse melanoma refers to iris melanomas presenting as
flat with infiltrating growth pattern or with seeding through the
anterior chamber with confluent or multifocal iris involvement.21
Only patients treated for a primary iris melanoma were included.
Ciliary body melanomas with iris involvement were excluded;
these presented as a main cellular nodule in the ciliary body,
infiltrating the iris root on ultrasound biomicroscopy.
Complications resulting from the extent of the iris tumor were
noted as well as best-corrected visual acuity (VA) using the
Monoyer’s scale. Our onco-ophthalmology proton therapy data-
base has been compiled prospectively in since 1991 to include
patient, tumor thickness, diameter, location, and treatment char-
acteristics. Meanwhile, associated signs and follow-up were filled
in by each of the 3 referring ophthalmologic centers. Informed
consent with information on the risks and benefits of proton ther-
apy was obtained from all of the patients. The ethical review board
of Centre Lacassagne, Nice approved the study, with consideration
to the tenets outlined in the Declaration of Helsinki.
Irradiation Technique
Implantation of 4 radio-opaque (tantalum) fiducial markers at the
surface of the sclera was performed under local anaesthesia 1 to 4
weeks before administration of proton therapy. Atropine drops
were prescribed for 1 month from the implantation of fiducial
placement by ophthalmologists from Genoa and Lyon, but not by
the third team from Nice. Using a treatment planning system
dedicated to ocular proton therapy, fiducials were used for ste-
reotactic modelling of eye orientation starting from the virtual eye
model in macular fixation proposed in the treatment planning
system. Fiducials also were used to define tumor borders. An
ocular planning computed tomography scan was used for eye, lens,
and nerve delineation. The treatment planning software EyePlan
version 1-3.6 (Douglas Cyclotron Centre, Clatterbridge, UK)
aimed to reproduce the true patient gaze and lens position and thus
to generate optimized sparing of anterior chamber structures rather
than those used from a virtual eye only. All patients were treated
with a 65-MeV proton fixed horizontal beam line in a seated po-
sition. For head immobilization, a custom-made thermoplastic
mask and a dental impression (bite block) were used for reliable
and reproducible positioning. Local anesthetic eye drops were used
before blepharostat placement to avoid cutaneous eyelid toxicity.
Being radio-opaque, the fiducials also were used to control eye
positioning using orthogonal radiographs in the treatment room so
as to ensure online image-guided irradiation immediately before
proton beam delivery. Then, during the 10 seconds of beam-on
treatment time, an infrared camera (with images projected on a
screen outside the treatment room for immediate interruption in
case of gaze deviation) was used for gaze monitoring. This full
process allows treatment delivery with submillimetric accuracy.
The beam range was usually selected to place a 90% isodose of the
distal fall of 2.5 mm beyond the tumor. The 90% isodose of the
Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
proximal spread-out-Bragg-peak was placed 2.5 mm in front of the
tumor. The collimator border defined the location of the 50%
isodose. The dose delivered was 60 relative biological effective-
ness in 4 fractions of 15 Gy over 4 days (relative biological
effectiveness is defined as the physical proton dose multiplied by
the relative biological effectiveness [relative to dose delivered with
photons] of 1.1).
Follow-up
Follow-up times were measured from the date of proton therapy to
the date of treatment failure or the last known status. Patients were
reviewed at 1 and 6 months after proton therapy, then twice yearly
for 5 years, and then annually. At each follow-up visit, examina-
tions included measurement of best-corrected VA, slit-lamp
examination, indirect ophthalmoscopy, and ultrasound bio-
microscopy. Patients alternately were examined by their general
ophthalmologists and onco-ophthalmologists. Local control was
defined as the absence of tumor growth and absence of any new
lesion on the treated eye. Ocular complications, such as cataract,
hyphema, elevated intraocular pressure, response to treatment,
palpebral sequelae, dry eye syndrome, keratitis, posterior syn-
echiae, uveitis, and tantalum ring exteriorization were collected.
The tumor was classified either as stable with partial response
(reduction of tumor thickness) or complete response (flat lesion) or
as progressive (increased tumor thickness or diameter). Detection
of metastases was based on 6-month hepatic ultrasound or
computed tomography scan.
Statistical Analysis
Quantitative parameters were described with median and range and
qualitative parameters with frequency and percentage. The cumu-
lative incidence of relapse was assessed using Kaplan-Meier sur-
vival estimates and was capped after 5 years. The log-rank test was
used to assess prognostic factors among initial patient and tumor
characteristics. The degradation of VA, defined by a decrease of
VA in logarithm of the minimum angle of resolution (logMAR)
units between baseline and last follow-up, was described by fre-
quency. Prognostic factors of VA were investigated through a lo-
gistic regression to adjust analyses on VA at baseline and time of
follow-up. Results were explained by an adjusted odds ratio and
95% confidence interval. P values of less than 0.05 were consid-
ered significant. All analyses were carried out using SAS software
version 9.2 (SAS Institute, Inc., Cary, NC).
Results
Patient and Tumor Characteristics
One hundred seven consecutive patients were treated between 1996
and 2015 and were included in the study. Fifty-nine were women
(55.6%; Table 1). Patients were referred by ophthalmologists from
Nice, Lyon, and Genoa in 27.1% (n ¼ 29), 33.6% (n ¼ 36), and
39.3% (n ¼ 42) of cases, respectively. This group represented
4.6% of all uveal melanomas in our population of ocular tumor
patients treated with proton therapy. Median age was 57.0 years
(range, 22.8e86.7 years). Iris color was blue, green, or brown in
46.2% (n ¼ 48), 15.4% (n ¼ 16), and 38.5% (n ¼ 40) of
patients, respectively. These distributions were different among
the centers, with more blue eyes in patients from Nice and Lyon,
representing 59.3% and 55.6% of patients, respectively, whereas
30% of patients from Genoa had blue eyes (P ¼ 0.003). Median
VA before treatment was 1.00 (range, 0e1.00).
All lesions were unilateral and equally distributed on either side
(51 right eyes, 56 left eyes). Tumors were centered on the inferior
607
 Ophthalmology Volume 125, Number 4, April 2018

Figure 1. Iris melanoma criteria shown on (A) slit-lamp examination and (B) gonioscopy: pigmented iris lesion more than 3 mm in diameter and more than
1 mm in thickness with mild uveal ectropion and pupillary deformation. Tumoral neovessels and a small hyphema were noted in the temporal iridocorneal
angle. Trabecular meshwork was invaded by the melanoma.

iris (clocks extending from hours 5 to 7) in 41.1% of the patients,
and in 72.9% of the patients tumors were below the horizontal
diameter (clocks extending below hours 9 to 3). Tumors occupied 1
clock hour in 25.7% (n ¼ 27), up to 2 clock hours in 38.7% (n ¼
41), or more than 2 clock hours in 36.5% (n ¼ 39) of patients.
Median number of clock hours of macroscopic iris involvement
was 2 (range, 1e6, including 3 patients with either iris tapioca in 1
patient from Lyon or diffuse angle involvement in 2 Italian pa-
tients). Thirteen underwent fine-needle aspiration biopsy. Ten pa-
tients underwent primary sectorial iridectomy and proton therapy
for incomplete resection in 9 patients. One patient had a history of
iridectomy for benign lesion for more than 20 years before diag-
nosis of iris melanoma and underwent proton therapy much later
for malignant lesion. Two patients underwent both fine-needle
aspiration biopsy and iridectomy. In patients undergoing iridec-
tomy, median time to proton therapy was 4.5 months (range,
1.1e304.9 months, the latter after iridectomy for a benign lesion).
Reviewing the histologic and cytologic documentation, 6 patients
demonstrated spindle-cell melanoma, 5 patients demonstrated
epithelioid cell melanoma, 3 patients demonstrated atypical cells, 2
patients demonstrated mixed epithelioid cells, and 1 patient
demonstrated either a spindle-cell melanoma or a nevus with
documented tumor growth. Rates of histologic and cytologic
documentation were similar among centers. In contrast, spindle
cells were reported in 0.0% and 14.3% of samples in patients from
Nice and Lyon, respectively, whereas spindle cells were reported in
87.5% of patients from Genoa. For 4 patients, the histologic and
cytologic analysis results were inconclusive. Ocular melanosis
(Ota’s nevus, a predisposing factor for uveal melanomas) was
observed in 1 patient (2.7%).
Clinical Criteria
Median diameter was 4.2 mm (range, 0.8e15.1 mm). Median tu-
mor thickness was 1.5 mm (range, 0.5e9.0 mm). Of available data,
6.6% of tumors were achromic (n ¼ 5/76), 82.9% of tumors (n ¼
63/76) were pigmented, whereas the remainder were at least
partially pigmented. A pre-existing nevus was reported in 33 pa-
tients (30.8%). However, the distribution varied among centers,
with 37.9%, 55.6%, and 49.0% of patients from Nice, Lyon, and
Genoa, respectively (P<0.001). Pupillary deformation was noted
in 68.6% of patients (n ¼ 72/105), uveal ectropion was noted in
40.0% of patients (n ¼ 42/105), neovessels were noted in 21.9% of
patients (n ¼ 23/105), and elevated intraocular pressure was noted
in 12.3% of patients (n ¼ 8/67 of patients with available intraocular
tension data), respectively. Median intraocular pressure before
proton therapy was 14 mmHg (range, 10e30 mmHg). Cataracts
were present at diagnosis in 31.7% of patients (n ¼ 34/101, with
the status unavailable in 6 patients). Thirteen eyes were pseudo-
phakic. Rates of pseudophakic patients were different between
centers, with 6.9%, 5.6%, and 25.7% in Nice, Lyon, and Genoa,
respectively (P ¼ 0.021). Pigmented cells were present in the iri-
docorneal angle in 51.4% of patients (n ¼ 54/105), with varying
proportions among centers: 55.2%, 47.2%, and 74.4% of patients
from Nice, Lyon, and Genoa, respectively (P ¼ 0.048). Ciliary
body involvements were present in 22.4% of patients (n ¼ 24/107),
and involvement of the trabecula was present in 59.1% of patients
(n ¼ 62/105; Fig 1). No extraocular extension was noted.
According to the seventh edition of the American Joint
Committee on Cancer classification, there were 78 T1N0M0
(72.9%) tumors, 23 T2N0M0 tumors (21.5%), and 6 T3N0M0
tumors. Median VA was 1 logMAR (mean, 0.83 logMAR;
range, 0e1 logMAR).
Tumor Control
All patients underwent proton therapy. Seventy-eight patients
(72.9%) received atropine drops and 29 patients (27.1%) did not.
Median follow-up was 49.6 months (range, 1.1e151.6 months),
including 12 patients lost to follow-up less than 1 year after proton
therapy. Median diameter at last follow-up was 2.7 mm (range,
0.0e7.9 mm), with a median decrease of 0.5 mm (range, e12.5 to

608

Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
 Thariat et al 
Proton Beam Therapy for Iris Melanoma

Table 1. Characteristics of the 107 Patients at Baseline

Characteristics Value
Age at diagnosis (yrs), meanSD 57.415.6
Gender (female), no. (%) 47 (44.3)
Center, no. (%)
Nice 29 (27.1)
Lyon 36 (33.6)
Genoa 42 (39.3)
Iris color, no. (%)
Blue 48 (46.1)
Green 16 (15.4)
Brown 40 (38.5)
Visual acuity, median (IQR) 1.00 (0.70e1.00)
Side (right), no. (%) 51 (47.7)
Clocks extending from hour 5 to 7, no. (%) 44 (41.1)
No. of clock hours occupied by the tumor
Median (IQR) 2 (1e3)
No. (%)
1 27 (25.7)
2 41 (38.7)
>2 39 (36.5)
Diameter (mm), median (IQR) 4.49 (2.70e6.40)
Tumour thickness (mm), median (IQR) 1.49 (1.00e20.1)
Pre-existing nevus, no. (%) 33 (30.8)
Pupillary deformation, no. (%) 72 (68.6)
Uveal ectropion, no. (%) 42 (40)
Neovessels, no. (%) 23 (21.9)
Elevated intraocular pressure, no. (%)* 8 (12.3)
Cataract at diagnosis, no. (%) 34 (31.7)
Pigmented (versus achromic), no. (%)y 63 (82.9)
Pigmented cells in the iridocorneal angle, no. (%) 54 (51.4)
Involvement of the trabecula, no. (%) 62 (59.1)
Ciliary body involvement, no. (%) 24 (22.4)
Tumor stage, no. (%)
T1 78 (72.9)
T2 23 (21.5)
T3 6 (5.6)
Atropine, no. (%) 78 (72.9)

IQR ¼ interquartile range; SD ¼ standard deviation.
*Missing data for 42 patients; missing data probably mean pigmented, but
no hypothesis was made.
y
Missing data for 31 patients.
1.3 mm) after treatment (P<0.001). Median tumor thickness was
0.9 mm (range, 0.0e3.5 mm), with a median decrease after proton
therapy of e0.3 mm (range, 5.4e2.5 mm; P<0.001; Fig 2). There
were 5 local relapses (Table 2). Median time to local relapse was
36.3 months (range, 13.1e50.9 months). At 5 years, cumulative
incidence of relapse was 7.5% (95% confidence interval, 3.1%e
17.5%). Among initial patient tumor characteristics, only the
local form (versus diffuse, i.e., with widespread pigments in the
angle) was associated with a risk for iris relapse (with 6.3%
versus 33.3% at 5 years, respectively; P ¼ 0.044, log-rank test).
All 5 patients who demonstrated relapse had an involvement of the
trabecula, and this was associated significantly with an increased
risk for local failure (P ¼ 0.056, log-rank test). None of the patients
with a histologically or cytologically confirmed diagnosis
demonstrated relapsed. There was no difference in terms of local
relapse with respect to treatment under atropine (P ¼ 0.667, log-
rank test). Four patients with relapses occurring in the proton
field were retreated and 1 patient with 2 successive angular relapses
was retreated with proton therapy. Two of these patients underwent
enucleation (1 patient with T2b melanoma and 1 with diffuse T3
disease). At last follow-up, 7 patients had died of causes other than
Figure 2. Follow-up images from patient with iris melanoma who under-
went proton beam therapy. A, Before proton beam therapy, a large ach-
romic lesion was seen in the nasal part of the iris with an underlying
pigmented lesion (insert, gonioscopy). B, Two years after proton beam
therapy, the voluminous achromic lesion had disappeared and only the
pigmented base remained on the iris surface (insert, gonioscopy).
iris melanoma, including 3 of other cancers and 2 of cardiovascular
comorbidities.
Functional Outcomes and Ocular Side Effects
Among 54 patients evaluable (not pseudophakic and without cat-
aracts at diagnosis), cataracts occurred in 31 patients (57.1%). Of
those, 6 patients had undergone phacoemulsification without sig-
nificant complications within a median of 40 months. They were
pseudophakic at last follow-up.
Vision improved or remained stable (as defined by a decrease in
logarithm of the minimum angle of resolution units between
baseline and last follow-up) in 59.4% of patients (n ¼ 60/101).
However, median best-corrected VA at last follow-up was only
slightly lower than at baseline (P<0.001), with 0.9 logMAR
(range, 0.0e1.0 logMAR) versus 1 logMAR at baseline. Addi-
tionally, among the 41 patients whose vision had deteriorated, the
deterioration was more than 0.3 logMAR in 15 patients. Prognostic
factors for any vision loss were more than 2-clock hour

609

Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
 Ophthalmology Volume 125, Number 4, April 2018

involvement (odds ratio, 2.79; 95% confidence interval,

Ciliary Body
Involvement
1.14e6.79; P ¼ 0.024) and involvement of the trabecula (odds

Yes

Yes
No

No
No
ratio, 4.06; 95% confidence interval, 1.67e9.93; P ¼ 0.002).
Twenty-five patients showed cataracts, in addition to the 6 patients
who had undergone phacoemulsification after proton therapy and
potentially were awaiting surgery. One patient demonstrated in-

Hypertonia
(mmHg)
flammatory cystoid macular edema, 5 patients demonstrated atro-

NM
NM
NM
16
14
phic neuropathy secondary to intraocular hypertonia, 1 patient
demonstrated band keratopathy, and 1 patient demonstrated pre-
existing age-related macular degeneration.
Of 58 patients from both Nice and Lyon (n ¼ 67) evaluable for
Cataract

Yes
No
No
No

No
intraocular pressure, 2 additional patients demonstrated elevated
intraocular pressure; it was resolved in 6 of the 8 patients with
elevated intraocular pressure at baseline. It was treated effectively
Neovessels

medically with hypotonic drops and laser in 1 patient or trans-

scleral cyclophotocoagulation in 3 other patients. Within the study
No
No
No
No
No

group, secondary glaucoma was present in 8 patients (7.6%) after

Table 2. Initial Diagnostic Patient and Tumor Characteristics in Patients Who Relapsed

proton therapy, including 4 with neovascular glaucoma. Five pa-

tients showed subsequent atrophic neuropathy. Hyphema was
Deformation

present in 5 patients (4.7%), including 1 patient with neovascular

Pupillary

glaucoma and 1 with secondary glaucoma after treatment. It was

Yes
Yes

Yes
No

No

deemed responsible for transient elevation of intraocular pressure

and resolved with medical treatments. Uveitis was noted in 4 pa-
tients (3.8%), including 2 patients with anterior granulomatous
Ectropion

uveitis treated with local steroids and 2 patients with cystoid

Yes
Yes
No

No

macular edema, 1 of whom showed persistent deterioration of VA.

1

Scleral necrosis was noted in 1 patient (0.9%). Transient dry

corneal syndrome with decreased break-up time was reported
among 35 patients (32.7%) patients from Nice and Lyon, and
Pigments
in angle

Yes
Yes
Yes
Yes

discomfort was treated efficiently with lubricants. Mild and tran-

No

sient conjunctival complications, including hyperemia or eyelid

complications, developed in 80.6% of patients from Lyon and in
13.8% of patients from Nice. One patient showed transient ble-
Nevus
Yes
No
No
No
No

pharitis and 2 showed sectorial madarosis. Chronic corneal com-

plications occurred in 6 patients (5.6%), including 2 with recurring
corneal ulcers requiring bandage contact lens and healing
Clock Hour

eyedrops. Also reported were band keratopathy, recurrent superfi-

Extent

cial punctuate keratitis, recurrent keratalgia, and transient periph-

3
2
2
5
4

eral corneal edema. There were no cases of ocular palsy, myositis,

or radiation retinopathy.
Thickness
(mm)
2.7
2.0
2.0
2.0
2.1

Discussion

This large series of 107 consecutive iris melanoma patients

Diameter (mm)

from 3 onco-ophthalmology centers and this study’s ocular

(T Stage)

10.0 (1a)

6.0 (1b)
2.6 (1b)
NM (3)
4.6 (2)

proton therapy facility corroborates previous pivotal results

by Damato et al15 in that proton therapy is an efficient
technique and has limited morbidity. Proton therapy has
been used as an alternative to surgery or brachytherapy
since 1994,15 but this method remains limited only to
Brown

Brown
Color
Iris

Blue

Blue
Blue

certain centers. In our series, the crude local control rate

was 95.3%. These oncologic outcomes are comparable
with the results of historical brachytherapy and proton
Age (yrs)
31.5
48.8
69.2
77.0
65.7

therapy series of patients with unresectable or resectable

tumors.14e18,22 However, indirect comparisons of the
sparsely published outcomes of iris melanomas treated
Gender

conservatively are of limited validity because selection and

Female
Female
Female

Female

NM ¼ not measured.
Male

interpretation biases cannot be excluded. As an illustration,

local control rates reported by Damato et al15 for localized
iris melanomas and those reported by Willerding et al23
Patient No.

for locally advanced ring melanomas (only resectable at

the price of enucleation) were roughly similar, although
one would expect poorer results for these tumors.
1
2
3
4
5

610

Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
 Thariat et al 
Proton Beam Therapy for Iris Melanoma

Figure 3. Images showing atropine-induced mydriasis and irradiated volume: (A) myosis and (B) atropine-induced mydriasis. By reducing the iris surface,
the area irradiated (yellow circle) also may be reduced. The aim of using atropine is to reduce the risk for synechiae and also may reduce the risk of cataract.
Pupillary monitoring can be performed easily during treatment.

Altogether, the oncologic and functional outcomes of iris
melanomas with proton therapy seem promising.
Ophthalmologists increasingly have been referring their
patients to our proton therapy center since 1996, which is
5 years after initiation of treatments for choroidal
melanomas. Proton therapy now is used systematically as
first-line treatment instead of a conventional sectorial iri-
dectomy or iridocyclectomy, even for small and circum-
scribed tumors of the iris. Complications of surgery, such as
incomplete resection, poor iris reconstruction, and risk of
hemorrhage, infection, or both highlight the limitations of
surgery in favor of proton therapy. Additionally, proton
therapy is associated minimally with photophobia and
diplopia compared with surgical intervention. When
compared with brachytherapy, the dose to the conjunctiva
and cornea can be reduced with proton therapy and the
proton beam may be reduced or enlarged in a more
personalized manner than can be achieved with plaques.
It has been suggested that the better prognosis and lower
metastatic potential of iris melanomas should be tied to
different natural history and genomics compared with
choroidal melanomas. Similar to the other series on iris
melanomas, our patients were younger than patients with
choroidal melanomas in an historical analysis.24e26 Previ-
ous publications showed that iris melanomas displayed
relatively high levels of chromosomal alterations, including
the formation of marker chromosomes, that were possibly
reminiscent of cutaneous melanomas.27 However, recent
research studies suggest that genomics may not be so
different.2,28 As in other series of iris melanomas undergo-
ing conservative treatments, only a minority of patients
underwent fine-needle aspiration biopsy. Consequently, we
could not show any influence of the histologic subtype.
However, similar to choroidal melanomas, an epithelioid
component may behave more aggressively.6 Histologic
confirmation is not recommended because of the risks and
failure rates of puncture resulting from limited tumor bulk.
As a consequence, such confirmation is obtained in less
than 10% of patients, leaving the question open of a
benign versus malignant lesion. In our series, histologic or
cytologic analysis using fine-needle aspiration biopsy was
performed in 15.9% of patients and results were conclusive.
Most fine-needle aspiration biopsies were performed in pa-
tients under observation for an atypically growing lesion.
However, the small quantity of material available is asso-
ciated with low performances overall.29 Thus, it is generally
recommended that an iris-pigmented lesion be observed at
every 3 to 6 months. If tumor growth is documented or if
glaucoma occurs as a complication of a pre-existing nevus,
then treatment may be indicated and a fine-needle aspiration
biopsy may be encouraged to confirm the diagnosis.27,29e31
In this series, almost one third of the patients had docu-
mented growth of a pre-existing nevus, whereas other pa-
tients had a de novo lesion. In de novo cases, there is even
less consensus regarding histologic documentation. In that
context, ultrasound biomicroscopy allows better documen-
tation of iris lesions.32 In particular, it can be used to
investigate growing lesions.33 More recently, OCT
angiography has been proposed as a cost-effective method
for monitoring iris melanocytic lesions for growth and
vascularity. This could be helpful in evaluating tumors for
malignant transformation and response to treatment. How-
ever, it is recognized that penetration of the OCT beam is
limited for highly pigmented tumors34 and also is
insufficient for documenting or excluding extension of the
tumor into the ciliary body.35
To overcome these limitations, Shields et al20 defined
criteria suggestive of the diagnosis of iris melanoma using
an ABCDEF rule similar to that used for cutaneous
melanomas. In our series, 27 patients had fewer than 3 of
the diagnostic criteria recently proposed by Shields et al.
Noteworthy, although the inferior location is not in the
clinical criteria, it is well known that iris melanomas
mostly arise in the inferior iris, as was the case for three
quarters of our population. With conservative treatment,
the possibility that some patients had benign tumors
cannot be excluded. Similar to other series with similar
control rates, some so-called iris melanomas may be only
benign tumors, which would explain the excellent out-
comes. All 5 patients who later demonstrated local relapses

611

Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
 Ophthalmology Volume 125, Number 4, April 2018
within a median of 3 years showed initial involvement of the
trabecula. The relapse either occurred outside of the treat-
ment field or in the iridocorneal angle. Such patterns of
relapse suggest that microscopic extensions may be under-
evaluated36 and that disease may be diffuse in the case of
involvement of the trabecula. Whether involvement of the
trabecula should indicate ring irradiation cannot be
proposed without further confirmation because the rate of
enucleation was low and such irradiation is associated
with a high risk for severe complications.17
We report a 1.9% rate of secondary enucleation for local
relapses. Similarly, Damato et al15 reported a 3.4% crude
rate and a corresponding 6.3% 5-year cumulative inci-
dence of secondary enucleation for local relapse. Similar to
our experience, diffuse melanoma seemed to be a risk factor
for ocular relapse,21,23,37 although statistics could not be
computed because of the low number of events. Rundle
et al18 reported a 6.6% crude rate of secondary enucleation.
Despite a poorer prognosis and higher risk for complications
with large iris melanomas, a few series of diffuse iris
melanomas have been reported that recommend irradiation
first based on their results.23
After proton beam therapy, visual outcomes were rather
good and most side effects were mild and transient with the
exception of glaucoma in 8 patients. Visual outcomes were
dependent on initial tumor characteristics (clock hour extent
and trabecula involvement) that may be associated with
trabecular scarring and lens clouding. Primary complica-
tions were cataracts in more than half of the patients. As
shown in a previous publication from our group by Rahmi
et al,38 one third of cataracts were operated on between the
third and fifth years after irradiation with significant visual
recovery. These cataract rates and outcomes after surgery
of radiation cataracts are consistent with those reported by
Damato et al15 and Shields et al.39 Overall, the surgical
management of radiation-induced cataracts is safe and un-
specific, although mild inflammation may occur.15,16,18,38
Glaucoma was the second most frequent complication. Eight
patients (7.6%) demonstrated secondary glaucoma, either by
angle obliteration or neovascular glaucoma. Rates have been
reported to be up to 20% after proton therapy of iris mela-
noma16,18 and 33% at 5 years after brachytherapy for advanced
unresectable iris melanomas.39 In such cases, glaucoma
occurred through trabecular scarring and rarely by rubeosis
iridis. Accurate reporting of the form of glaucoma is critical
because treatments and outcomes are different. High
intraocular pressure has been reported as a poor prognostic
factor.40 Hyphema associated with elevated intraocular
pressure was noted in less than 5% of patients and resolved
after appropriate medical treatment. Inflammation and uveitis
also were reported in less than 5% of patients and were
reversible in all but 5 patients. As detailed by Rahmi et al,38
corneal complications such as opacities or stromal dystrophies
were rare and often peripheral, occurring in patients with
large tumors compressing the corneal endothelium. These
complications appear less frequently with proton therapy than
a historical series of brachytherapy.14,39,41 In case of extensive
or ring melanomas, corneal complications may be more sig-
nificant. In such situations, limbal stem cell preservation is a
promising technique.42
612
Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
The management of complications can be curative, but
also may be preventable. The current series is interesting in
that it included patients who received atropine systematically
based on their center of origin and unbiased by their own
tumor characteristics. The main objective of atropine was to
prevent posttherapeutic pain (after fiducial placement and
after proton therapy) and to prevent posterior synechiae.
Although not evaluable in this retrospective study, by pre-
venting posterior synechiae, atropine seemed to play a role in
the occurrence of radiation-induced cataracts and should be
evaluated prospectively. Atropine-induced mydriasis also can
be used to reduce the irradiated surface.15 By reducing the
irradiated iris surface, the surface of limbus, lens, and
ciliary body to be irradiated also may be reduced (Fig 3).
This was prescribed systematically by 2 centers. The third
referring center did not for fear of geometric miss. If using
atropine or shorter-life mydriatic drops, proton therapy care-
takers must take iris contraction and volume changes into
account during proton therapy planning.43 They also must
monitor pupillary dilation online to guarantee proper proton
beam projection onto the iris tumor volume. To that extent,
atropine is more comfortable than drops with a shorter half-
life. Further investigations of the benefit of mydriatic drops
will be necessary.
None of the patients exhibited metastases at last follow up.
Similar to the study by Willerding et al,23 some patients were
lost to follow-up. We did not exclude these patients because it
highlights the critical issue of compliance to follow-up visits
for patients with rare diseases carrying a relatively good
prognosis yet requiring specialized treatments that take place
at a distance from their homes. The development of patient-
reported outcomes may be a useful aid to gather informa-
tion in the long term. Underreporting cannot be excluded
because late occurrences of metastases have been reported by
Shields et al,8 with 5-year, 10-year, and 20-year rates of 3%,
5%, and 10%, respectively. Eleven patients were lost to
follow-up within 1 year of treatment. Despite that, median
follow-up in our series was more than 4 years, which is more
than that reported in several other published series.15,16,18
Other study limitations include potential heterogeneity
in patient populations referred by the 3 ophthalmology
centers because different rates were noted by iris color and
indications, and rates of complications were slightly
different. This also may emphasize different selection
criteria used by ophthalmologists to refer patients for
treatment after making a diagnosis of iris melanoma.
Additionally, reporting of complications was not system-
atic, in particular for dry eye and conjunctival complica-
tions. Common toxicity criteria (CTC) grading or any other
grading system was lacking, making assessment of the
severity of complications potentially more subjective. This
study also shows the need for standardization of compli-
cations, such as could be achieved using the international
CTCv4AE (common toxicity criteria version 4 for adverse
events Proton therapy of iris melanoma with 50 CGE: In-
fluence of target volume on clinical outcome) classification.
Standardization of reporting also would improve interseries
comparisons. This is critical in such rare diseases as iris
melanomas not only to assess toxicities and to guarantee
optimized tumor control and finally the risk-benefit ratio.
 Thariat et al 
Proton Beam Therapy for Iris Melanoma
Because this study seems to show a prognostic value of
pigments in the angle and a trend for trabecular involve-
ment, further investigations may be necessary to refine
proton therapy target volumes further in such situations. In
conclusion, our study demonstrated that proton therapy can
be proposed safely and efficiently as first-line conservative
treatment of localized iris melanomas. Continued long-term
assessment is warranted.
Acknowledgment
The authors thank Richard Mauer for his careful review and En-
glish editing.
References
1. Henderson E, Margo CE. Iris melanoma. Arch Pathol Lab
Med. 2008;132:268-272.
2. Krishna Y, Kalirai H, Thornton S, et al. Genetic findings in
treatment-naive and proton-beam-radiated iris melanomas. Br
J Ophthalmol. 2016;100:1012-1016.
3. Jensen OA. Malignant melanoma of the iris. A 25-year anal-
ysis of Danish cases. Eur J Ophthalmol. 1993;3:181-188.
4. McGalliard JN, Johnston PB. A study of iris melanoma in
Northern Ireland. Br J Ophthalmol. 1989;73:591-595.
5. Arentsen JJ, Green WR. Melanoma of the iris: report of 72
cases treated surgically. Ophthalmic Surg. 1975;6:23-37.
6. Conway RM, Chua WC, Qureshi C, Billson FA. Primary iris
melanoma: diagnostic features and outcome of conservative
surgical treatment. Br J Ophthalmol. 2001;85:848-854.
7. Rones B, Zimmerman LE. The production of hetero-
chromia and glaucoma by diffuse malignant melanoma of
the iris. Trans Am Acad Ophthalmol Otolaryngol. 1957;61:
447-463.
8. Shields CL, Shields JA, Materin M, et al. Iris melanoma: risk
factors for metastasis in 169 consecutive patients. Ophthal-
mology. 2001;108:172-178.
9. Naumann GO, Rummelt V. Block excision of tumors of the
anterior uvea. Report on 68 consecutive patients. Ophthal-
mology. 1996;103:2017-2027. discussion 27-28.
10. Gupta M, Puri P, Rennie IG. Iris seeding following
iridocyclectomy for localised iris melanoma. Eye (Lond).
2001;15(Pt 6):808-809.
11. Augsburger JJ, Schneider S, Freire J, Brady LW. Survival
following enucleation versus plaque radiotherapy in statisti-
cally matched subgroups of patients with choroidal mela-
nomas: results in patients treated between 1980 and 1987.
Graefes Arch Clin Exp Ophthalmol. 1999;237:558-567.
12. Collaborative Ocular Melanoma Study Group. The COMS
randomized trial of iodine 125 brachytherapy for choroidal
melanoma: V. Twelve-year mortality rates and prognostic
factors: COMS report no. 28. Arch Ophthalmol. 2006;124:
1684-1693.
13. Finger PT. Plaque radiation therapy for malignant melanoma
of the iris and ciliary body. Am J Ophthalmol. 2001;132:
328-335.
14. Shields CL, Shields JA, De Potter P, et al. Treatment of non-
resectable malignant iris tumours with custom designed plaque
radiotherapy. Br J Ophthalmol. 1995;79:306-312.
15. Damato B, Kacperek A, Chopra M, et al. Proton beam
radiotherapy of iris melanoma. Int J Radiat Oncol Biol Phys.
2005;63:109-115.
Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.
16. Lumbroso-Le Rouic L, Delacroix S, Dendale R, et al. Proton
beam therapy for iris melanomas. Eye (Lond). 2006;20:1300-
1305.
17. Riechardt AI, Karle B, Cordini D, et al. Proton therapy of iris
melanoma with 50 CGE: influence of target volume on clinical
outcome. Strahlenther Onkol. 2017;193(11):943-950.
18. Rundle P, Singh AD, Rennie I. Proton beam therapy for
iris melanoma: a review of 15 cases. Eye (Lond). 2007;21:
79-82.
19. Hrbacek J, Mishra KK, Kacperek A, et al. Practice patterns
analysis of ocular proton therapy centers: the International
OPTIC Survey. Int J Radiat Oncol Biol Phys. 2016;95:
336-343.
20. Shields CL, Kaliki S, Hutchinson A, et al. Iris nevus growth
into melanoma: analysis of 1611 consecutive eyes: the ABC-
DEF guide. Ophthalmology. 2013;120:766-772.
21. Demirci H, Shields CL, Shields JA, et al. Diffuse iris
melanoma: a report of 25 cases. Ophthalmology. 2002;109:
1553-1560.
22. Agraval U, Sobti M, Russell HC, et al. Use of Ruthenium-106
brachytherapy for iris melanoma: the Scottish experience. Br J
Ophthalmol. 2018;102:74-78.
23. Willerding GD, Cordini D, Hackl C, et al. Proton beam
radiotherapy of diffuse iris melanoma in 54 patients. Br J
Ophthalmol. 2015;99:812-816.
24. Bensoussan E, Thariat J, Maschi C, et al. Outcomes after
proton beam therapy for large choroidal melanomas in 492
patients. Am J Ophthalmol. 2016;165:78-87.
25. Thariat J, Grange JD, Mosci C, et al. Visual outcomes of
parapapillary uveal melanomas following proton beam ther-
apy. Int J Radiat Oncol Biol Phys. 2016;95:328-335.
26. Thariat J, Maschi C, Lanteri S, et al. Dry eye syndrome after
proton therapy of ocular melanomas. Int J Radiat Oncol Biol
Phys. 2017;98:142-151.
27. Sisley K, Brand C, Parsons MA, et al. Cytogenetics of iris
melanomas: disparity with other uveal tract melanomas.
Cancer Genet Cytogenet. 1998;101:128-133.
28. Scholz SL, Moller I, Reis H, et al. Frequent GNAQ, GNA11,
and EIF1AX mutations in iris melanoma. Invest Ophthalmol
Vis Sci. 2017;58:3464-3470.
29. Schalenbourg A, Uffer S, Zografos L. Utility of a biopsy in sus-
picious pigmented iris tumors. Ophthalmic Res. 2008;40:267-272.
30. Char DH, Kemlitz AE, Miller T, Crawford JB. Iris ring mela-
noma: fine needle biopsy. Br J Ophthalmol. 2006;90:420-422.
31. Inoue R, Saishin Y, Shima C, et al. A case of iris melanocy-
toma transformed to malignant melanoma. Jpn J Ophthalmol.
2009;53:271-273.
32. Conway RM, Chew T, Golchet P, et al. Ultrasound bio-
microscopy: role in diagnosis and management in 130
consecutive patients evaluated for anterior segment tumours.
Br J Ophthalmol. 2005;89:950-955.
33. Giuliari GP, Krema H, McGowan HD, et al. Clinical and ul-
trasound biomicroscopy features associated with growth in iris
melanocytic lesions. Am J Ophthalmol. 2012;153:1043-1049.
34. Skalet AH, Li Y, Lu CD, et al. Optical coherence tomography
angiography characteristics of iris melanocytic tumors.
Ophthalmology. 2017;124:197-204.
35. Razzaq L, Emmanouilidis-van der Spek K, Luyten GP, de
Keizer RJ. Anterior segment imaging for iris melanocytic tu-
mors. Eur J Ophthalmol. 2011;21:608-614.
36. Sandinha MT, Kacperek A, Errington RD, et al. Recurrence of
iris melanoma after proton beam therapy. Br J Ophthalmol.
2014;98:484-487.
37. Brown D, Boniuk M, Font RL. Diffuse malignant melanoma
of iris with metastases. Surv Ophthalmol. 1990;34:357-364.
613
 Ophthalmology Volume 125, Number 4, April 2018

38. Rahmi A, Mammar H, Thariat J, et al. Proton beam therapy for
presumed and confirmed iris melanomas: a review of 36 cases.
Graefes Arch Clin Exp Ophthalmol. 2014;252:1515-1521.
39. Shields CL, Naseripour M, Shields JA, et al. Custom-designed
plaque radiotherapy for nonresectable iris melanoma in 38
patients: tumor control and ocular complications. Am J Oph-
thalmol. 2003;135:648-656.
40. Shields CL, Materin MA, Shields JA, et al. Factors associated
with elevated intraocular pressure in eyes with iris melanoma.
Br J Ophthalmol. 2001;85:666-669.
41. Rouberol F, Duquesne N, Kodjikian L, et al. [Sarcoid granu-
loma simulating amelanotic melanoma of the iris: a case
report]. J Fr Ophtalmol. 2002;25:831-834.
42. Singh AD, Dupps Jr WJ, Biscotti CV, et al. Limbal stem cell
preservation during proton beam irradiation for diffuse iris
melanoma. Cornea. 2017;36:119-122.
43. Aptel F, Chiquet C, Beccat S, Denis P. Biometric evaluation of
anterior chamber changes after physiologic pupil dilation using
Pentacam and anterior segment optical coherence tomography.
Invest Ophthalmol Vis Sci. 2012;53:4005-4010.

Footnotes and Financial Disclosures

Originally received: August 15, 2017. HUMAN SUBJECTS: Human subjects were included in this study.
Final revision: October 1, 2017. Informed consent with information on the risks and benefits of proton
Accepted: October 4, 2017. therapy was obtained from all of the patients. The ethical review board of
Available online: November 8, 2017. Manuscript no. 2017-1905. Centre Lacassagne, Nice approved the study, with consideration to the te-
1
Department of Radiation Oncology, Cancer Centre Francois Baclesse, nets outlined in the Declaration of Helsinki.
Normandie UniversitedUnicaen, Caen, France. Author Contributions:
2
Department of Radiation Oncology, Cancer Centre Antoine Lacassagne, Conception and design: Thariat, Rahmi, Butet, Caujolle
Nice, France. Analysis and interpretation: Thariat, Rahmi, Salleron, Mosci, Butet,
3
Department of Ophthalmology, Croix-Rousse University Hospital, Uni- Maschi, Lanza, Baillif, Herault, Mathis, Caujolle
versity Claude Bernard Lyon 1, Lyon, France. Data collection: Thariat, Rahmi, Mosci, Butet, Lanza, Lanteri, Herault,
4
Department of Biostatistics and Data Management, Institut de Cancer- Caujolle
ologie de Lorraine, Vandoeuvre-Les-Nancy, France. Overall responsibility: Thariat, Rahmi, Mosci, Herault, Caujolle
5
Department of Ophthalmology, Ocular Oncology Center, E.O. Ospedali Abbreviations and Acronyms:
Galliera, Genoa, Italy. logMAR ¼ logarithm of the minimum angle of resolution; VA ¼ visual
6
Department of Ophthalmology, University Hospital Pasteur 2, Nice, acuity.
France.
Correspondence:
Presented at: Particle Therapy Cooperative Group Annual Meeting, 2017, Juliette Thariat, MD, PhD, Centre Baclesse, 3 av General Harris, Caen
Yamamoto, Japan. 14000, France. E-mail: jthariat@gmail.com.
Financial Disclosure(s):
The author(s) have no proprietary or commercial interest in any materials
discussed in this article.

614

Downloaded for Hendry Agustian (h3ndry_agustian@yahoo.com) at Universitas Tarumanagara from ClinicalKey.com by Elsevier on May 10, 2018.
For personal use only. No other uses without permission. Copyright ©2018. Elsevier Inc. All rights reserved.