Causes of Chemoreduction Failure in

Retinoblastoma and Analysis of Associated

Factors Leading to Eventual Treatment with
External Beam Radiotherapy and Enucleation
Kaan Gündüz, MD,1 İlhan Günalp, MD,1 Nilüfer Yalçındağ, MD,1 Emel Ünal, MD,2 Nurdan Taçyıldız, MD,2
Esra Erden, MD,3 Pınar Özdemir Geyik, PhD1
Purpose: To evaluate the causes of chemoreduction failure in retinoblastoma and to analyze the associated
factors for eventual treatment with external beam radiotherapy and enucleation.
Design: Prospective noncomparative case series.
Participants: Seventy-one patients with 105 eyes with intraocular retinoblastoma that underwent chemore-
duction therapy between October 1998 and January 2003.
Intervention: A 6-treatment cycle of chemoreduction therapy with vincristine, etoposide, and carboplatin
was administered at monthly intervals. Unresponsive disease was defined as persistence of retinal tumors,
vitreous seeds, or subretinal seeds after the second treatment cycle, with no appreciable sign of regression. Eyes
with unresponsive disease were enucleated after the second treatment. Eyes that responded to chemoreduction
therapy received focal treatment, including indirect laser photocoagulation, transpupillary thermotherapy, cryo-
therapy, and ruthenium 106 episcleral plaque radiotherapy after the second chemoreduction treatment, if
necessary, to achieve complete tumor regression. Recurrence was defined as the regrowth of retinal tumors,
vitreous or subretinal seeds after an initial favorable response, and regression. Recurrent retinal tumor, vitreous
seeds, or subretinal seeds were treated with focal treatments and 2 to 3 additional chemoreduction treatments.
When these methods failed or were not applicable, external beam radiotherapy and/or enucleation was administered.
Main Outcome Measures: The use of external beam radiotherapy and enucleation for chemoreduction
failure, which was defined as unresponsive or recurrent disease.
Results: The mean follow-up was 25.7 months (range: 6 – 49). Ten of 105 eyes (9.5%) with unresponsive
disease were enucleated after the second treatment. Of the remaining 95 eyes, 42 (44.2%) developed recurrence
after chemoreduction. Recurrent disease failing to be treated successfully by other methods was treated with
external beam radiotherapy in 26 of 95 eyes (27.4%) and enucleation in 22 of 95 eyes (23.2%). External beam
radiotherapy was successful in preventing enucleation in 20 of 26 eyes (76.9%). Overall, the globe salvage rate
was 69.5%, ranging from 36.1% for Reese–Ellsworth group V disease to 87.0% for groups I to IV disease.
Histopathologically, 29 of 31 enucleated eyes (93.5%) had poorly differentiated or moderately differentiated
retinoblastoma. Using multivariate logistic regression analysis, factors predictive of eventual treatment with
external beam radiotherapy were female gender (P ⫽ 0.010), presence of subretinal seeds (P ⫽ 0.023), and a
greater number of chemoreduction treatments (P ⫽ 0.027). By multivariate analysis, the factors associated with
the need for eventual treatment with enucleation were recurrence of retinal tumors (P ⫽ 0.004), presence of
vitreous seeds (P ⫽ 0.008), greater tumor thickness (P ⫽ 0.015), presence of subretinal fluid (P ⫽ 0.040), and
older patient age (P ⫽ 0.042).
Conclusions: Chemoreduction failure in this article was defined as unresponsive or, more commonly,
recurrent retinoblastoma. Older patient age, greater tumor thickness, presence of vitreous seeds and subretinal
fluid at baseline, and retinal tumor recurrence after chemoreduction were factors associated with the need for
enucleation. Ophthalmology 2004;111:1917–1924 © 2004 by the American Academy of Ophthalmology.

For many years, the standard treatment for unilateral reti- eye with the more advanced tumor was enucleated, and
noblastoma was enucleation. In bilateral retinoblastoma, the external beam radiotherapy was given in an attempt to

3
Originally received: September 26, 2003. Department of Pathology, Ankara University Faculty of Medicine, An-
Accepted: April 6, 2004. Manuscript no. 230645. kara, Turkey.
1
Ocular Oncology Service, Department of Ophthalmology, Ankara Uni- The authors have no proprietary interest in the instruments or products used
versity Faculty of Medicine, Ankara, Turkey. in this article.
2
Department of Pediatrics, Ankara University Faculty of Medicine, An- Reprint requests and correspondence to Kaan Gündüz, MD, G.M.K.Bulvarı
kara, Turkey. 116/3, Maltepe 06570, Ankara, Turkey. E-mail: eyemd@ada.net.tr.

© 2004 by the American Academy of Ophthalmology ISSN 0161-6420/04/$–see front matter 1917
Published by Elsevier Inc. doi:10.1016/j.ophtha.2004.04.016
 Ophthalmology Volume 111, Number 10, October 2004
Table 1. Chemoreduction Regimen for Patients with
Intraocular Retinoblastoma Receiving 6-Treatment Cycle of
Chemotherapy*
Day Vincristine† Etoposide‡ Carboplatin§
0 X X X
1 X
*This regimen is repeated once a month for 6 mos.
†
1.5 mg/m2 (0.05 mg/kg for children ⱕ36 months old and maximum dose
2 mg).
‡
150 mg/m2 (5 mg/kg for children ⱕ36 months old).
§
560 mg/m2 (18.6 mg/kg for children ⱕ36 months old).
preserve the less affected eye. During the past decade,
chemoreduction has become a more commonly used com-
ponent of the initial management of retinoblastoma.1–5 Che-
moreduction of tumor volume allows focal, less damaging
therapeutic measures to be used subsequently. The chemo-
therapy agents differ according to the pediatric oncologist;
most centers presently employ vincristine, etoposide, and
carboplatin. The chemotherapy regimen consists of 6 treat-
ments to allow for adequate tumor reduction.1 Focal treat-
ment methods, including laser photocoagulation, transpupil-
lary thermotherapy, cryotherapy, and plaque radiotherapy,
are applied to individual tumors, usually after the second or
third treatment, with the goal of inducing complete regres-
sion of the tumor.1–3,5– 8
The ocular salvage rate has improved with the addition
of chemoreduction to treatment regimens. Many patients
with retinoblastoma that would otherwise have been treated
with enucleation or external beam radiotherapy9 now un-
dergo chemoreduction and focal treatment. However, che-
moreduction and focal treatment methods are not uniformly
successful in controlling intraocular retinoblastoma, with
chemoreduction failure being defined as unresponsive or
recurrent retinoblastoma. Recurrences can occur at 3 ana-
tomic locations as retinal tumor, vitreous seeds, and sub-
retinal seeds.10 Eyes that fail chemoreduction therapy even-
tually require external beam radiotherapy and/or
enucleation.
The purpose of this study was to evaluate the outcome of
chemoreduction treatment at our center. We studied the
causes of chemoreduction failure and analyzed the factors
predictive of eventual treatment with external beam radio-
therapy and enucleation.
Materials and Methods
Between October 1998 and January 2003, 105 eyes of 71 patients
with intraocular retinoblastoma underwent chemoreduction treat-
ment. The chemoreduction protocol was approved by the pediatric
oncology review board. Informed consent was obtained from all
patients. The chemotherapeutic agents employed in the protocol
included IV vincristine, etoposide, and carboplatin, as shown in
Table 1. Chemoreduction therapy consisted of 6 treatments admin-
istered at monthly intervals. Examination under anesthesia by one
of the authors (KG) was performed at baseline and every month
until complete regression of the intraocular disease was achieved.
1918
Cranial magnetic resonance imaging was done at baseline and
twice yearly thereafter until age 5 years to detect pinealoblastoma
or trilateral retinoblastoma.
Any patient whose tumor(s) could be adequately controlled
with conservative methods alone (cryotherapy, laser photocoagu-
lation, transpupillary thermotherapy, and plaque radiotherapy) was
not included in this study. Exclusion criteria for treatment with
chemoreduction included iris neovascularization, neovascular
glaucoma, and tumor invasion into the anterior chamber, iris, optic
nerve, and choroid, as documented by clinical and ultrasono-
graphic modalities. These are risk factors for systemic metastasis
of retinoblastoma,11 and patients having these risk factors were
treated with enucleation. Systemic exclusion criteria included ev-
idence of metastasis or inadequate renal or hepatic function.
All data were collected in a prospective fashion. A detailed
fundus drawing was made, indicating the location of each tumor or
seed. In selected cases, fundus photographs were obtained. Each
retinoblastoma was measured for the greatest basal dimensions (in
millimeters) by indirect ophthalmoscopy and for thickness (in
millimeters) by A-scan and B-scan ultrasonography and indirect
ophthalmoscopy. The presence of subretinal fluid, subretinal seeds,
and vitreous seeds was recorded.
Unresponsive disease was defined as persistence of retinal
tumors, vitreous seeds, or subretinal seeds with no appreciable sign
of regression after the second treatment cycle. Eyes with unre-
sponsive disease were enucleated after the second treatment. In
eyes that responded to chemoreduction, focal treatment methods,
including indirect diode laser photocoagulation, transpupillary
thermotherapy, cryotherapy, and ruthenium 106 (Ru106) episcleral
plaque radiotherapy, were applied after the second chemoreduction
treatment, if necessary, to achieve complete tumor regression. The
end point of treatment was to obtain either a type 1 or type 4
regression of retinal tumors (type 1: completely calcified mass
looking like the cottage cheese type; type 4: flat atrophic scar
tissue). For tumors displaying a type 2 or 3 regression (type 2:
completely noncalcified mass looking like fish flesh; type 3: mixed
calcified and noncalcified mass), additional focal treatments were
given to achieve the desired regression pattern.
Recurrence was defined as the regrowth of original tumors
(retinal tumors, vitreous seeds, or subretinal seeds) after an initial
favorable response and regression. Recurrent disease was treated
with focal treatment methods. In selected cases, 2 to 3 additional
chemoreduction treatments were given to reduce tumor volume so
that focal methods could be used, provided the general health
status of the child allowed additional chemotherapy after the
standard 6 treatments. When these methods failed or were not
applicable, external beam radiotherapy and/or enucleation was
administered. New tumor was defined as the development of a
tumor during or after chemoreduction that was not noted at base-
line examination. New tumors were treated with focal methods,
including transpupillary thermotherapy and cryotherapy.
The indications for various focal treatment methods have been
previously described.12 Indications for cryotherapy included equa-
torial and peripheral retinoblastoma of ⬍4 mm in base diameter
and ⬍3 mm in thickness, without vitreous seeds. Thermotherapy
was used for retinoblastoma of ⬍3 mm in base diameter and ⬍3
mm in thickness, without vitreous seeds located at or posterior to
the equator. Occasionally, fish flesh portions of larger diameter
retinoblastoma regressed in a type 3 fashion were also treated with
thermotherapy. A continuous mode was employed. Indications for
Ru106 plaque radiotherapy comprised retinoblastoma of ⬍15 mm
in base diameter and ⬍6 mm in thickness, with minimal overlying
vitreous seeds. A total dose of 40 Gy was delivered to the apex of
the tumor.12 Tumors ⬎6 mm thick can not be treated with Ru106
plaque radiotherapy because of the limited tissue penetration of
this radioisotope. Laser photocoagulation was used for retinoblas-
 Gündüz et al 䡠 Chemoreduction in Retinoblastoma
Table 2. Response to Chemoreduction and Additional Treatment Given, According to Reese–Ellsworth Group
Controlled by Focal
Reese–Ellsworth Chemotherapy Treatment
Group (No) Alone Given
I (4) — 4
II (23) — 23
III (7) — 7
IV (35) 4 27
V (36) 1 14
EBRT ⫽ external beam radiotherapy.
toma of ⬍3 mm in base and ⬍3 mm in thickness, without vitreous
seeds located posterior to the equator. The goal of laser photoco-
agulation was to surround the tumor and close off all feeder
vessels.
The external beam radiotherapy dose ranged from 35 to 45 Gy,
based on previous studies. Either a whole eye or a lens-sparing
technique was used, depending on the stage of the intraocular
disease.9,13 External beam radiotherapy or enucleation was
deemed necessary if the recurrent retinal tumor was ⬎7 mm thick
and ⬎15 mm in base diameter and if the recurrent subretinal and
vitreous seeds involved ⬎2 quadrants of the fundus. In bilateral
cases, enucleation was offered for the first eye and external beam
radiotherapy for the remaining eye after enucleation of the first
eye. We did not offer external beam radiotherapy in unilateral
cases after chemoreduction failure. However, the final treatment
decision was complex and based on discussion with the parents;
external beam radiotherapy was sometimes substituted for enucle-
ation.
Pathologic Examination
The enucleated eyes were immersed in 10% buffered formalde-
hyde solution. After macroscopic examination, the eye was cut
along a plane that included the bulk of the tumor in the main
pupil– optic nerve block. Serial sections were also obtained in
several levels. Paraffin-embedded samples were cut in sections 4
␮m thick and stained with hematoxylin– eosin.
The degree of tumor differentiation was based on the percent-
age of differentiated areas with Flexner–Wintersteiner rosettes
relative to the total tumor area.14 The tumor was classified as well
differentiated when Flexner–Wintersteiner rosettes were present in
⬎80% of the tumor area, and moderately differentiated if Flexner–
Wintersteiner rosettes were present in ⬍80% of the tumor area.
The tumor was considered poorly differentiated if Flexner–Win-
tersteiner rosettes were absent.
Statistical Analysis
The factors leading to eventual treatment with external beam
radiotherapy and enucleation after chemoreduction of intraocular
retinoblastoma were analyzed using univariate and multivariate
logistic regression analysis. The variables that were entered in the
univariate logistic regression analysis for eventual treatment with
external beam radiotherapy were patient age (as continuous vari-
able), gender, Reese–Ellsworth category of the tumor (groups
IV–V vs. I–III), largest basal tumor diameter (in millimeters as a
continuous variable), tumor thickness (in millimeters as a contin-
uous variable), number of retinal tumors (as a continuous vari-
able), presence of subretinal seeds, presence of vitreous seeds,
presence of subretinal fluid, number of focal treatment sessions (as
a continuous variable), number of chemoreduction treatments (as a
Recurrence
EBRT Enucleation 1 yr 2 yrs 3 yrs
— — 1 — —
6 1 3 2 1
2 1 3 1 —
7 7 9 2 —
11 23 17 3 —
continuous variable), recurrence of retinal tumors, recurrence of
subretinal seeds, and recurrence of vitreous seeds. The univariate
logistic regression analysis for eventual treatment with enucleation
included all the variables listed above plus previous administration
of external beam radiotherapy. Of those variables that showed a
high degree of correlation, only one from the set of associated
variables was entered at a time into subsequent multivariate mod-
els. A final multivariate model fitted variables that were identified
as significant predictors (P⬍0.05) from the initial model, as well
as variables deemed clinically important for the 2 outcome events.
Results
Of the 71 patients included in this study, 40 had bilateral retino-
blastoma, and 31 had unilateral retinoblastoma. A total of 105 eyes
were included in the study. Ten patients had a family history of
retinoblastoma. Six eyes of 6 patients with bilateral retinoblastoma
had been enucleated elsewhere because of advanced intraocular
disease before being referred to us. Of the 71 patients, 35 were
male and 36 were female. The mean age of the patients was 20.8
months (range: 2–96).
The distribution of eyes according to the Reese–Ellsworth
classification system is shown in Table 2. Of 105 eyes, 4 were
Reese–Ellsworth group I; 23, group II; 7, group III; 35, group IV;
and 36, group V. The mean number of retinal tumors per eye was
2.0 (range: 1–14). The mean maximum tumor base diameter was
13.0 mm (range: 2.0 – 40.0), and the mean tumor thickness was 5.0
mm (range: 0 –15.0). Vitreous seeds were present in 27 of 105 eyes
(25.7%), subretinal seeds in 26 of 105 eyes (24.8%), and subretinal
fluid in 43 of 105 eyes (41.0%) at baseline.
The mean number of chemoreduction treatments administered
was 6.3 (range: 2–9). All patients developed transient cytopenia
and alopecia after chemotherapy. Fever and severe neutropenia
were observed in 26 patients (36.6%). However, none of the
patients developed neuropathy, hearing loss, renal or hepatic tox-
icity, or secondary tumors related to chemotherapy.
The mean follow-up was 25.7 months (range: 6 – 49). Table 3
shows the end result of chemoreduction treatment with or without
additional focal treatment. Overall, focal treatments were used in
75 eyes (Fig 1). The most commonly used focal treatment was
transpupillary thermotherapy (71 eyes), followed by cryotherapy
(10 eyes), Ru106 plaque radiotherapy (2 eyes), and indirect laser
photocoagulation (2 eyes) (Table 3). The mean number of focal
treatments applied per eye was 3.9 (range: 2–12). In the remaining
30 eyes, focal treatments were not used (Table 3, Fig 2). These 30
eyes comprised 3 subgroups: (1) eyes with unresponsive disease,
which were enucleated for tumor control after the second treatment
(10 eyes); (2) eyes that showed some improvement after 2 che-
moreduction treatments but could not be treated with focal meth-
ods because of extensive retinal tumor, vitreous seeds, or subreti-
1919
 Ophthalmology Volume 111, Number 10, October 2004

Table 3. End Result of Chemoreduction Therapy with or without Additional Focal Treatment
No. of
Management Eyes Recurrence (yrs) End Result (No. of Eyes)
1 2 3 NR AFT EBRT Enucleation
Focal treatment added at 2 months 75 19 7 1 48 22 18 14
Laser photocoagulation 2 — — — — 0 — —
Transpupillary thermotherapy 71 — — — — 14 — —
Cryotherapy 10 — — — — 12 — —
Ruthenium 106 plaque radiotherapy 2 — — — — 8 — —
No focal treatment 30 14 1 0 5 0 8 18
Enucleation after 2 cycles 10 — — — — — 0 10
Large macular tumor controlled by 5 — — — 5 — 0 0
chemotherapy alone
Extensive tumor/seeds initially 15 — — — — — 8 8
responsive to chemotherapy but not
amenable to focal treatment

AFT ⫽ additional focal treatment; EBRT ⫽ external beam radiotherapy; NR ⫽ no recurrence.

Dashes indicate nonapplicability.

nal seeds (15 eyes); and (3) eyes with large macular tumors that
responded well to chemotherapy alone with a type 1 regression
pattern after 2 treatments (5 eyes).
Excluding the 10 eyes that were enucleated for tumor control
after the second cycle, retinal tumors were present in 95 eyes,
subretinal seeds in 22 eyes, and vitreous seeds in 21 eyes. Retinal
tumors recurred in 34 of 95 eyes (35.8%), subretinal seeds recurred
in 4 of 22 eyes (18.2%), and vitreous seeds recurred in 11 of 21
eyes (52.4%). Some eyes had more than one type of tumor or seed
recurrence. Overall, recurrence was noted in 42 of 95 eyes
(44.2%). New tumors were noted in 15 of 95 eyes (15.8%).
Recurrent disease was treated with focal treatments including
transpupillary thermotherapy, cryotherapy, and Ru106 plaque ra-
diotherapy in 22 of 95 eyes (23.2%), and 2 to 3 additional che-
moreduction treatments in 20 of 95 eyes (18.0%). Eventually,
external beam radiotherapy was necessary in 26 of 95 eyes
(27.4%) and enucleation in 22 of 95 eyes (23.2%) (Table 3)
because other methods failed or were not applicable. External
beam radiotherapy was successful in avoiding enucleation in 20 of
26 eyes (76.9%). Six eyes (23.1%) had external beam radiotherapy
first and were later enucleated.
New tumors were successfully treated with focal methods,
including transpupillary thermotherapy (13 eyes) and cryotherapy
(6 eyes) in all 15 eyes. However, 2 eyes with new tumors under-
went external beam radiotherapy, and 9 eyes with new tumors
were enucleated because of coexisting recurrent disease.
Twenty of 42 eyes (47.6%) with recurrent disease had group V
disease, 11 eyes (26.2%) had group IV disease, and the remaining
11 eyes (26.2%) had groups I to III disease (Table 2). Tables 2 and
3 show that most retinal tumor, vitreous, and subretinal seed
recurrences occurred during the first 2 years after initiation of
chemoreduction in all Reese–Ellsworth groups.
Table 2 shows the response to chemoreduction and additional
treatment according to Reese–Ellsworth group. Of 26 eyes under-
going external beam radiotherapy, 6 were Reese–Ellsworth group
II, 1 was group III, 8 were group IV, and 11 were group V. Overall,
15 of 69 groups I to IV eyes (21.7%) required external beam
radiotherapy, versus 11 of 36 group V eyes (30.6%). Using uni-
variate logistic regression analysis, the factors associated with the
eventual need of external beam radiotherapy were greater number
of chemoreduction treatments (P ⫽ 0.013), female gender (P ⫽
0.023), and presence of subretinal seeds (P ⫽ 0.041) (Table 4).
Using multivariate logistic regression analysis, the factors predic-
tive of eventual treatment with external beam radiotherapy were
female gender (P ⫽ 0.010), presence of subretinal seeds (P ⫽
0.023), and greater number of chemoreduction treatments (P ⫽
0.027) (Table 4).
Eventually, 32 eyes, including 10 with unresponsive or persis-
tent disease and 22 with recurrent disease, were enucleated. Over-
all, the globe salvage rate was 69.5%. The distribution of the
enucleated eyes according to the Reese–Ellsworth grouping is
shown in Table 2. Globe salvage rates were 100% in group I
retinoblastoma, 95.7% in group II retinoblastoma, 85.7% in group
III retinoblastoma, 80.0% in group IV retinoblastoma, and 36.1%
in group V retinoblastoma. Nine of 69 eyes (13.0%) with groups I
to IV disease required enucleation, as opposed to 23 of 36 eyes

Figure 1. Left, Inferiorly located retinoblastoma focus in the right eye

measuring 6⫻6 mm in base dimension and 3 mm in thickness (by ultra- Figure 2. Left, B-mode ultrasonogram shows a 10-mm-thick macular
sonography). Right, After 6 chemoreduction treatments and 2 sessions of retinoblastoma. Right, After 6 chemoreduction treatments and no focal
transpupillary thermotherapy, the tumor is completely atrophic (type IV treatment, the tumor is completely calcified, with orbital shadowing. The
atrophy). tumor thickness remains at 4 mm.

1920
 Gündüz et al 䡠 Chemoreduction in Retinoblastoma

Table 4. Significant Variables Leading to Eventual Treatment with External Beam Radiotherapy with Use of Univariate and
Multivariate Logistic Regression Analyses

95% Confidence
Significant Variable P Odds Ratio Interval
Univariate analysis
Greater no. of chemoreduction treatments 0.013 1.5 1.1–2.0
Female vs. male gender 0.023 0.3 0.1–0.9
Presence of subretinal seeds 0.041 2.7 1.0–6.9
Multivariate analysis
Female vs. male gender 0.010 0.2 0.1–0.7
Presence of subretinal seeds 0.023 3.6 1.2–10.7
Greater no. of chemoreduction treatments 0.027 1.5 1.1–2.3

(63.9%) with group V disease. Histopathologic examination of the
eyes that came to enucleation showed that the tumor was poorly
differentiated in 21 of 31 eyes (67.7%), moderately differentiated
in 8 of 31 eyes (25.8%), and well differentiated in 2 of 31 eyes
(6.5%). Histopathologic data on 1 eye could not be obtained
because the patient underwent enucleation elsewhere.
Using univariate logistic regression analysis, the factors asso-
ciated with the need for eventual enucleation were presence of
vitreous seeds (P⬍0.001), presence of subretinal fluid (P⬍0.001),
recurrence of vitreous seeds (P⬍0.001), larger tumor base diam-
eter (P⬍0.001), greater tumor thickness (P⬍0.001), presence of
retinal tumor recurrence (P ⫽ 0.001), groups IV or V (vs. I–III)
(P ⫽ 0.001), lesser number of chemoreduction treatments (P ⫽
0.010), lesser number of focal treatment sessions (P ⫽ 0.036),
older patient age (P ⫽ 0.045), and greater number of retinal tumors
(P ⫽ 0.046) (Table 5). Using multivariate logistic regression
analyis, the factors predictive of eventual enucleation were recur-
rence of retinal tumors (P ⫽ 0.004), presence of vitreous seeds
(P ⫽ 0.008), greater tumor thickness (P ⫽ 0.015), presence of
subretinal fluid (P ⫽ 0.040), and older patient age (P ⫽ 0.042)
(Table 5).
Two patients had trilateral retinoblastoma characterized by the
presence of bilateral intraocular retinoblastoma and pinealoblas-
toma. One of the 2 patients died of intracranial pinealoblastoma at
4 months after initiation of chemoreduction. An autopsy could not
be performed. No patient developed second malignant neoplasms
during the follow-up period. There were no other deaths from
retinoblastoma during the follow-up period.
Discussion
The use of chemotherapy in the treatment of intraocular
retinoblastoma dates back nearly 50 years. Intravenous ni-
trogen mustard was the first chemotherapeutic drug used in
the management of this condition.15,16 Later, improved oc-
ular salvage rates using intracarotid triethylenemelamine in
combination with external beam radiotherapy were report-
ed.17 Since that time, there have been several reports on the
use of chemotherapy, particularly for extraocular retinoblas-
toma.18 –20 The use of chemoreduction for intraocular reti-
noblastoma was reevaluated in the early 1990s, and the
pioneering articles on the subject were published in
1996.1– 4
Previous studies generally reported favorable ocular sal-
vage rates (95%8 to 100%2,3) with chemoreduction for
Reese–Ellsworth groups I to IV disease. In a study by
Shields et al of 158 eyes, 15% of groups I to IV eyes
required enucleation by 5 years.21 In our series, 9 of 69
groups I to IV eyes (13.0%) required enucleation. The need

Table 5. Significant Variables Leading to Eventual Treatment with Enucleation with Use of Univariate and Multivariate Logistic
Regression Analyses

95% Confidence
Significant Variable P Odds Ratio Interval
Univariate analysis
Presence of vitreous seeds ⬍0.001 8.4 3.3–21.8
Presence of subretinal fluid ⬍0.001 10.2 3.8–27.3
Larger tumor base diameter ⬍0.001 1.2 1.1–1.3
Greater tumor thickness ⬍0.001 1.4 1.2–1.6
Recurrence of vitreous seeds ⬍0.001 0.3 2.8–70.0
Presence of retinal tumor recurrence 0.001 4.6 1.9–11.2
Groups IV–V vs. I–III retinoblastoma 0.001 11.7 2.6–52.6
Lesser no. of chemoreduction treatments 0.010 0.6 0.4–0.9
Lesser no. of focal treatment sessions 0.036 0.8 0.7–1.0
Older patient age 0.045 1.0 1.0–1.0
Greater no. of retinal tumors 0.046 0.6 0.4–0.9
Multivariate analysis
Recurrence of retinal tumors 0.004 8.4 2.0–34.9
Presence of vitreous seeds 0.008 6.6 1.6–26.4
Greater tumor thickness 0.015 1.3 1.1–1.7
Presence of subretinal fluid 0.040 5.2 1.1–25.0
Older patient age 0.042 1.0 1.0–1.1

1921
 Ophthalmology Volume 111, Number 10, October 2004
for external beam radiotherapy in groups I to IV eyes was
reported as 31.8% by Wilson et al.16 Shields et al found that
10% of groups I to IV eyes required external beam radio-
therapy by 5 years.21 In our series, 15 of 69 eyes (21.7%)
with groups I to IV disease required external beam radio-
therapy.
Compared with groups I to IV disease, group V disease
more often requires external beam radiotherapy and enucle-
ation after chemoreduction. Considering the potential side
effects of external beam radiotherapy, including secondary
cancers22–24 and facial deformity, some centers have been
more reluctant than others to administer external beam
radiotherapy. Therefore, the proportion of eyes receiving
external beam radiotherapy has varied. The numbers re-
ported in various series were 19%,2 57%,8 70%,4 and 75%5
for group V retinoblastoma eyes. Shields et al found that
47% of group V eyes required external beam radiotherapy
by 5 years.21 In our series, 11 of 36 (30.6%) group V
retinoblastoma eyes underwent external beam radiotherapy.
The globe salvage rate for group V retinoblastoma also
varied among different studies, with the following rates
reported: 0%,4 64%,8 75%,25 and 78%.3,5 Shields et al
reported an ocular salvage rate of 47% for group V eyes by
5 years.21 The globe salvage rate was 36.1% for group V
retinoblastoma in our series.
The 2 main outcome measures of our study were the use
of external beam radiotherapy and enucleation because of
chemoreduction failure. Chemoreduction failure was de-
fined as unresponsive or recurrent disease. Unresponsive
disease in 10 of 105 (9.5%) study eyes was treated with
enucleation after the second chemoreduction treatment. Re-
current disease in 42 of the remaining 95 eyes (44.2%)
required treatment with external beam radiotherapy in 26
eyes (27.4%) and with enucleation in 22 eyes (23.2%), after
failure of other methods.
Recurrence of vitreous seeds was noted in 52.4% of the
eyes, recurrence of retinal tumors in 35.8% of the eyes, and
recurrence of subretinal seeds in 18.2% of the eyes after
chemoreduction. That the recurrence rate of subretinal seeds
is lower than those of retinal tumor and vitreous seeds is
interesting. It is possible that we may have underestimated
subretinal seed recurrence in advanced group V retinoblas-
toma patients because of other confounding factors, such as
massive tumor, extensive subretinal fluid, and vitreous seed-
ing. Microscopic subretinal seeds may be difficult to detect
in eyes with bullous subretinal fluid. Although most recur-
rences occur in the first year after initiation of chemoreduc-
tion, recurrence may take place as long as 2 years after
treatment. Shields et al similarly found that most retinal
tumor, vitreous seed, and subretinal seed recurrences oc-
curred at mean periods of 2 to 4 months after completion of
chemoreduction.10 They also reported that tumor and seed
recurrences may develop at extended periods, ranging up to
3 years.10 These results attest to the importance of continued
follow-up for at least 3 years after chemoreduction in reti-
noblastoma management.
New tumors were noted in 15.8% of the retinoblastoma
eyes in our study. Shields et al reported that new tumors
developed in 23% of the eyes by 1 year after chemoreduc-
tion.26 Wilson et al reported the rate of development of new
1922
tumors after chemoreduction as 5.6%.16 All new tumors in
our series were observed in eyes with recurrent disease. The
new tumors were treated successfully with focal methods,
and in no eye was external beam radiotherapy or enucle-
ation necessary because of the new tumor. The fact that all
new tumors developed in eyes with recurrent disease may
indicate a generalized chemoresistance problem in these
patients.
Based on multivariate analysis, factors predictive of ex-
ternal beam radiotherapy in our study were female gender,
presence of subretinal seeds, and greater number of che-
moreduction treatments. A different subset of multivariate
factors, including non-Caucasian race, male gender, and
Reese–Ellsworth group V disease, was found to be predic-
tive of eventual treatment with external beam radiotherapy
in Shields et al’s study.21 In our study, the presence of
subretinal seeds was a factor predictive of treatment with
external beam radiotherapy for 2 reasons. First, subretinal
seeds may reflect the discohesive nature of the retinal tumor.
Such tumors may undergo fragmentation after chemoreduc-
tion, releasing several tiny foci of retinoblastoma in the
subretinal space. Second, inadequate penetration of chemo-
therapy to the relatively avascular subretinal space may
cause undertreatment of these tumor foci. Shields et al
reported that the presence of subretinal seeds may be a risk
factor for retinal tumor recurrence.10 In our study, a greater
number of chemoreduction treatments (as a continuous vari-
able) did not seem to protect the eye from the need for
external beam radiotherapy. Based on this finding, it seems
unlikely that administration of additional chemoreduction
treatments will spare retinoblastoma that recurs after the
standard 6-treatment chemoreduction protocol from exter-
nal beam radiotherapy. Therefore, we stopped giving addi-
tional chemoreduction treatments for recurrent retinoblas-
toma after the standard 6-treatment protocol. Furthermore,
excessive chemotherapy and subsequent external beam ra-
diotherapy may be toxic to the eye.4
In our series, multivariate factors predictive of eventual
treatment with enucleation were older patient age, greater
tumor thickness, presence of vitreous seeds and subretinal
fluid at baseline, and recurrence of retinal seeds. In the study
of Shields et al, a different subset of multivariate factors was
found to be predictive of eventual treatment with enucle-
ation, including patient age over 12 months, single tumor in
the eye, and tumor proximity to foveola within 2 mm.21 In
our study, older patient age (as a continuous variable) was
a factor associated with the eventual need for enucleation
because larger tumors diagnosed at a later age are more
difficult to control with chemoreduction. Greater tumor
thickness (as a continuous variable) and presence of vitre-
ous seeds indicate advanced tumors and represent a threat
for enucleation. Furthermore, thick tumors are prone to
tumor dispersion related to chemotherapy, undergoing frag-
mentation and releasing seeds into the vitreous cavity.8,16
The presence of vitreous seeds is a risk factor for globe loss,
because the vitreous cavity is avascular and the vitreous
seeds may not have sufficient access to chemotherapy. The
presence of subretinal fluid may signify the presence of
microscopic subretinal seeds that may not be visible oph-
thalmoscopically. These subretinal seeds constitute a prob-
 Gündüz et al 䡠 Chemoreduction in Retinoblastoma
lem because, in the relatively avascular subretinal space,
they may not have sufficient access to chemotherapy. Re-
currence of retinal tumors after chemoreduction and focal
treatments indicates chemoresistance, leading to enucle-
ation.
Histopathologically, 93.5% of enucleated eyes in our
study had poorly differentiated (67.7%) or moderately dif-
ferentiated (25.8%) retinoblastoma, attesting to the relation-
ship between poor differentiation and worse globe progno-
sis. Previous studies have found that the clinical regression
pattern of retinoblastoma is closely correlated with histopa-
thology.27,28 Types 2 and 3 regression patterns are usually
associated with viable retinoblastoma cells. However, a
tumor with a type 3 regression pattern may be well differ-
entiated, with benign retinocytomalike features.28 There-
fore, a stable type 3 regressed tumor can be observed
cautiously in the absence of other risk factors, including
recurrent subretinal and vitreous seeds.
Macular tumors may be amenable to cure by chemother-
apy alone because of the rich vascular supply in the macular
region provided by the short posterior ciliary arteries.29
Therefore, some macular retinoblastomas may regress com-
pletely in a type 1 pattern with chemoreduction alone,
without focal treatment, as in our study and as reported
previously.29
The 3 chemotherapeutic drugs used in chemoreduction
treatment are associated with a number of hematologic and
nonhematologic side effects. Carboplatin may be associated
with hematologic toxicity, primarily thrombocytopenia.30
The nonhematologic toxicities of carboplatin include neph-
rotoxicity, ototoxicity, and neurotoxicity.30 These toxic ef-
fects are seen at doses of carboplatin exceeding 800 mg/m2;
therefore, they are are unusual in retinoblastoma patients.
The dose-limiting toxic effect of vincristine is neurotoxicity
involving peripheral motor and sensory nerves, cranial mo-
tor nerves, and autonomic nerves.30 The neurotoxicity of
vincristine is related to the cumulative dose used. Nausea,
vomiting, and myelosuppression are rarely encountered
with vincristine.
The primary dose-limiting toxicity of etoposide is my-
elosuppression.30 Other toxicities include alopecia, nausea,
vomiting, phlebitis, and mucositis.30 However, the most
important toxicity of etoposide is the development of acute
myelogenous leukemia. Etoposide has been known to cause
acute myelogenous leukemia at doses higher than those
used for the chemotherapy of retinoblastoma.31 Smith et al
investigated the relationship between the incidence of sec-
ondary leukemia after etoposide treatment and the cumula-
tive etoposide dose in multiagent chemotherapy regimens
including alkylating agents and doxorubicin.32 They found
that the 6-year risk for secondary leukemia was greater in
the low (⬍1.5 g/m2) cumulative etoposide group than in the
moderate (1.5–2.99 g/m2) and higher (⬎3.0 g/m2) cumula-
tive etoposide dose groups.32 The authors concluded that the
etoposide cumulative dose does not seem to be of primary
importance in determining the risk of secondary leukemia.32
Other factors, including the presence of cytogenetic abnor-
malities, may increase the risk of secondary leukemia after
etoposide use.32 Recently, the occurrence of acute myelog-
enous leukemia in retinoblastoma patients who received
chemotherapy including etoposide has also been reported.
Gombos reported 15 patients with retinoblastoma who de-
veloped acute myelogenous leukemia. Over a third of these
patients received chemotherapy including etoposide as part
of their retinoblastoma treatment (Gombos D. Secondary
AML in patients with retinoblastoma. Presented at: 11th
International Retinoblastoma Symposium, May 19 –22,
2003; Paris, France). Lee et al reported another patient who
developed acute myeloid leukemia after undergoing chemo-
therapy including etoposide as the primary treatment.33
None of our patients developed secondary malignancy re-
lated to radiotherapy or chemotherapy during the limited
follow-up period of this study.
For bilateral retinoblastoma, we currently offer chemore-
duction in all cases, provided that the tumors cannot be
cured with focal treatment methods alone. One potential
advantage of administering chemoreduction in bilateral
cases may be to lower the frequency of developing trilateral
retinoblastoma, as suggested by Shields et al.34 Another
advantage is that the risk of secondary cancers may be
significantly reduced by delaying external beam radiother-
apy until the patient is over 1 year of age.35 For unilateral
retinoblastoma, we offer chemoreduction in eyes with
Reese–Ellsworth groups I to IV disease and enucleation in
eyes with group V disease.36 Regardless of the Reese–
Ellsworth group of the disease, we do not offer external
beam radiotherapy to patients with unilateral retinoblas-
toma. However, there may be occasional changes in this
recommended treatment plan, according to the parents’ dis-
cretion.
In summary, unresponsive and, more commonly, recur-
rent retinoblastoma comprises failure of chemoreduction
therapy. Factors predicting enucleation are older patient
age; advanced disease at baseline, including thicker tumors;
presence of vitreous seeds and subretinal fluid; and recur-
rence of retinal tumors after chemoreduction.
References
1. Shields CL, De Potter P, Himelstein B, et al. Chemoreduction
in the initial management of intraocular retinoblastoma. Arch
Ophthalmol 1996;114:1330 – 8.
2. Murphree AL, Villablanca JG, Deegan WF III, et al. Chemo-
therapy plus local treatment in the management of intraocular
retinoblastoma. Arch Ophthalmol 1996;114:1348 –56.
3. Gallie BL, Budning A, DeBoer G, et al. Chemotherapy with
focal therapy can cure intraocular retinoblastoma without ra-
diotherapy. Arch Ophthalmol 1996;114:1321– 8.
4. Kingston JE, Hungerford JL, Madreperla SA, Plowman PN.
Results of combined chemotherapy and radiotherapy for ad-
vanced intraocular retinoblastoma. Arch Ophthalmol 1996;
114:1339 – 43.
5. Shields CL, Shields JA, Needle M, et al. Combined chemore-
duction and adjuvant treatment for intraocular retinoblastoma.
Ophthalmology 1997;104:2101–11.
6. Greenwald MJ, Strauss LC. Treatment of intraocular retino-
blastoma with carboplatin and etoposide chemotherapy. Oph-
thalmology 1996;103:1989 –97.
7. Levy C, Doz F, Quintana E, et al. Role of chemotherapy alone
or in combination with hyperthermia in the primary treatment
1923
 Ophthalmology Volume 111, Number 10, October 2004
of intraocular retinoblastoma: preliminary results. Br J Oph-
thalmol 1998;82:1154 – 8.
8. Beck MN, Balmer A, Dessing C, et al. First-line chemother-
apy with local treatment can prevent external-beam irradiation
and enucleation in low-stage intraocular retinoblastoma. J Clin
Oncol 2000;18:2881–7.
9. Toma NM, Hungerford JL, Plowman PN, et al. External beam
radiotherapy for retinoblastoma. II. Lens sparing technique.
Br J Ophthalmol 1995;79:112–7.
10. Shields CL, Honavar SG, Shields JA, et al. Factors predictive
of recurrence of retinal tumors, vitreous seeds, and subretinal
seeds following chemoreduction for retinoblastoma. Arch
Ophthalmol 2002;120:460 – 4.
11. Finger PT, Harbour JW, Karcioglu ZA. Risk factors for me-
tastasis in retinoblastoma. Surv Ophthalmol 2002;47:1–16.
12. Shields CL, Shields JA. Recent developments in the manage-
ment of retinoblastoma. J Pediatr Ophthalmol Strabismus
1999;36:8 –18.
13. Hungerford JL, Toma NM, Plowman PN, Kingston JE. Ex-
ternal beam radiotherapy for retinoblastoma: I. Whole eye
technique. Br J Ophthalmol 1995;79:109 –11.
14. Khelfaoui F, Validire P, Auperin A, et al. Histopathologic risk
factors in retinoblastoma. A retrospective study of 172 pa-
tients treated in a single institution. Cancer 1996;77:1206 –13.
15. Kupfer C. Retinoblastoma treated with intravenous nitrogen
mustard. Am J Ophthalmol 1953;36:1721–3.
16. Wilson MW, Rodriguez-Galindo C, Haik BG, et al. Mul-
tiagent chemotherapy as neoadjuvant treatment for multifocal
intraocular retinoblastoma. Ophthalmology 2001;108:2106 –
15.
17. Ellsworth RM. The practical management of retinoblastoma.
Trans Am Ophthalmol Soc 1969;67:463–534.
18. Doz F, Neuenschwander S, Plantaz D, et al. Etoposide and
carboplatin in extraocular retinoblastoma. A study by the
Societe Francaise d’Oncologie Pediatrique. J Clin Oncol
1995;13:902–9.
19. Pratt CB, Fontanesi J, Chenaille P, et al. Chemotherapy for
extraocular retinoblastoma. Pediatr Hematol Oncol 1994;11:
301–9.
20. White L. Chemotherapy in retinoblastoma: current status and
future directions. Am J Pediatr Hematol Oncol 1991;13:189 –
201.
21. Shields CL, Honavar SG, Meadows AT, et al. Chemoreduc-
tion plus focal therapy for retinoblastoma: factors predictive of
need for treatment with external beam radiotherapy or enucle-
ation. Am J Ophthalmol 2002;133:657– 64.
22. Wong FL, Boice JD Jr, Abramson DH, et al. Cancer incidence
1924
after retinoblastoma. Radiation dose and sarcoma risk. JAMA
1997;278:1262–7.
23. Tucker MA, D’Angio GJ, Boice JD Jr, et al. Bone sarcomas
linked to radiotherapy and chemotherapy in children. N Engl
J Med 1987;317:588 –93.
24. Roarty JD, McLean IW, Zimmerman LE. Incidence of second
neoplasms in patients with bilateral retinoblastoma. Ophthal-
mology 1988;95:1583–7.
25. Gündüz K, Shields CL, Shields JA, et al. The outcome of
chemoreduction treatment in patients with Reese-Ellsworth
group V retinoblastoma. Arch Ophthalmol 1998;116:1613–7.
26. Shields CL, Shelil A, Cater J, et al. Development of new
retinoblastomas after 6 cycles of chemoreduction for retino-
blastoma in 162 eyes of 106 consecutive eyes. Arch Ophthal-
mol 2003;121:1571– 6.
27. Bechrakis NE, Bornfeld N, Schueler A, et al. Clinicopatho-
logic features of retinoblastoma after primary chemoreduc-
tion. Arch Ophthalmol 1998;116:887–93.
28. Demirci H, Eagle RC Jr, Shields CL, Shields JA. Histopatho-
logic findings in eyes with retinoblastoma treated only with
chemoreduction. Arch Ophthalmol 2003;121:1125–31.
29. Gombos DS, Kelly A, Coen PG, et al. Retinoblastoma treated
with primary chemotherapy alone: the significance of tumor
size, location, and age. Br J Ophthalmol 2002;86:80 –3.
30. Balis FM, Holcenberg JS, Poplack DG. General principles of
chemotherapy. In: Pizzo PA, Poplack DG, eds. Principles and
Practice of Pediatric Oncology. 3rd ed. Philadelphia: Lippin-
cott-Raven; 1997:215–72.
31. Winick NJ, McKenna RW, Shuster JJ, et al. Secondary acute
myeloid leukemia in children with acute lymphoblastic leuke-
mia treated with etoposide. J Clin Oncol 1993;11:209 –17.
32. Smith MA, Rubinstein L, Anderson JR, et al. Secondary
leukemia or myelodysplastic syndrome after treatment with
epipodophyllotoxins. J Clin Oncol 1999;17:569 –77.
33. Lee V, Hungerford JL, Bunce C, et al. Globe conserving
treatment of the only eye in bilateral retinoblastoma. Br J
Ophthalmol 2003;87:1374 – 80.
34. Shields CL, Meadows AT, Shields JA, et al. Chemoreduction
for retinoblastoma may prevent intracranial neuroblastic ma-
lignancy (trilateral retinoblastoma). Arch Ophthalmol 2001;
119:1269 –72.
35. Abramson DH, Frank CM. Second nonocular tumors in sur-
vivors of bilateral retinoblastoma. A possible age effect on
radiation-related risk. Ophthalmology 1998;105:573–9.
36. Shields CL, Honavar SG, Meadows AT, et al. Chemoreduc-
tion for unilateral retinoblastoma. Arch Ophthalmol 2002;120:
1653– 8.