Article Text

Primary intravenous chemotherapy for group D retinoblastoma: a 13-year retrospective analysis
  1. Ido D Fabian1,2,
  2. Andrew W Stacey1,2,3,
  3. Kenneth P Johnson2,
  4. Zerrin Onadim2,
  5. Tanzina Chowdhury2,4,
  6. Catriona Duncan2,4,
  7. M Ashwin Reddy1,2,
  8. Mandeep S Sagoo1,2,5
  1. 1Ocular Oncology Service, Moorfields Eye Hospital NHS Foundation Trust, London, UK
  2. 2Retinoblastoma Service, Royal London Hospital, London, UK
  3. 3Department of Ophthalmology, University of Washington, Seattle, Washington, USA
  4. 4Paediatric Oncology Department, Great Ormond Street Hospital, London, UK
  5. 5University College London, Institute of Ophthalmology, London, UK
  1. Correspondence to Dr Ido Didi Fabian, Moorfields Eye Hospital, 162 City Road, London EC1V 2PD, UK; didifabian{at}


Background Eye salvage rate for group D retinoblastoma using intravenous chemotherapy (IVC) as a primary modality is <50%. To report on 13 years' experience with the use of primary IVC for group D retinoblastoma.

Methods A retrospective analysis of 64 group D eyes (52 patients) treated with primary IVC, from 2002 to 2014.

Results The median age at presentation was 11.0 months (mean: 18.6, range: 0.6–144.0), 35 (67%) patients had bilateral disease, 38 (73%) germline disease and 8 (15%) cases were familial. In addition to IVC, patients received a median number of three treatments (mean: 6, range: 0–24), including thermotherapy/cryotherapy, plaque radiotherapy, intra-ophthalmic artery chemotherapy (IAC) and/or intravitreous chemotherapy. External beam radiotherapy (EBRT) was used in five eyes, all of which were eventually enucleated. In a median follow-up time of 55 months (mean: 64, range: 14–156), 63% of eyes were salvaged. By the Kaplan-Meier survival analysis, globe salvage rate was 83%, 70%, 59% and 45% at 1, 3, 5 and 10 years, respectively. There were no cases of metastatic spread from intraocular retinoblastoma and no deaths. IVC-related adverse events included febrile neutropenia in 21 (40%) patients and anaphylactic reaction to carboplatin in 2 (4%), all conservatively resolved. Of the patients receiving IAC, third and sixth nerve palsies were documented in two (10%) and one (5%) eyes, respectively.

Conclusions Primary IVC for group D eyes, with adjuvant treatments as required, was found to be a safe and efficient approach, achieving 63% eye salvage rate, no metastatic spread from intraocular retinoblastoma and no deaths. IAC has now replaced EBRT as a successful salvage treatment.

  • Neoplasia

Statistics from


The treatment of intraocular retinoblastoma has advanced over the years and a wide range of options is available. Over the last two decades, one of the major treatment options has been intravenous chemotherapy (IVC) and the classification of intraocular retinoblastoma has changed from the Reese-Ellsworth (R-E),1 which predicted success from radiotherapy, to the International Classification of Retinoblastoma (ICRB),2 which predicts success in the era of chemotherapy. In this schema, the most advanced cases, group E, are usually enucleated and groups A to C demonstrate over 90% chance of success of globe salvage using local treatments with IVC.3 It is the group D eyes that are the challenging cases in terms of eye salvage.

Since external beam radiotherapy (EBRT) was widely abandoned and IVC for intraocular retinoblastoma introduced, the latter has been the main primary treatment for group D eyes in many centres,4 achieving up to 47% eye salvage rate.3 With development and popularisation of intra-ophthalmic artery chemotherapy (IAC),5 ,6 this treatment has found more recent favour as a primary treatment for group D eyes due to its selectivity and reported success in achieving tumour control, even being used in some centres for bilateral cases.7 However, first-line IAC has possible drawbacks, the most prominent of which relates to its selectivity for the eye, potentially allowing systemic spread.8 ,9 In addition, as it is a modality requiring advanced infrastructure and expertise, it is not available in all centres.

In the London Retinoblastoma Service, the primary treatment used for the majority of patients with group D retinoblastoma has been IVC. Selected patients are treated by primary enucleation, but attempts to preserve the globe without compromise to patient survival are being increasingly favoured. In the UK, IAC is used mainly as a salvage treatment.

The goal of this study was to report on the management course and outcomes of group D eyes treated initially in our centre with IVC.


This was a retrospective chart review of consecutive group D retinoblastoma cases, classified according to the ICRB,2 which presented to the Retinoblastoma Service at the Royal London Hospital, from 27 November 2002 to 17 December 2014, and treated initially with IVC. The study was approved by the Barts Health NHS Trust institutional review board (number 6622).

Data retrieved from clinical notes included patients' age, gender, inheritance status (ie, retinoblastoma family history), presenting signs, clinical variables, genetic analysis, treatment specifications and extraocular complications. Two of the authors (IDF and KPJ) examined all RetCam images, including fluorescein angiograms when available, of all patients, analysed the scans and recorded the results.

Cryotherapy, transpupillary thermotherapy (TTT) and ruthenium plaque brachytherapy were performed at the discretion of the treating clinician. For cryotherapy, a cryosurgical system (Mira, Waltham, Massachusetts, USA) was used, and each treatment session included three freeze–thaw cycles, and for TTT (Iris Medical OcuLight SLx, Iridex, Mountain View, California, USA), an 810 nm indirect large laser spot beam was used. During the study period, overall three physicians treated the study patients, using the same treatment protocols. Decision to enucleate an eye or treat with chemotherapy was made after a joint consultation between the treating ophthalmologist/s and medical oncologist/s. Implants used after enucleations were vicryl mesh-coated hydroxyapatite, to which the extraocular muscles were attached, or acrylic, in which cases a myoconjunctival technique was used.10 The presence of high-risk features for systemic spread on histopathological evaluation11 prompted treatment with adjuvant IVC. IVC for intraocular tumours included six courses of vincristine, etoposide and carboplatin (VEC), given via a central venous line, approximately every 3 weeks. In cases in which second-line IVC was indicated, or in cases of high-risk features on histopathology in enucleated eyes, the protocol included ifosfamide, vincristine and doxorubicin (IVAd), given for four to six courses. The technique of intravitreous chemotherapy (IViC) was previously described and the agents used were melphalan and/or topotecan.12 For IAC, the agents used were melphalan, topotecan and/or carboplatin, and the technique has been described by our centre previously.13

Statistical analysis and study definitions

All calculations were performed using Microsoft Excel 2013 software (Microsoft, Redmond, Washington, USA) and SPSS software V.17.0 (SPSS, Chicago, Illinois, USA).

p Values were calculated using the χ2 function and t-test, and survival estimates using the Kaplan-Meier analysis. Early treatment failure was defined as insufficient tumour response (main tumour and/or seeds), requiring change in management, during or immediately after IVC cycles.


Patient characteristics and clinical presentation

There were 104 group D retinoblastoma eyes of 92 patients diagnosed and managed in our service during the study period. Of these, 40 (37.5%) underwent primary enucleation. The remaining 64 eyes of 52 patients (28 (54%) males) were treated by means of primary IVC and these comprise this study cohort. The presentation details and genetic analysis of the study patients are summarised in table 1 and results of clinical examination at presentation in table 2.

Table 1

Demographics, variables at presentation and genetic analysis of 52 patients with group D retinoblastoma (64 eyes) treated by primary intravenous chemotherapy

Table 2

Clinical examination at presentation of 64 group D retinoblastoma eyes (52 patients) treated by primary intravenous chemotherapy

Leukocoria was the most common presenting feature (56%), followed by strabismus (25%). Of the 52 study patients, 34 (65%) presented with bilateral retinoblastoma and the remaining 18 (35%) with unilateral disease. One familial case that presented as unilateral group D retinoblastoma later developed retinal tumours in the non-D eye. Of the bilateral cases, bilateral D was the most common combination (34%).

The median age of diagnosis was 11.0 months (mean: 18.6, range: 0.6–144.0). Eight patients (15%) had positive family history, one (12.5%) of which presented with trilateral retinoblastoma. On genetic analysis, available approximately 3 months after first diagnosis, in 38 patients (73%) a blood RB1 mutation was found.

Multifocal tumours were found in 23 (36%) eyes, a total retinal detachment in 29 (45%) eyes, the optic disc was obscured in 47 (73%) eyes, foveal tumour involvement in 42 (66%) eyes and retinoblastoma seeds, subretinal and/or vitreous, were detected in 55 (86%) eyes, in ≥97% eyes seeds were spread in a diffuse manner.

Management and outcomes

Table 3 summarises the management course of the study patients. Of the bilateral cases, there were six that presented with group D in one eye and group E in the fellow eye. All of these had the group E eye enucleated and histology evaluated prior to commencing IVC.

Table 3

Management summary of 64 group D retinoblastoma eyes (52 patients) treated with primary intravenous chemotherapy

Standard protocol of six cycles of intravenous VEC was used in all patients, except three, whose tumours failed to respond to IVC. Of these, two were switched to IAC, but underwent enucleation after two treatments due to further tumour progression, and one was enucleated after five VEC cycles.

Altogether, early treatment failure occurred in eight (12.5%) eyes, five of which underwent secondary enucleation immediately after completion of the sixth IVC cycle. On statistical analysis, tumour focality and horizontal corneal diameter of the D eye were the only variables associated with early treatment failure (table 4). In detail, 100% of eyes with multifocal retinoblastoma showed good response to primary IVC compared with 81.5% of eyes with unifocal tumours (p=0.043), and larger horizontal corneal diameter was found to associate with higher chances for early treatment failure (p=0.012). Of note, in no case was a difference found in corneal horizontal diameter between eyes of the same patient.

Table 4

Primary intravenous chemotherapy in 64 group D eyes of 52 patients: favourable response versus early treatment failure

Of the 64 eyes treated with IVC, 95% required additional treatments, mainly in the form of TTT and/or cryotherapy. Additional chemotherapy treatments included IAC in 20 (31%) eyes, IViC in 4 (6%) eyes and second-line IVC in 2 (3%). Five (8%) were treated with lens-sparing EBRT; however, all eventually had to undergo secondary enucleation (in four (80%) cases, the underlying reason was uncontrolled tumours and in one (20%) development of iris neovascularisation and non-clearing vitreous haemorrhage).

The median follow-up time for the whole cohort was 55 months (mean: 64, range: 14–156), in which time 40 (63%) eyes were salvaged. The median time from last intervention to last follow-up visit was 24 months (mean: 38, range: 0–156). Kaplan-Meier survival analysis showed an overall globe salvage rate of 83%, 70%, 59% and 45% at 1, 3, 5 and 10 years, respectively. Figure 1 shows the cumulative eye survival according to the Kaplan-Meier analysis, and in addition a subanalysis of eyes that resulted with early treatment failure and those treated with salvage IAC. On statistical analysis, none of the clinical variables emerged as significant risk factors for secondary enucleation. Of the 24 enucleated eyes, 11 (46%) had been treated with salvage IAC, but IAC was not found to be a significant risk factor, nor a protective one, for secondary enucleation (p=0.118). Of the four eyes treated with IViC, only one underwent secondary enucleation.

Figure 1

Kaplan-Meier estimates of globe salvage in (A) the whole cohort, (B) favourable initial response to primary intravenous chemotherapy (IVC) versus early treatment failure and (C) primary IVC versus primary IVC and salvage intra-ophthalmic artery (IAC).

Of the 24 secondary enucleations, 42% were performed during the first year of follow-up, 58% after the first year, 38% after the second year and 25% after the third year. Indications for secondary enucleation included tumour relapse at the optic nerve head, neovascularisation of the iris with vitreous haemorrhage, tumour relapse in the anterior chamber and multiple subretinal/vitreous seeds with/without subtotal/total retinal detachment. Most cases showed a combination of these indications. Six (25%) secondary enucleations were performed before IAC was available and the rest (75%) after. To note, 20 (31%) eyes presented before 2009, that is, first IAC use; however, all were further followed up during the period that IAC was available, and in four (20%) of this early cohort, IAC was used. Of the 22 patients (42%) who underwent secondary enucleation, 4 (8%) who had bilateral retinoblastoma eventually underwent bilateral enucleation.

Of the 24 secondary enucleated eyes, 5 (21%) showed high-risk features on histopathology and were treated with adjuvant IVC because of risk of secondary spread.

Extraocular-related adverse events

IVC-related adverse events included grade 3 febrile neutropenia in 21 (40%) patients and grade 3 anaphylaxis after carboplatin administration in 2 (4%). IAC-related adverse events included third and sixth nerve palsies in two (10%) and one (5%) patients, respectively. In all of these cases, conservative management or observation resulted with complete resolution. In two (10%) cases, IAC was abandoned due to technical difficulties in catheterisation. No complications or related adverse events were recorded after IViC.

No case of metastatic spread from intraocular retinoblastoma and no cases of death occurred during the study period. However, one case with trilateral disease was also treated for pinealoblastoma that had metastasised to the spine, and was alive at the end of this study period.


The last three decades have witnessed huge developments in the management of retinoblastoma, with several new modalities becoming available. This has necessitated a rethink of the classification of intraocular disease from the R-E classification that predicted globe salvage in the era of EBRT1 to the ICRB, which was developed in the era of chemotherapy.2 This study bridges an era in our centre after the establishment of IVC for retinoblastoma,4 ,14–16 when EBRT was used for salvage, to the emergence of IAC for salvage. This was a large cohort of 64 group D eyes followed up for a median time of nearly 5 years with an eye salvage rate of 63%. This is significantly higher than previous similar studies.3 ,17–25 Before IAC, Shields et al3 reported a globe salvage rate of 47% in group D eyes treated by primary IVC, the highest success rate achieved in the pre-IAC era. Of note, Shields et al3 used the Philadelphia version of the IRCB, whereas we used slightly different criteria to define D eyes, according to the Children's Hospital Los Angeles scheme,2 and it was shown that such discrepancies may adversely impact attempts to compare treatment outcomes.26 Nevertheless, this does not fully explain the disparity in salvage rates between the two studies.

A previous study from our centre, from 1995 to 2003, on group D heritable retinoblastoma treated with primary IVC, with median follow-up of 30 months, found that 11/18 eyes (61%) were salvaged, with 5 eyes requiring salvage with EBRT.23 Since then, the role of EBRT has been relegated to being akin to treatment failure, especially with increased risk of second cancers in patients with germline retinoblastoma.27 Interestingly, in this study, salvage EBRT was used in five eyes, all in the early cases, prior to the use of IAC, and in all, treated eyes were eventually enucleated. Presumably, the longer follow-up in this report allowed for more definitive outcomes. These findings, in addition to the known risk associated with EBRT,27 strengthen the approach that EBRT should have a diminished role in treating intraocular retinoblastoma.

IAC and more recently IViC injections are important adjuncts and are important in achieving the relatively high salvage rate in the present cohort. Interestingly, neither was found to be a significant protective factor for eye salvage. As for IViC, this is not surprising, as it was first used in the current cohort in 2014 and only in four eyes. IAC however was first introduced in 2009 and used in 20 eyes in this study. A reasonable explanation would be that IAC was used as a salvage treatment initially in more resistant cases, and 11 of these had to undergo secondary enucleation.

Yousef et al,8 in a systematic review on IAC, reported a total of 67% success rate, when used as second-line therapy, but for all eye groups, clearly different to this study that concentrates on D eyes only. Eye salvage for advanced retinoblastoma in their analysis was found to be 57%, when IAC was used as first and second line and for ICRB groups D and E.8 Shields et al28 reported 67% eye retention in a small case series of six D eyes treated with primary IVC followed by IAC. Abramson et al recently published their results on D eyes treated with first-line IAC (n=47) and reported 85.1% success rate.9 In their study, the median follow-up time was 30 months, in 12 (24%) eyes it was under a year, and in 7 (14%) 2–6 months. In our study, nearly 60% of secondary enucleations were performed after the first year of follow-up and nearly 40% after the second year, so it is possible that later relapse might lead to more enucleations and a lower overall globe salvage. Conversely, it is possible that we may be able to increase our globe retention rates by earlier use of salvage IAC, as our initial cases were treated with this modality as a last resort. Overall then, results from the present and other studies show that IAC, when indicated, is a powerful tool as an adjunct to primary IVC for group D retinoblastoma, enabling retention of a greater number of eyes compared with IVC alone or with the use of additional local consolidation.

We tried to understand whether certain features would subclassify group D into eyes that were more likely to succeed with IVC or would be higher risk for enucleation. None of the variables emerged as significant risk factors for secondary enucleation. This is not surprising, as per definition, D eye tumours hold common features, including tumour size and the presence of retinal detachment and seeding. To note, in two relatively large cohort studies on retinoblastoma eyes treated by primary IVC,20 ,29 group D eyes were analysed as part of a larger cohort containing additional ICRB groups and no subanalysis was performed. In the literature, there are no reports of specific risk factors for secondary enucleation of D eyes treated initially with IVC.

Eight eyes resulted with early treatment failure and eventually secondary enucleation. Salvage IAC was not available for all, but when used for two of these cases, it also failed to achieve a beneficial effect. There is only scant literature on retinoblastoma that does not respond to IVC and that necessitates immediate change in management. In a retrospective analysis by Gunduz et al,30 unresponsive disease was defined as persistence of retinal tumours, vitreous or subretinal seeds after the second IVC cycle, with no appreciable sign of regression. Recurrent disease was defined as regrowth of tumours any time after an initial favourable response. Of 105 eyes (all R-E classification groups), 10 (10%) eyes were unresponsive to treatment. On statistical analysis, however, unresponsive cases were combined with recurrent ones, precluding specific insight into the former subgroup. Interestingly, in this study, eyes with unifocal tumours and those with relatively larger horizontal corneal diameters (in both eyes) were found to be at significant risk for early treatment failure. We have no reasonable explanations for these associations. It should also be mentioned that detection of numerous tumours, usually in cases of genetic disease, is not always possible, especially in D eyes, as in most a total retinal detachment can limit a clear view and ultrasound is only partially helpful in such cases.

In three eyes (5%), no additional treatments to IVC were needed to reach tumour control and in more than half TTT and/or cryotherapy were given after completion of the chemotherapy cycles. There is no consensus on the use of adjuvant therapies. In some reports, routine use of consolidating local modalities was superior to chemotherapy alone,31 but caution is sometimes necessary in case of adverse impact.32 ,33 This argues for adjuvant treatments given only when indicated, as they may also have an adverse impact on the treated eyes.

The extraocular complications and adverse events related to IVC reported in this study are in the spectrum found in the literature.34 This is also the case regarding IAC.13 For both modalities, no serious or life-threatening events were recorded, but twice there was an anaphylactic reaction to systemic carboplatin that responded to conservative management.

Metastatic disease from intraocular retinoblastoma did not occur in any of the cases, an inherent protective feature of IVC over primary IAC.8 The high globe retention rates reported with primary IAC for group D retinoblastoma (85%) are impressive, but there was a 6% metastatic rate in that study.9 In the IVC era, metastatic rates have been low, so the metastatic risks of primary IAC may temper the enthusiasm for this technique in some centres.

Limitations of this study include mainly its retrospective and non-randomised design. Nearly 40% of eyes during the study period underwent primary enucleation (hence were excluded from this analysis of chemotherapy-treated eyes). This could potentially result in selection bias if the more advanced group D eyes were enucleated. Although not the subject of the current report, we compared the primary enucleation with the primary IVC cohort (data not shown), and found that the enucleation cases were selected not for more advanced or complicated disease, but rather older age of presentation and usually unilateral disease. These reservations should be taken into account when interpreting the results of this study. Additional limitations are inherent to the clinical examination of group D retinoblastoma eyes and include, in some cases, difficulty in assessing the number of tumours and the presence of subretinal seeds in case of a large retinal detachment. With regard to insufficient tumour response cases, the definite criterion to be included in this subgroup was change in management during or immediately after IVC. There are several tumour features that can change in respond to treatment and these include, among others, size, colour, vascularisation, degree of calcification, number of tumours and density. It is the combination of all of the above features summed up on clinical examination that dictates whether change of management is required.

In summary, in this large cohort of group D retinoblastoma eyes, primary IVC, followed by additional local therapies, tailored as per clinical need, was found to be an efficient and safe approach, achieving an eye salvage rate of 63% in a median follow-up time of nearly 5 years. In this time period, no case of metastatic spread from intraocular retinoblastoma was recorded and no cases of death. These results are an improvement on those reported previously for group D eyes treated with primary IVC, and are partly a result of IAC, that was added as a salvage modality, replacing EBRT. With the recent addition of IViC, both locally delivered chemotherapy modalities, used as powerful adjuncts to primary IVC, are expected to further improve eye salvage rates and control of the disease, replacing the use of EBRT.


The authors gratefully acknowledge the contribution of the late Dr Judith Kingston, Consultant Paediatric Oncologist for the introduction of intravenous chemotherapy for intraocular retinoblastoma, including in the patients reported herein. The authors would also like to thank Maxine Fraser, Laura Reynolds, Charlotte Clifton and Gemma Melisi for their assistance with this study.


View Abstract


  • Twitter Follow Maddy Reddy @childrens eyes

  • Contributors IDF and AWS had substantial contribution to the design of the work, collection and analysis of the data, drafting the work, final approval of the version published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. KPJ had substantial contribution to the acquisition, collection and analysis of the work, drafting the work, final approval of the version published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. OZ had substantial contribution to the collection and analysis of the data, drafting the work, final approval of the version published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. TC and CD had substantial contribution to the conception of the work, collection and analysis of the data, drafting the work, final approval of the version published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. MSS and MAR had substantial contribution to the design of the work and analysis of the data, significantly drafting the work, final approval of the version published and agrees to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

  • Competing interests None.

  • Ethics approval Barts Health NHS Trust Institutional Review Board (number 6622).

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

Request permissions

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

Linked Articles

  • At a glance
    Keith Barton James Chodosh Jost Jonas