Article Text

Download PDFPDF

Predicting local control of choroidal melanomas following 106Ru plaque brachytherapy
  1. K I Papageorgiou1,
  2. V M L Cohen1,2,
  3. C Bunce3,
  4. M Kinsella1,
  5. J L Hungerford1,2
  1. 1Department of Ophthalmology, St Bartholomew's and the London NHS Trust, West Smithfield, London, UK
  2. 2Department of Ophthalmology, Moorfields Eye Hospital NHS Foundation Trust, London, UK
  3. 3Medical Statistics, Research and Development, Moorfields Eye Hospital NHS Foundation Trust, London, UK
  1. Correspondence to Dr Victoria Cohen, Consultant in Ocular Oncology, Department of Ophthalmology, Barts and the London NHS Trust, West Smithfield, London EC1A 7BE, UK; victoria.cohen{at}bartsandthelondon.nhs.uk

Abstract

Background We evaluated the control rate of choroidal melanomas treated with 106Ru plaque brachytherapy to identify the risk factors associated with local recurrence and lack of response.

Methods A retrospective review of 106Ru plaque brachytherapy for patients with choroidal melanoma treated at St Bartholomew's Hospital, London. Survival analysis was used to assess associations between evaluated age, sex, location, foveal proximity, tumour base and height, presence of lipofuscin and subretinal fluid, apex dose, radiation rate and type of plaque with time to local recurrence. Logistic regression analysis was used to assess to evaluate the association between the same set of variables and lack of tumour response.

Results From January 2002 to December 2006 189 patients were treated. The follow-up ranged from 12 to 78 (median 33) months. None of the patients received adjuvant diode laser thermotherapy. The control rate was 85.7% (14 recurred while 13 did not respond). Of the patients who had local recurrence, univariate survival analysis demonstrated an association with younger patients, foveal proximity, preoperative subfoveal fluid and tumour base >11 mm. Age and foveal proximity remained significant in a Cox multiple variable model (p=0.03). Of the patients who did not respond, logistic regression analysis showed that lack of response was associated with a tumour height >5 mm, confirmed through multiple variable analysis (p=0.027).

Conclusions Tumours that are close to the fovea in young patients appear more likely to show local recurrence. Tumour height >5 mm was the only prognostic factor that determined lack of response. These results may be used to select which tumours require adjuvant therapy.

  • Choroid
  • neoplasia
  • diagnostic tests/investigation
  • treatment surgery

Statistics from Altmetric.com

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.

Introduction

Published data on the growth rate of uveal melanoma and effect on tumour control of different treatment modalities have raised important concerns over the appropriate choice of treatment.1–7 Since the introduction of 106Ru plaque brachytherapy by Lommatzsch and Vollmar in 1964,8 this type of globe-conserving procedure has been widely accepted in many centres throughout Europe.9–11

In 1991 the Collaborative Ocular Melanoma Study demonstrated that for medium-sized choroidal melanoma, patient survival was comparable with enucleation.12 Plaque brachytherapy offers the patient a perceived better quality of life with the potential to preserve vision and the eye.13 However, the application of 106Ru plaque brachytherapy is limited by a steep dose-gradient in tissue, with the upper limit of penetration judged to be approximately 5 mm.14 Therefore, this radioisotope is only suitable for small to medium-sized melanomas. Adjuvant treatment with diode laser thermotherapy may allow better tumour control for more elevated melanomas with the potential to reduced radiation complications.15–19

During the study period, however, diode laser thermotherapy was not used as routine adjuvant therapy for uveal melanoma and was only performed for local recurrence.

The aim of this study was to identify the clinical factors associated with local tumour recurrence and lack of tumour response following treatment of choroidal melanoma with 106Ru plaque brachytherapy in order to provide an evidence-based strategy to select patients who may benefit from adjuvant diode laser thermotherapy.

Patients and methods

A retrospective review was undertaken on patients with posterior uveal melanomas, treated with 106Ru plaque brachytherapy from January 2002 to December 2006, at the Ocular Oncology Unit of St Bartholomew's Hospital in London.

The inclusion criteria were patients with posterior uveal melanomas (choroidal tumours with no ciliary body component) treated solely with 106Ru plaque brachytherapy, and who had at least 1 year of follow-up in the service. Patients were excluded from the study if there was insufficient follow-up information in the medical records or if they had received adjuvant diode laser thermotherapy.

The diagnosis of posterior uveal melanoma was based on the findings of comprehensive ophthalmic examination supplemented by B scan ultrasonography and fundal photography. Tumours were treated with 106Ru brachytherapy, using commercial radioactive plaques (BEBIG, Berlin, Germany). Both indirect ophthalmoscopy and transillumination were used to demarcate tumour margins intraoperatively. Dummy plaques were placed initially to ensure that all tumour margins were within the field of irradiation. All tumours received at least a 2 mm safety margin. Active plaques were subsequently positioned and left in place until the prescribed treatment dose had been delivered to the tumour apex. After irradiation, follow-up visits took place at 1, 3, 6 and 12 months post-brachytherapy, then twice a year until the second post-operative year and annually thereafter. Follow-up evaluation included best corrected visual acuity, ophthalmoscopic and ultrasonographic assessment of local tumour control and brachytherapy-related complications.

All data were collected retrospectively. Demographic data included age, sex and race of the patient. Ocular parameters included laterality (right, left), presence of cataract, presence of co-existing retinal disease such as hypertensive or diabetic retinopathy, or age-related macular degeneration. Tumour characteristics included height, maximal base diameter, tumour location (macular, peripapillary, pre-equatorial and post-equatorial), foveal proximity (distance from the fovea (mm), proximity to the optic nerve (mm)), and the presence of lipofuscin or subretinal and subfoveal fluid. Both foveal and optic nerve distances were either retrieved from the initial assessments as documented in the notes, or when possible calculated from the fundus photograph of the tumour. Radiation parameters included type of 106Ru plaque (20 mm CCB, 20 mm notched COB and 15 mm CCA), prescribed apex dose (Gy) and radiation rate (Gy/h). The follow-up period from the time of treatment to the latest date at which the ocular status was known was noted.

The three outcome measures were tumour control, non-response (tumours who failed to show signs of regression or continued to increase within the first year from treatment) and tumour recurrence (documented expansion of the tumour margin or an increase in tumour height confirmed by sequential ultrasound examination following an initial period of tumour control).

Statistical analysis

Survival analyses (Cox proportion hazards models) were used to examine associations between the hazard of recurrence and the putative risk factors. Only patients who showed an initial response were included in this analysis. All factors were examined in univariate analyses and then stepwise regression was conducted on all variables. Similarly, logistic regression was used to examine associations between putative risk factors and the odds of non-response. All analyses were conducted using the STATA version 10 statistical software (StataCorp, Texas, USA).

Results

Overall, 214 patients diagnosed with posterior uveal melanoma underwent treatment solely with 106Ru plaque brachytherapy from January 2002 to December 2006. Of these patients, six were excluded as they were advised on enucleation but opted for brachytherapy, 13 patients had inadequate follow-up, five had missing documentation on plaque parameters or tumour size, while one patient underwent enucleation within the first year from treatment. Therefore 189 patients were analysed. The patient demographics, tumour characteristics and radiation parameters are summarised in table 1.

Table 1

Demographics of 189 patients with posterior uveal melanoma treated with 106Ru brachytherapy from January 2002 to December 2006, with characteristics of posterior uveal melanomas and plaque radiotherapy parameters

The median follow-up period from the initiation of therapy was 33 months (mean 33.6, SD 15.6, range 12–78  months). The overall control rate was 85.7%. Fourteen of 189 patients recurred and 13 of 189 patients did not respond within the first year from treatment. Time to recurrence ranged between 12 and 71 months with a median of 25.5 months.

Tumour height ranged from 0.8 to 10.3 mm (mean 4, SD 1.7 mm). The median maximal base diameter was 9.5 (IQR 8.4–2.5) mm. Regarding tumour location, 28% of tumours involved the macula, 25% were peripapillary, 41% were pre-equatorial while the 6% of tumours were post-equatorial. Peripapillary tumours were treated with a 20 mm COB-notched plaque. The median radiation rate was 1.4 (IQR 1.08 to 1.76) Gy/h with 65% of choroidal melanomas receiving 110–110 Gy, 19.5% receiving 80 Gy and 15.5% 120 Gy to the tumour apex.

The results of Cox univariate survival analysis on 176 patients (excluding the non-responders from the study group) are summarised in table 2. Univariate evaluation of clinical factors related to the risk of local recurrence revealed an association of decreasing hazard with increasing age (HR per year increase 0.96; 95% CI 0.92 to 0.99; p=0.019) and with increasing distance from the fovea (HR 0.71; 95% CI 0.51 to 0.98; p=0.038). There was a greater hazard of recurrence in patients with pre-treatment subfoveal fluid than without (p=0.03, HR 3.25) and evidence of increasing hazard with increasing basal diameter (p=0.032, HR 1.25). There was some suggestion that men had an increased risk of local recurrence, although this was not statistically significant with these data (p=0.08). Subsequent Cox stepwise analysis, including all variables (except plaque type shape, which was collinear with location), confirmed the association with age (p=0.096, adjusted HR 0.043) and proximity to the fovea (p=0.04, adjusted HR 0.069), as seen in table 2.

Table 2

Cox univariate hazard model on 176 patients undergoing 106Ru brachytherapy (survival analysis excluding the non-responders and evaluation of significant clinical factors associated with time to recurrence) and Cox multiple variable analysis on 176 patients (excluding the non-responders from the study group) with significant clinical factors related to the time of recurrence

Univariate logistic regression (table 3) looking at the same set of factors associated with non-response showed evidence of association with tumour height (p=0.006, OR 1.55). Successive multiple variable logistic regression (table 3), confirmed that the only factor associated with the likelihood of non-response was the height of the tumour (p=0.027, OR 1.52).

Table 3

Univariate logistic regression analysis on 189 patients undergoing 106Ru brachytherapy with evaluation of clinical factors associated with lack of initial response to treatment, and multiple variable regression analysis with significant clinical factors associated with lack of initial response to treatment

Figure 1 shows the temporal relationship of tumour recurrence. Even after 60 months of follow-up, the risk of local recurrence was still present. The Kaplan–Meier curve shows a gradual decline in local control with time, as there appears to be no plateau. Figure 2 demonstrates that tumours more than 4 mm from the fovea had a reduced risk of local recurrence.

Figure 1

Kaplan–Meier survival estimate on 176 patients with posterior uveal melanoma and showing the temporal relationship of tumour recurrence. Even after 60 months (5 years) of follow-up, the risk of local recurrence was still present. There is a gradual decline in local control with time, but there is no plateau.

Figure 2

Kaplan–Meier estimate on 176 patients in relation to foveal distance. Cox multivariate analysis showed increased risk of recurrence with foveal proximity (p=0.03).

Discussion

Over the past 10 years 106Ru has been the sole radioisotope used for plaque brachytherapy in the London ocular oncology service. This study was designed to report the success of local tumour control using 106Ru plaque brachytherapy alone without adjuvant diode laser thermotherapy. The selection of a particular radioisotope may influence local control rate. In 1999, we reported our experience with 106Ru, 125I and proton beam radiotherapy. Five-year melanoma recurrence rates were 11%, 4% and 5%, respectively.20 In this study we report a 5-year melanoma recurrence rate of 14% for 106Ru plaque brachytherapy. This is similar to the 5-year recurrence rate published by Karlsson et al for 60Co plaque brachytherapy.21 However, Lommatzsch et al4 reported a higher recurrence rate of 37% at 15 years using 106Ru plaque brachytherapy. This recurrence rate may be reflected by the longer period of follow-up in this paper. In addition, our study demonstrates that the Kaplan–Meier curve for local tumour control does not plateau within the first 5 years of treatment and therefore long-term follow-up is required in these patients.

Adjuvant diode laser thermotherapy can be used to improve local recurrence rates significantly. Damato et al22 reported a 5-year recurrence rate of only 2.1% using 106Ru plaque brachytherapy combined with diode laser when appropriate. Irradiated tumours received adjuvant diode laser thermotherapy to the posterior tumour margin. However, this low rate of recurrence may be related to case selection, as many more tumours are selected for proton beam radiotherapy by this centre. Similarly, Shields and associates15 were able to reduce the 5-year recurrence rate to 3% by combining 125I plaque brachytherapy with adjuvant diode laser thermotherapy. However, comparing local control rates between different institutions may not be reliable, as there is considerable inter-study variation and case mix differences. An important outcome of this study was the development of the ability to select which tumours were more likely to show local recurrence.

In our series, Cox multivariate analysis of prognostic factors for local recurrence identified a higher risk of recurrence in tumours close to the fovea (p=0.04). This increased risk of recurrence for juxtafoveal tumours was not due to plaque placement to spare vision, as during the period of the study none of the macular plaques were decentred. Some juxtafoveal tumours were treated with lower radiation rates in order to preserve vision and this could be an explanation for the high risk of recurrence in this group. However, multivariate analysis demonstrated that the rate of delivery of radiation did not affect the risk of local recurrence. It is most likely that the increased blood supply to this area allows tumour cells to survive even the face of radiation. It is well established that an effective vasculature supply is an independent prognostic factor for tumour growth23 Several clinical trials using colour Doppler imaging in patients with choroidal melanoma have identified an increase in blood flow velocity in untreated tumours.24 25

Surprisingly in this series, proximity to the optic disc did not emerge as a significant predictive factor for local recurrence. Twenty-five per cent of melanomas in this study were peripapillary and therefore notched plaque was successful without diode laser thermotherapy.

Increasing tumour height was identified as a risk factor for lack of response to 106Ru plaque brachytherapy. Tumour height is a predictive factor for local relapse as well as for loss of visual acuity.17 26 27 In our series the median tumour height was 3.7 mm. However, 27% of tumours had a thickness of greater than 5 mm and the median tumour height in the non-response group was 5.6 mm, while in the recurrence group it was 4.9 mm. In our retrospective series reported in 1999, treatment with 106Ru plaque brachytherapy was found to be associated with a higher rate of local recurrence than either the 125I or proton beam radiation therapy.20 However, fewer side effects and better visual results were seen in the 106Ru plaque brachytherapy-treated patients. Therefore the use of 106Ru was continued, although we attempted to improve the local tumour control rate by both using wider treatment margins (through a larger 20 mm plaque) and by varying the radiation dose according to tumour thickness (120 Gy calculated at 5 mm depth).

The upper limit of 106Ru plaque brachytherapy is assumed to be a tumour thickness of 5 mm as β radiation is limited by the steep dose-gradient in tissue.14 This limitation can be overcome by adopting ‘sandwich therapy’, where plaque brachytherapy is combined with up to three sessions of adjuvant diode laser thermotherapy.28 Several centres have reported improved control rates with this treatment approach.16 17

There are several limitations that should be realised in this retrospective series, the main one being the follow-up interval for local recurrence and the risk of bias arising from the retrospective collection of data. We have demonstrated that recurrence may occur more than 5 years after treatment. The data regarding lack of initial tumour response and its association with increasing tumour height is therefore most valid. We have identified a reduced 5-year local control rate within our institution for 106Ru plaque brachytherapy: 14% in 2009 versus 11% in 1999 (statistically significant difference, Fisher's exact=0.404, one-sided Fisher's exact=0.215). This may have occurred because thicker tumours were selected for 106Ru plaque brachytherapy since 125I plaques were abandoned. An important observation from this study is that 106Ru brachytherapy as a sole modality of treatment should not be used for tumours exceeding the 5 mm in height. Currently only patients with a tumour height of less than 5 mm are selected for 106Ru plaque brachytherapy.29 Plaque brachytherapy is combined with adjuvant diode laser thermotherapy for tumours greater than 5 mm and up to 7 mm in height. Diode laser thermotherapy may also be considered for juxtafoveal tumours although this approach is likely to result in early visual loss or a paracentral scotoma.

References

Footnotes

  • Competing interests None.

  • Ethics approval The study was approved by the St Bartholomew's Hospital Clinical Governance Department and the Research Government Committee of Moorfields Eye Hospital (Reference no. PAPK100).

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