Article Text
PDF
Abstract
Background/aims We have previously introduced electroretinography (ERG) as a proxy for visual function to monitor for retinal toxicity due to intra-arterial chemotherapy for retinoblastoma in young children. In this paper, we report ERG results for patients with retinoblastoma receiving initial treatment with systemic chemotherapy.
Methods Inclusion criteria were patients presenting with retinoblastoma at <3 months of age or <6.0 kg in weight, with large tumours not amenable to local laser treatment, cryotherapy or plaque brachytherapy. Patients received intravenous carboplatin 18.7 mg/kg every 3–5 weeks, contingent on recovery of blood counts, until they had grown sufficiently to receive intra-arterial chemotherapy. ERG was performed during examination under anaesthesia at monthly intervals, using contact lens electrodes and a hand-held ganzfeld stimulator. 30-Hertz flicker responses are reported.
Results Four patients were treated for bilateral retinoblastoma. All eyes responded well to systemic chemotherapy. 30-Hertz flicker ERGs improved during treatment in all eyes, significantly in six of eight eyes, and at least in one eye of each patient.
Conclusion Effective systemic chemotherapy for retinoblastoma in children who are too small for intra-arterial chemotherapy is compatible with significant increases in ERG amplitudes, even in eyes presenting with extinguished ERGs. ERG signals may increase independent of resolution of retinal detachment.
- Retinoblastoma
- chemotherapy
- electroretinography
- carboplatin
- cancer
- electrophysiology
- telemedicine
- psychophysics
- optics and refraction
- retina
- macula
- neoplasia
- iris
- treatment other
- conjunctiva
- ciliary body
- treatment lasers
Statistics from Altmetric.com
- Retinoblastoma
- chemotherapy
- electroretinography
- carboplatin
- cancer
- electrophysiology
- telemedicine
- psychophysics
- optics and refraction
- retina
- macula
- neoplasia
- iris
- treatment other
- conjunctiva
- ciliary body
- treatment lasers
Introduction
Management of patients with advanced retinoblastoma has historically emphasised preservation of life and, secondarily, preservation of the eye, if necessary at the expense of vision. Indeed, reports of visual outcomes in eyes treated for retinoblastoma have been rare, reflecting the priorities in the underlying management.1 2 Recently, we have introduced routine monitoring of retinal function by electroretinography (ERG) as a proxy for measurements of visual acuity in eyes treated for retinoblastoma, primarily by intra-arterial chemotherapy infusion.3 4 ERG monitoring is particularly appropriate in this setting, where retinal toxicity of chemotherapeutic agents is of primary concern as systemic toxicity is rare. Furthermore, ERG testing is particularly practical, as most of our patients are pre-verbal, precluding standard subjective acuity testing.
Intra-arterial chemotherapy has substantially expanded the spectrum of cases eligible for eye- or vision-sparing treatment in lieu of primary enucleation. Nevertheless, primary systemic chemotherapy, administered by intravenous infusion, remains an important therapeutic option in situations where intra-arterial treatment is unavailable or impractical, as in the very youngest and smallest patients. In this paper, we provide the first report of ERG monitoring of retinal function after systemic chemotherapy for retinoblastoma.
Methods
Patients were selected for initial systemic chemotherapy from those referred to our centre for primary treatment of retinoblastoma. Selection criteria included large tumours inappropriate for management with only local measures such as trans-pupillary thermotherapy, laser treatment or cryotherapy, and age <3 months or weight <6.0 kg, which, in our experience, frequently precludes successful catheterisation of the ophthalmic artery for intra-arterial infusion treatment.
Each patient received single-agent carboplatin (18.7 mg/kg/dose intravenously over 1 h) in each cycle of chemotherapy. Treatment cycles and examination under anaesthesia were repeated every 3–5 weeks depending on adequate blood count (absolute neutrophil count and platelet count) recovery.5 All patients subsequently received intra-arterial chemotherapy for consolidation of tumour control. The details of the choice of intra-arterial chemotherapy regimens and the ERGs after the initial systemic phase of treatment are beyond the scope of this report. The intra-arterial treatment and subsequent outcomes are summarised below for narrative completeness.
ERG recordings were obtained at baseline and at subsequent examinations under anaesthesia as described previously.3 Briefly, ERGs were obtained after indirect ophthalmoscopy using a hand-held ganzfeld stimulator and ERG-jet contact lens electrodes, according to a modified protocol based on ISCEV (International Society for Clinical Electrophysiology of Vision) standards. In contrast to the ISCEV standard, photopic recordings were obtained first to take advantage of the light adaptation afforded by the previous indirect ophthalmoscopy. Single-flash light-adapted 3.0 responses were elicited from the right eye and then averaged together in groups of 10. Light-adapted 3.0 flicker responses were obtained with a stimulus frequency of 30 Hz, singly, and then averaged together in groups of 10. (Averaged waveforms were used for all analyses and are shown below.) The electrode was then switched to the left eye, and the stimulus sequence was repeated. Room lights were then extinguished. Dark adaptation lasted 5 min (in contrast to the 20-minute ISCEV standard) as dictated by the need to minimise the total duration of general anaesthesia. In our experience, the 5-minute adaptation provides about 80–90% of the adaptation effect obtained by the full 20-minute adaptation period. Dark-adapted 0.01 and 3.0 responses were obtained from the left eye, then the electrode was transferred to the right eye and the dark-adapted 0.01 and 3.0 responses were obtained. Standard responses and oscillatory potentials were recorded simultaneously.
The response to light-adapted 3.0 flicker stimulation was adopted as a proxy for the complete set of ERG responses, as photopic and scotopic responses were highly correlated, and the 30-hertz stimulus facilitated signal detection in severely impaired retinas. A change in the 30-hertz response amplitude of 25 μV was considered clinically significant, based on analysis of ERGs during examination under anaesthesia of normal eyes (as described below). B-wave implicit times obtained during these recordings are typically prolonged by the effects of the general anaesthetic agent. As this effect varies unpredictably, due to the variations in the level of anaesthesia, we are unable to analyse the effects of chemotherapy on the ERG-implicit times.
Results
The 25 μV criterion for significant change in the ERG was based on an analysis of test–retest variability for the response to light-adapted 3.0 flicker stimulation obtained from the clinically normal fellow eyes of patients with unilateral retinoblastoma or Coats' disease seen in our clinic. The average SD between recording sessions for the mean response amplitude in 11 patients (each with ERGs obtained at between 2 and 10 sessions, for a total of 58 sessions) was 24.7 μV.
We report ERG findings in the initial four patients who underwent primary intravenous chemotherapy for bilateral retinoblastoma in our clinic, for whom baseline and follow-up ERG data are available:
Case 1. A 10-day-old infant presented with bilateral retinoblastoma diagnosed during the newborn examination. The right eye was classified as Reese Ellsworth group Va, International Classification Group E; the left eye was classified as Reese Ellsworth group Ia, International Classification Group B. The child was treated with four cycles of single-agent carboplatin and focal laser treatment until the age of 4 months, when one cycle of intra-arterial chemotherapy with melphalan was given to the right eye. Tumours and seeding showed a response to treatment with no recurrent disease at 6 months follow-up.
Case 2. A 2-week-old infant presented with unilateral leukocoria and was referred for treatment of retinoblastoma. The right eye was classified as Reese Ellsworth Group IIIa, International Classification Group D. A tumour was later found in the left eye, which was classified as Reese Ellsworth Group IIIa, International Classification Group B. The child was treated with two cycles of single-agent intravenous carboplatin and focal cryotherapy and laser treatment until the age of 3 months when intra-arterial chemotherapy with melphalan, topotecan and carboplatin was begun. The patient received a total of five cycles to the right eye, with additional laser and/or cryotherapy after each cycle. The left eye required no intra-arterial treatment. Tumours and subretinal seeding showed a response to treatment with no recurrent disease at 6 months follow-up. A large retinal detachment in the right eye resolved with treatment.
Case 3. A 3-day-old infant presented with familial bilateral retinoblastoma diagnosed during the newborn examination. The right eye was classified as Reese Ellsworth Group Ia, International Classification Group B; the left eye was classified as Reese Ellsworth Group IIb, International Classification Group C. The child was treated with two cycles of single-agent carboplatin and focal cryotherapy and laser treatment until the age of 3 months when intra-arterial chemotherapy with melphalan and topotecan was given for a total of two cycles to the right eye, and three cycles of melphalan and carboplatin to the left eye. Tumours and subretinal seeding showed a response to treatment with no recurrent disease at 8 months follow-up.
Case 4. A 2-week-old infant presented with familial bilateral retinoblastoma diagnosed during paediatric examination and was referred for treatment. The right eye was classified as Reese Ellsworth Group IIIa, International Classification Group B; the left eye was classified as Reese Ellsworth Group Vb, International Classification Group D. The child was treated with three cycles of single-agent carboplatin and focal cryotherapy and laser treatment until the age of 4 months when intra-arterial chemotherapy with melphalan was given for a total of one cycle to the left eye, with additional cryotherapy and focal laser. Tumours and seeding showed a response to treatment with no recurrent disease at 6 months follow-up.
ERG amplitudes at baseline, and as of the last recordings obtained during intravenous chemotherapy treatment, are shown in table 1. ERGs improved in all eight eyes, significantly in six eyes (at least in one eye of each patient). The time course of the ERG amplitudes for each patient, together with the timing of the chemotherapy cycles, is shown in figure 1. Representative ERG waveforms in response to the 30-hertz photopic flicker stimulation are shown for Case 2 (figure 2). Fundus photographs for Case 2 before and after intravenous chemotherapy are shown in figure 3.
Electoretinography (ERG) outcomes after intravenous chemotherapy
30-Hertz flicker ERG amplitudes in μV are plotted against age for the duration of systemic chemotherapy treatment. Squares, right eyes; triangles, left eyes; diamonds, chemotherapy cycles.
30-Hertz flicker ERG waveforms for Case 2. Top, baseline ERGs; bottom, ERGs following intravenous carboplatin treatment. Right eye, extinguished at baseline, has recovered to the lower limit of normal amplitude; responses of the left eye have more than doubled. The vertical tick marks indicate the timing of the 30-hertz stimulus flashes.
Fundus photographs for Case 2 before (top row) and after (bottom row) intravenous chemotherapy. Note type-1 regression pattern of large tumour in the right eye, with resolution of retinal detachment. Small peripheral tumour in the left eye (arrow) was treated with cryotherapy, resulting in a small retinal scar.
Discussion
We report the results of ERG monitoring of retinal function in four cases of retinoblastoma during primary intravenous treatment. All four cases were bilateral, as might be expected for most neonate patients presenting with retinoblastoma. ERGs improved in all eight eyes, significantly in six eyes (at least in one eye of each patient).
This small cohort demonstrates that retinal function can improve after initial treatment of retinoblastoma with systemic chemotherapy, quite apart from improvements that may be associated with resolution of retinal detachment. As the ERG normally increases with age during the first several months of life (reaching adult amplitudes shortly before 1 year of age), it is impossible to know with certainty the extent to which the increases in ERG amplitude observed in this cohort of patients might be related to the suppression of tumours by chemotherapy.
However, it should be noted that the normal increases in the amplitude of the ERG response to the 30-hertz flicker with age are seldom as dramatic as those seen here. Compare, for example, the logistic curve for the normal maturation of the 30-hertz flicker ERG based on the parameters reported by Westall et al,6 shown in figure 4. This comparison suggests at least the possibility that the presence of active retinoblastoma tumours may have a deleterious effect on ERG responses, even in the absence of retinal detachment, which may be ameliorated with successful chemotherapy.
Model logistic growth curve for the 30-hertz flicker ERG response for the first several months of life, plotted on the same scale as the time courses shown in figure 1, after Westall et al6: Y/Ymax=Agen/(Agen+Cn), where Y is the flicker ERG amplitude, Age is the patient's age in months, Ymax is the maximal ERG amplitude in μV, C is the age-semisaturation constant (in months) and n is a dimensionless exponent. Parameter values are Ymax=89.6 μV, C=5.12 months, n=0.41.
Apart from serous or mechanical retinal detachment, it is unclear how the presence of active retinoblastoma might suppress the ERG. We are not in a position to comment on the possible effects of tumour metabolism, secretion of tissue factors or autoimmune responses to tumours that might act on the physiology of retinal neurons or Mueller cells. To the best of our knowledge, there are no reports of visual abnormalities or abnormal ERGs in the normal fellow eyes of patients with unilateral retinoblastoma who carry the germ-line RB1 mutation. We have reported one such unilateral retinoblastoma patient with normal ERGs in the fellow eye previously.7
This capacity for improvement in the retinal function supports continued efforts to develop treatment strategies that will preserve globes, even with extensive retinoblastoma, as they often carry significant potential for useful vision, and because it may not be possible to predict, at presentation, which eye may ultimately be relied upon for seeing.
Footnotes
Funding This work was supported by the Fund for Ophthalmic Knowledge, Inc. grant no. FFOK 18-11, and by Research to Prevent Blindness. Partial funding was also provided by the Ophthalmic Oncology Center, MSKCC, and departmental support was provided by the Mount Sinai Department of Ophthalmology.
Competing interests None.
Ethics approval Ethics approval was provided by Memorial Sloan-Kettering Cancer Center Institutional Review Board.
Provenance and peer review Not commissioned; externally peer reviewed.