Plot of the ratio of histopathologic measure to clinical measure at the time of randomization for the longest basal diameter vs height. Each point represents tumor measurements in an eye enucleated because of local treatment failure. Open circles represent 2 tumors with extrascleral extension.
Avery RB, Diener-West M, Reynolds SM, Grossniklaus HE, Green WR, Albert DM. Histopathologic Characteristics of Choroidal Melanoma in Eyes Enucleated After Iodine 125 Brachytherapy in the Collaborative Ocular Melanoma Study. Arch Ophthalmol. 2008;126(2):207-212. doi:10.1001/archophthalmol.2007.50
To describe the histopathologic findings in eyes with uveal melanoma that had secondary enucleation after failed brachytherapy plaque treatment.
Histopathologic findings in eyes that had secondary enucleation after plaque radiation therapy in the Collaborative Ocular Melanoma Study (COMS) were reported on a standardized data form. The findings were compared with eyes that had primary enucleation for uveal melanoma.
Seventy-five eyes that had secondary enucleation were studied. Compared with primary enucleations, tumors in the irradiated eyes had lower mitotic activity, a smaller proportion of histologically intact tumor, more inflammation, more fibrosis, and more vascular damage within the tumor. In addition, compared with primary enucleations, eyes previously irradiated had a higher frequency of retinal invasion by the tumor and greater damage to the retinal vasculature, consistent with radiation retinopathy; neovascularization of the iris; and vitreous hemorrhage. Tumor growth or extrascleral extension was confirmed histopathologically in 25 of 42 eyes (60%) enucleated because of a reported failure of local control.
Eyes with secondary enucleation after brachytherapy differ histopathologically from eyes with primary enucleation for uveal melanoma. These histopathologic differences may be due to the effects of radiation, tissue conditions related to plaque failure, and, in some cases, tumor growth. In 40% of eyes enucleated because of suspected failure of local control, increased tumor size could not be histologically confirmed.
The Collaborative Ocular Melanoma Study (COMS) Group showed no difference in survival between patients randomized to primary enucleation compared with iodine 125 plaque brachytherapy.1,2 A potentially important subgroup of the radiotherapy arm are those patients who had secondary enucleation. Of the 650 COMS patients who received brachytherapy, the Kaplan-Meier estimate of secondary enucleation within 5 years of treatment was 12.5%.3 The clinical features of these “plaque failures” have been previously reported.3 Enucleations within the first 3 years after brachytherapy are likely due to treatment failure, and this is weakly associated with reduced survival after adjusting for other baseline variables (adjusted risk ratio, 1.5; 95% confidence interval, 0.94-2.52; P = .08).3 This study describes the histopathologic findings in eyes with failed plaques as compared with a control group with primary enucleation.
Histopathologic analysis of eyes with secondary enucleation can identify changes in eyes following radiation and/or features associated with plaque failure, particularly when compared with eyes enucleated for failure of local control compared with other reasons (blind, painful eye). Of interest in this analysis was consistency of histopathologic findings with the clinically stated reason for enucleation. We were able to evaluate histopathologic changes following brachytherapy and to identify microscopic features that could predispose brachytherapy plaques to fail as well as to address other questions of interest.
A summary of the COMS protocol has been published previously.4 Histopathologic processing and evaluation were performed as described in the COMS Manual of Procedures and elsewhere.5,6 Members of the Pathology Review Committee (D.M.A., H.E.G., and W.R.G.) performed independent evaluations and recorded microscopic features of the eyes and tumors on standard data collection forms.7 Differences among the 3 independent reviewers were adjudicated according to an established protocol.
In eyes that were randomized to iodine 125 brachytherapy, treatment failures were identified during follow-up examination by the following ophthalmologist. Identification of treatment failure required completion of a separate data collection form documenting the failure of local control. Treatment failure was defined as tumor growth, extrascleral extension, or local recurrence according to preestablished criteria. Whenever an eye assigned to brachytherapy subsequently was enucleated for any reason (including treatment failure), an additional report regarding the enucleation surgery was submitted to the COMS Coordinating Center. Eyes enucleated during follow-up were processed and evaluated histopathologically according to the COMS protocol.5
The histopathologic characteristics of the failed-plaque eyes were then compared with individually matched controls from the overall group of eyes assigned to primary enucleation and removed. Eyes were matched exactly for tumor cell type (Armed Forces Institute of Pathology–modified Callendar classification8) and as closely as possible for tumor height, proximity to foveal avascular zone, and baseline visual acuity.
The failed-plaque eyes and their matched control eyes were then reevaluated for additional histopathologic parameters that were not included in the original protocol: tumor shape, amount of viable tumor remaining, fibrosis, tumor blood vessel damage, and retinal damage. These additional features were independently evaluated by 2 examiners (R.B.A. and D.M.A.) who were masked to the reason for enucleation, using a supplemental standard evaluation form. The reason for enucleation of the failed-plaque eyes was obtained from the enucleation surgery report form submitted to the COMS Coordinating Center at the time of the secondary enucleation.
The relative frequency of each histopathologic parameter within the plaque-failure group was compared with that of the overall group of primary enucleations using the χ2 statistic or Fisher exact test. The matched analysis of plaque failures with matched controls was performed using simple conditional logistic regression to account for the individual matching. The failed-plaque group was further subdivided into 2 groups: (1) those eyes enucleated for a failure of local control vs (2) all other reasons. The characteristics of these 2 subgroups were compared using the χ2 statistic or Fisher exact test.
Specimens from secondary enucleations that were cited as “failures of local control” were further analyzed microscopically to confirm evidence of tumor growth. Baseline dimensions were obtained from clinical measures of tumor height and largest basal diameter at the time of enrollment and randomization (before plaque therapy). Corresponding postenucleation measures were made microscopically from the processed specimens. The ratio of the postenucleation histopathologic measure to the baseline clinical measure was calculated separately for tumor height and largest basal diameter. A tumor that measured larger after enucleation than at baseline would have a ratio with a value greater than 1 for either height or diameter.
All analyses were based on data received at the COMS Coordinating Center by September 30, 2000. At that time, all patients had been enrolled in the COMS for at least 2 years and most patients had been followed up for at least 5 years.
The 650 patients treated with brachytherapy were followed up for a median of 67 months. Of these, 85 forms reporting secondary enucleation of a plaqued eye were received at the COMS Coordinating Center. Seventy-eight eyes were available for histopathologic analysis. Two of those eyes were excluded from this study because of misdiagnosis (melanocytoma), and an additional eye was not used because extensive necrosis precluded determination of tumor cell type, leaving 75 plaque failures for analysis.
The histopathologic characteristics of melanomas in the failed-plaque group were compared with all primary enucleations (Table 1). Irradiated lesions were more likely to show rupture of the Bruch membrane (93% vs 81%), invasion of the retina (70% vs 30%), tumor cells in the vitreous (37% vs 13%), and scleral extension (68% vs 60%). Irradiated lesions showed less mitotic activity (19% vs 40% with ≥ 2 total mitoses in 40 high-power field) and more inflammation (68% vs 29% with moderate or marked), macrophages (73% vs 60% with moderate or marked), and hemorrhage (63% vs 37%). In addition, the failed-plaque group exhibited more pigmentation and necrosis. There was a higher proportion with prominent tumor vessels (75% vs 61%) and neovascularization of the iris (NVI) (40% vs 2%) in the failed-plaque group. (All P values for comparisons mentioned were ≤ .02.)
The histopathologic characteristics of melanomas in the failed-plaque group were compared with matched-control primary enucleations (Table 2). Compared with matched controls, the failed-plaque tumors were more likely to be juxtapapillary (11%
vs 1%). Irradiated lesions had a smaller proportion of histologically intact tumor (ie, suggestive of viable tumor cells), less mitotic activity, and were less likely to be dome shaped or oval (without extension). They were more likely to exhibit invasion of the retina by the tumor, inflammation, fibrosis, balloon cells, and damage to blood vessels, specifically thickening and thrombosis. Irradiated tumors also were more likely to show extravasation of blood. (P values ≤ .05 for all comparisons mentioned.)
The extratumoral findings in eyes with failed plaques were compared with matched-control primary enucleations (Table 3). Eyes with failed plaques were more likely to show loss of bipolar and ganglion cells, damage to the retinal vasculature, retinal neovascularization, and intraretinal and vitreous hemorrhage. Neovascularization of the iris was found in 39% of the irradiated eyes but in none of the matched controls.
The reasons cited for secondary enucleation are given in Table 4. Forty-two eyes were removed for a perceived failure of local control. Thirty-three eyes in this study were enucleated because of poor vision or pain, without a reported concern of tumor growth.
Failed-plaque eyes were further subdivided based on the reasons for enucleation that were reported to the COMS Coordinating Center. The histopathologic features of those enucleated because of a failure of local control were compared with those enucleated strictly for pain and/or low vision (Table 5). Eyes enucleated for local control failure had a greater proportion of presumed viable tumor in the lesion and more had retinal detachment. Eyes removed because of pain or low vision had more fibrosis within the lesion and more inflammation in ocular structures outside the lesion. The retinal vasculature was more likely to show thrombosis and neovascularization, signs of radiation retinopathy. Vitreous hemorrhage and NVI were more common in eyes enucleated because of pain or low vision only. Eyes enucleated for local control failure had tumors with larger mean baseline diameters and apical height. (All comparisons yielded P values ≤ .05.)
We attempted to confirm tumor growth in eyes enucleated because of concerns of local failure. Of the 42 eyes that were enucleated because of concerns of local failure, 23 had histopathologic measures that were greater in height or had larger basal diameters than their clinical measures at the time of randomization (Table 6). Two more demonstrated histopathologic evidence of extrascleral extension. Neither of those 2 eyes had evidence of enlarged tumor dimensions. In 17 of 42 eyes, we could not confirm enlarged tumor dimensions or scleral extension. The Figure displays a comparison of histopathologic vs clinical tumor dimensions for the 42 eyes.
Compared with either all primary enucleations or matched controls, tumors in eyes undergoing secondary enucleations had less mitotic activity; this finding is consistent with prior studies of postradiation uveal melanomas6,9- 11 and probably represents direct cytotoxicty.12 Alternatively, if the lower mitotic activity represents a slower-growing tumor at baseline, those tumors may have been relatively resistant to radiation, thus leading to continued tumor growth and secondary enucleation.
Following radiation, vascular damage and hemorrhage were found in tumors in enucleated eyes of COMS patients. These changes are likely due to postradiation vasculopathy, as has been described in other studies.11,12 In this study, 55% of eyes also demonstrated vascular abnormalities suggestive of radiation retinopathy in the adjacent retina. This finding is consistent with the findings of Gündüz et al,13 who reported a 42% incidence of radiation retinopathy following plaque radiotherapy for posterior uveal melanoma. Neovascularization of the iris (39% of failed plaques) probably represents radiation-induced ischemia.14
The host response to radiation and/or its effects was evidenced by more inflammation and more fibrosis in the irradiated tumors. Increased pigmentation and increased numbers of plasma cells, neutrophils, and macrophages were seen in failed plaques compared with all primary enucleations. However, there were no statistically significant increases when failed plaques were compared with controls matched for cell type. This finding supports previous reports that associated pigmentation and macrophages with cell type.6
Secondary enucleations also occurred more often in eyes with tumor near the optic nerve, likely because of radiation optic neuropathy,15 despite study exclusion criteria excluding patients in whom the tumor was within 2 mm of the nerve head at diagnosis. Compared with all primary enucleations, secondary enucleations more often had a rupture in the Bruch membrane, invasion of the retina by the tumor, and tumor cells in the vitreous. Compared with matched controls, secondary enucleations were more likely to show invasion of the retina. We cannot distinguish whether these findings are due to more aggressive tumors at baseline that continued to grow after radiation or to an effect of the radiation that, at least in some cases, resulted in more invasive tumor activity.
We compared the histopathologic findings in eyes enucleated because of concerns of local control (ie, tumor growth) with those in which enucleation was strictly for pain or vision loss. Eyes enucleated for failure of local control had a greater proportion of presumed viable tumor. Eyes removed because of pain or low vision had more thrombosis, neovascularization, NVI, and vitreous hemorrhage. Vascular damage to structures outside the tumor should not be discounted when counseling patients about the effects of radiation therapy.
The rate of serous retinal detachment was 92% in the failed-plaque group, compared with 99% in matched-control primary enucleations. The total cohort of COMS primary enucleations (medium and large melanomas) also reported a 99% prevalence of retinal detachment.6 Within the subgroup of failed plaques, eyes enucleated because of a reported failure of local control had a higher likelihood of retinal detachment. It is possible that coincident retinal detachment causes an overestimation of tumor growth, or at least a level of discomfort in evaluating tumor growth that encourages secondary enucleation.
The decision to enucleate involved many factors, of which perceived tumor growth was only one. Nevertheless, we tried to quantify some tumor growth by comparing histopathologic measurements on irradiated tumors with baseline measurements. With this analysis, we could confirm tumor growth (in either height or longest basal diameter) in 25 of 42 eyes that were enucleated because of perceived failure of local control. In 40% of eyes, we could not confirm failure of local control. However, tumors may have grown after initial shrinkage following brachytherapy.
Admittedly, there are many pitfalls in comparing histopathologic with clinical measures. These include the fixation shrinkage, inadequate orientation during sectioning, nonperpendicular ultrasonography, and shadowing caused by highly elevated or mushroom-shaped tumors. Most of these errors tend to underestimate the histologic measure relative to the clinical measure, as confirmed by a recent study comparing clinical measures of COMS tumors immediately preceding enucleation with histopathologic measures after fixation.16 That report showed that histopathologic measures had a mean discrepancy of 0.8 mm less than the clinical estimate for height and 1.6 mm less for largest basal diameter. If those numbers are incorporated as a “compensation” for different measuring techniques, we would still be unable to confirm growth in 9 of 42 of the local failure tumors. Other characteristics—such as hemorrhage, retinal detachment, and tumor shape—may lead to a perception of tumor growth when none has actually occurred.
Examination of secondary enucleations after brachytherapy has shown that these tumors and eyes differ histopathologically from those that had primary enucleation for melanoma. These histopathologic differences may represent effects of radiation, predisposing factors to plaque failure, or characteristics that generate a perception of tumor growth.
Correspondence: Daniel M. Albert, MD, MS, 3310 University Ave, Ste 202, Madison, WI 53705 (firstname.lastname@example.org).
Submitted for Publication: March 20, 2007; final revision received June 7, 2007; accepted June 22, 2007.
Financial Disclosure: None reported.
Funding/Support: The COMS has received support from the National Eye Institute and the National Cancer Institute through cooperative agreements EY06253, EY06257, EY06258, EY06259, EY06020, EY06264, EY06265, EY06266, EY06268, EY06269, EY06270, EY06274, EY06275, EY06276, EY06279, EY06280, EY06282, EY06283, EY06284, EY06287, EY06288, EY06289, EY06291, EY06839, EY06843, EY06844, EY06848, EY06858, and EY06899 with the National Institutes of Health, Bethesda, Maryland.