High-Frequency Ultrasound Characteristics of 24 Iris and Iridociliary Melanomas: Before and After Plaque Brachytherapy

Objective  To evaluate size, characteristics, and regression of iris and iridociliary melanomas on high-frequency ultrasound images before and after plaque brachytherapy.

Methods  Retrospective review of high-frequency ultrasound characteristics of 24 consecutive iris and iridociliary melanomas before and after radiation therapy.

Results  The median tumor thickness before radiation therapy was 2.3 mm (range, 1.4-4.3 mm). Nineteen iris melanomas (79%) involved the ciliary body, 18 (75%) involved the iris pigment epithelium, 11 (46%) were club shaped, and 4 (17%) caused disinsertion of the iris root. At a median follow-up of 30 months after plaque brachytherapy, the mean tumor thickness had diminished to 1.2 mm (median, 1.2 mm; range, 0.9-1.9 mm). While all tumors exhibited a reduction in thickness, no tumors showed additional regression after 30 months past treatment. Fourteen tumors (58%) were noted to have increases in internal reflectivity. There was 1 failure of local control (at 6 years), successfully treated by a second application of plaque brachytherapy.

Conclusion  High-frequency ultrasonography revealed unique tumor characteristics, quantified tumor size, and demonstrated tumor response to radiation therapy.

High-frequency ultrasonography allows for evaluation of the size, occult margins, internal reflectivity, and growth of iris and iridociliary melanomas. It is an invaluable tool for initial diagnosis and for subsequent treatment. High-resolution cross-sectional ultrasound images allow visualization of a tumor's surface characteristics, interstitial borders, and reflectivity.^1-4 Histopathologic correlations of high-frequency ultrasound images of anterior segment tumors have confirmed its usefulness.^4-6 High-frequency ultrasonography is used to identify tumor growth, vascularity, sector cataract, and disturbance of the iris pigment epithelium (IPE).^3-5 Although findings from studies^2,5 suggest that changes in internal reflectivity and tumor size predict malignancy, the usefulness of evaluating involvement of the IPE needs to be proven.

Although iridectomy or iridocyclectomy previously was the procedure of choice for iris and ciliary body melanomas, ophthalmic plaque radiation therapy offers the advantages of larger treatment margins and retained iris function.^7-10 As an extraocular procedure, plaque radiation therapy is associated with little risk of hyphema, infection, or retinal detachment.^11,12 Although iridociliary plaque radiation therapy is likely to cause cataract, it is unlikely to induce radiation maculopathy.^10,11 High-frequency ultrasound tumor measurements facilitate plaque radiation therapy dose calculations and assist in planning for surgical resection.^5,10,13

Ultrasound characteristics of iris and ciliary body melanomas have been described.^2-6,13-15 Most recently, the high-frequency ultrasound features of 4 anterior melanomas before and after brachytherapy were reported.¹⁴ The objective of our study was to evaluate the high-frequency ultrasound characteristics of 24 iris and iridociliary melanomas before and after plaque radiation therapy.

We conducted a retrospective review of 24 patients diagnosed as having iris (n = 5) and iridociliary (n = 19) melanomas and subsequently treated with palladium 103 plaque brachytherapy. Patients selected for this study were consecutive cases with at least 4 months of follow-up. All patients were referred to The New York Eye Cancer Center, New York, where a detailed medical history was followed by an ophthalmic examination. These evaluations included but were not limited to best-corrected visual acuity (Early Treatment of Diabetic Retinopathy Study chart), slitlamp biomicroscopy with photography, Goldmann tonometry, gonioscopy, scleral transillumination, ultrasonography, and indirect ophthalmoscopy. Phakic patients were examined for the presence of sector cataract. This study adhered to the tenets of the Declaration of Helsinki and the Health Insurance Portability and Accountability Act of 1996. High-frequency (20-, 35-, or 50-MHz) ultrasonography was typically performed at the initial visit, during periods of observation for tumor growth, and every 4 months after plaque brachytherapy. Within the framework of The New York Eye Cancer Center, we used 3 commercially available units. Early scans were performed using a 50-MHz transducer (Paradigm Medical Industries, Salt Lake City, Utah). This highest-frequency ultrasound transducer provides the greatest resolution and the least intraocular penetration. Examinations of larger lesions typically required the examiner to tape together sequential images (to include the entire lesion). Subsequently, an ophthalmic 20-MHz transducer (Innovative Imaging, Inc, Sacramento, California) became available. This lower-frequency unit provides a larger field, integrated calipers, and ease of use through a miniature water-filled latex condom. Our most recently acquired 35-MHz transducer-based machine (Ophthalmic Technologies Inc, Toronto, Ontario) requires the use of an isotonic sodium chloride solution–filled eye cup (water bath). However, the newer machines have the advantage of computerization. This allows for recording of the examination as a video, with subsequent image capture, magnification, and measurement (with integrated calipers). In addition to tumor evaluations, we routinely examine all 360° of the anterior segment to rule out additional tumors (eg, ring melanoma). In this study, we evaluated the initial and subsequent ultrasonograms. There may have been some variation in tumor size related to the capability and resolution of the 3 high-frequency ultrasound instruments. However, all reported tumor sizes represent the best possible measurements (by P.T.F.). In comparing serial ultrasonograms for the same patient using different machines, every effort was made to account for the inherent instrument-related differences in analyzing ultrasound features. Regardless of the instrument used, tumor thicknesses were measured at the thickest portion of the tumor, whether in the ciliary body or in the iris. Width was typically determined by evaluation of adjacent tissues for tumor invasion. Longitudinal and transverse tumor diameters were measured (Figure 1). The largest transverse dimension was recorded in the iris or in the ciliary body (depending on its location). Other high-frequency ultrasound characteristics evaluated included the tumor shape, scleral invasion, involvement of the IPE, internal tumor reflectivity, disinsertion of the iris root, angle morphological features, and presence of intratumoral hypoechoic spaces.

In this study, there were 3 types of melanomas: (1) ciliary body tumor with iris displacement or extension, (2) diffuse iris melanoma that extended into the supraciliary space, and (3) pure iris melanomas that did not extend posterior to the iris root. Assessment of tumor growth was used to help determine the stage of malignancy for all 3 types of melanomas. However, because the metastatic potential for focal iris melanomas is low, criteria for treating these tumor were growth suggestive of malignancy or biopsy-proven malignant melanoma.

Tumor shapes were defined as club, dome, or irregular. The internal reflectivity was graded as low, moderate, or high. The iris was evaluated for attachment to the ciliary body at the iris root or for disinsertion by the tumor. The IPE was determined to be involved if it was displaced or eroded by the tumor. Ultrasonographic evidence of subjacent cataract (lens hyperreflectivity) was noted. Tumor thickness was evaluated by serial high-frequency ultrasound examinations every 4 months after plaque brachytherapy.

Analysis of 24 iris and iridociliary melanomas revealed that their median follow-up was 30 months (mean, 35.5 months; range, 4 months to 10 years). The median patient age was 63.5 years (mean age, 60 years; age range, 29-90 years) (Table 1). Fourteen patients (58%) were female, and 10 patients (42%) were male. Clinical data were available for all 24 patients. Eleven melanomas (46%) were present in the right eye, and 13 melanomas (54%) were present in the left eye. Eleven of 24 irises (46%) exhibited corectopia. Nine of 24 irises (38%) had tumor-associated ectropion uveae (Figure 2). Before plaque brachytherapy, 6 patients had biopsy-proven malignant melanoma, 1 by surgical iridectomy and 5 by the minimally invasive iridectomy technique described by Finger et al.¹⁶ After healing for 3 weeks, tumors were remeasured by high-frequency ultrasonography for radiation dosimetry planning.

Evaluation of visual acuity revealed that 19 patients (79%) retained visual acuity within 1 line of their preoperative visual acuity on the Early Treatment of Diabetic Retinopathy Study chart (Table 2). Of 5 patients who did not, 1 had age-related macular degeneration, 2 had cataracts, and 2 required enucleation for melanomalytic glaucoma. Fourteen patients (58%) were noted to have cataract formation following surgery, 11 of whom underwent cataract surgery at a mean of 34.5 months (range, 9-68 months) after brachytherapy. Four patients (17%) required other postsurgical procedures (1 each required endolaser, argon laser trabeculoplasty, and enucleation, and 1 patient required both cyclocryotherapy with subsequent enucleation). Two patients (8%) developed iris neovascularization. Seven patients (29%) had elevated intraocular pressure secondary to pigment dispersion alone (n = 5) or to a combination of neovascularization of the angle (n = 2). Three patients had glaucoma before brachytherapy.

On high-frequency ultrasonography, the initial mean ± SD tumor thickness before radiation therapy was 2.6 ± 0.8 mm (median, 2.3 mm; range, 1.4-4.3 mm) (Table 3). Nineteen melanomas (79%) involved the ciliary body, and 5 tumors (21%) were confined to the iris (Figure 3). Eighteen tumors (75%) had low reflectivity, and 6 tumors (25%) had moderate reflectivity. Eleven iris melanomas(46%) were club shaped, 7 (29%) were dome shaped, and 6 (25%) were irregularly shaped (Figure 4). Four tumors (17%) caused disinsertion of the iris root (Figure 5), 18 tumors (75%) involved the IPE (Figure 6), and 12 tumors (50%) caused blunting of the angle (Figure 7). Two tumors (8%) had intratumoral hypoechoic spaces, and no tumors had invaded the sclera.

After analysis of comparative intraocular dosimetry, we chose to use palladium 103 plaques (vs iodine 125).^12,17-19 In every case, comparative dosimetric analysis (palladium 103 vs iodine 125 plaques) revealed that the use of palladium 103 resulted in less or decreased irradiation of the optic disc and macula. Therefore, all patients were treated with palladium 103 plaques. At a median of 30 months after plaque brachytherapy, the mean ± SD tumor thickness was reduced to 1.2 ± 0.4 mm (median, 1.2 mm; range, 0.9-1.9 mm) (Table 2). The mean ± SD change in thickness of an iris or iridociliary melanoma was 1.4 ± 0.9 mm (median, 1.3 mm; range, 0.1-3.3 mm). All tumors were reduced in thickness after treatment.

Of 5 tumors that were followed up beyond 3 years (mean follow-up, 84.4 months; median follow-up, 74 months; follow-up range, 72-120 months), no tumor showed additional regression after 30 months. Fourteen tumors were followed up beyond 2 years (mean follow-up, 50.4 months; median follow-up, 36 months; follow-up range, 26-120 months), and 8 melanomas did not show additional regression after 2 years. Six tumors continued to regress at between 24 and 36 months. The mean rate of reduction in tumor thickness during the first 30 months was 0.4 mm (median, 0.3 mm; range, 0.1-1.7 mm) per 6-month interval. Seventeen tumors (71%) were reduced by at least 33% of the initial thickness by 24 months (Figure 8). Fourteen tumors (58%) increased in internal reflectivity after therapy (Figure 1), while 10 tumors (42%) exhibited no change in reflectivity (Figure 9).

One patient (4%) died of metastatic disease within 3 years of radiation therapy, and 1 patient (4%) was found to have local recurrence 6 years after initial treatment. The latter tumor was irregularly shaped, involved the ciliary body and the IPE, caused blunting of the angle and iris root disinsertion, and exhibited an increase in reflectivity after therapy. It was initially considered a local cure, having decreased in thickness by 0.6 mm (30%), with an increase in internal reflectivity. When the tumor enlarged, it was treated by an additional application of palladium 103 plaque radiation therapy (8800 rad [88 Gy]). Although the tumor was subsequently controlled for 48 months, the patient died of myocardial infarction.

The diagnosis and treatment of malignant iris and iridociliary melanomas are important because the tumors can metastasize and induce secondary glaucoma.^20-22 Before the advent of high-frequency ultrasonography, these tumors were evaluated using slitlamp photography, gonioscopy, and conventional B-scan ultrasonography. High-frequency ultrasonography has expanded our capability to serially evaluate various tumor characteristics. Subtle interstitial areas of ciliary body invasion of an iris melanoma can be detected by high-frequency ultrasonography (Figure 3). Decreased internal tumor reflectivity can help discriminate an iris melanoma from a nevus.⁵ High-frequency ultrasonography can visualize IPE and iris root involvement that may be missed using older diagnostic modalities (Figures 3 and 4). Furthermore, high-frequency ultrasonography has shown usefulness in delineating the borders of a tumor by helping distinguish benign scleral pigmentation from invasive melanoma.¹⁴

Investigators have reported on the high-frequency ultrasound characteristics of iris melanomas and the usefulness of this imaging modality in showing tumor growth.^2,4,5,14,15 Conway et al¹⁵ used high-frequency ultrasonography to demonstrate that a high percentage (81%) of iris melanomas distorted surrounding structures, such as the ciliary body and posterior iris plane. Other reports evaluated the frequency of cavitations within iris tumors detected by high-frequency ultrasonography and found them to be present in up to 8.4% of cases.^23-25 To our knowledge, only 1 previous study¹⁴ reported on the high-frequency ultrasound features of anterior uveal melanomas after plaque brachytherapy, and Torres et al found a mean reduction in thickness of 1.12 mm after a median follow-up of 23 months in 4 tumors.

In this series of 24 iris and iridociliary melanomas, most tumors were club shaped (11 of 24 [46%]), involved the ciliary body (19 of 24 [79%]), distorted the IPE (18 of 24 [75%]), and caused blunting of the angle (12 of 24 [50%]). Four tumors (17%) had disinserted the iris root, and no melanomas had invaded the sclera. We observed 2 patients (8%) who had iridociliary melanomas with intratumoral hypoechoic spaces, which may represent cystic degeneration or vascularity.

Tumors in this study were clinically diagnosed uveal melanomas, exhibited growth while under observation, or were biopsy-proven malignant melanoma.¹⁶ Plaque brachytherapy provided local control in 23 patients (96%) and preservation of initial visual acuity in 19 patients (79%). In comparison, there was 90% local control in an iridocyclectomy series, with 57% of eyes achieving 20/50 visual acuity or better.⁹ Our increased control may be explained by the use of larger tumor margins and shorter follow-up. However, since plaque brachytherapy is an extraocular procedure, it all but eliminates the risks inherent to intraocular surgery (as well as pupil deformation).^7,8,12

Following therapy, high-frequency ultrasonography was a valuable tool for documenting regression. In our study of 24 iris and iridociliary melanomas treated with plaque brachytherapy during a median follow-up of 30 months, there was a mean reduction in tumor thickness of 1.4 mm. In addition, tumor regression occurred during 30 months by a mean 0.25 mm for every 6 months of follow-up, and one-third of the tumors had regressed to normal iris and ciliary body thickness by the first year after treatment. Fourteen tumors (58%) in our study increased in reflectivity after treatment. A similar tendency was reported by Torres et al.¹⁴ Pavlin et al² suggested that high reflectivity was histopathologically correlated with poorly cohesive cells with resultant large intercellular spaces. Therefore, our finding of increased internal reflectivity after plaque brachytherapy may represent a decrease in the density and cohesion of the uveal melanoma cells within treated tumors. We found that changes in reflectivity did not always correlate with reductions in tumor thickness. Therefore, tumor response to plaque brachytherapy should be monitored by observation for a decrease in tumor size (thickness and width) and for changes in internal reflectivity. As noted in our 1 case of failure of local control despite an initial response to therapy (shrinkage and an increase in reflectivity), the potential for recurrence continues and underscores the need for follow-up vigilance.

Our study findings demonstrate that the diagnosis and classification of iris and iridociliary melanomas can be obtained by observation of their visually apparent and high-frequency ultrasound characteristics. High-frequency ultrasound characteristics include involvement of the ciliary body, IPE distortion, disinsertion of the iris root, and tumor shape and growth. After plaque brachytherapy, this imaging modality should again be used to serially document a tumor's response to therapy (shrinkage and reflectivity). High-frequency ultrasonography revealed unique tumor characteristics, quantified tumor size, and demonstrated tumor response to radiation therapy.

Correspondence: Paul T. Finger, MD, The New York Eye Cancer Center, 115 E 61st St, New York, NY 10065 (pfinger@eyecancer.com).

Submitted for Publication: July 17, 2006; final revision received December 8, 2006; accepted December 11, 2006.

Financial Disclosure: None reported.

Funding/Support: This study was supported by The EyeCare Foundation, Inc.