Pars Plana Vitrectomy in Eyes Containing a Treated Posterior Uveal Melanoma WILLIAM J. FOSTER, MD, PHD, J. WILLIAM HARBOUR, MD, NANCY M. HOLEKAMP, MD, GAURAV K. SHAH, MD, AND MATTHEW A. THOMAS, MD
● PURPOSE: To determine the safety of pars plana vitrectomy in eyes containing a treated posterior uveal melanoma. ● DESIGN: Interventional case series. ● METHODS: Retrospective case series of patients with posterior uveal melanoma who underwent pars plana vitrectomy. Complications, vitreous cytology, local tumor control, and metastasis were assessed. ● RESULTS: Nine patients met study criteria. Tumors were treated with 125I plaque radiotherapy (seven patients) or transpupillary thermotherapy (two patients). Vitrectomy was performed for vitreous hemorrhage (five patients), macular pucker (two patients), macular hole (one patient), and rhegmatogenous retinal detachment (one patient). Vitrectomy was performed at a mean of 24.7 months (range, 7– 47 months) after melanoma treatment. Dispersion of tumor cells at vitrectomy was not observed in any patients. Melanoma cells were detected in the vitreous aspirate in one of seven cases examined cytologically. This patient had intratumoral and vitreous hemorrhage before plaque radiotherapy, underwent combined vitrectomy/cataract extraction, and developed intraocular tumor dissemination 56 months after vitrectomy. No other patients developed intraocular tumor dissemination. At mean follow-up of 24 months (range, 3– 63 months) after vitrectomy, none of the nine patients developed systemic metastasis. ● CONCLUSIONS: Pars plana vitrectomy rarely may lead to intraocular tumor dissemination, although the risk of this complication is probably low if the tumor has been treated and has responded to therapy before vitrectomy. Vitrectomy should be approached with caution if a vitreous hemorrhage is present, especially if the hemorAccepted for publication Feb 24, 2003. InternetAdvance publication at ajo.com June 18, 2003. From the Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, and Barnes Retina Institute, St. Louis, Missouri. This study was presented in part at the combined meeting of the Vitreous and Retina Societies, San Francisco, California, October 2002. Inquiries to J. William Harbour, MD, Department of Ophthalmology and Visual Sciences, Washington University School of Medicine, Campus Box 8096, 660 South Euclid Ave, St. Louis, MO 63110; fax: (314) 747-5073; e-mail:
[email protected] 0002-9394/03/$30.00 doi:10.1016/S0002-9394(03)00244-7
©
2003 BY
rhage occurred before tumor treatment, as this may seed tumor cells into the vitreous cavity. (Am J Ophthalmol 2003;136:471– 476. © 2003 by Elsevier Inc. All rights reserved.)
P
OSTERIOR UVEAL MELANOMA OCCASIONALLY MAY
coexist in the same eye with a visually significant retinal disorder that is treatable by pars plana vitrectomy, such as rhegmatogenous retinal detachment, macular hole, macular pucker, and vitreous hemorrhage.1 However, vitrectomy potentially may induce intraocular or extraocular tumor dissemination, and it remains unclear whether vitreous surgery is safe to perform in this setting. Several studies have established the safety of cataract surgery in eyes containing a melanoma treated with plaque or charged particle radiotherapy.2– 4 However, cataract surgery typically does not require entry into the posterior segment or manipulation of the vitreous in proximity to the melanoma. Repair of rhegmatogenous retinal detachments by scleral buckling techniques in melanoma-containing eyes appears to be safe, especially if external drainage can be avoided.1,5 There are few reports of pars plana vitrectomy in eyes containing uveal melanomas. Radtke and coworkers6 described a single patient who underwent vitrectomy for persistent exudative retinal detachment after plaque radiotherapy; this patient had no evidence of tumor dissemination at 37 months after surgery. Another case report described a patient who underwent vitrectomy for excision of a presumed subfoveal choroidal neovascular membrane that proved histopathologically to be a uveal melanoma. No tumor dissemination was reported, but there was limited follow-up of only 7 months. A multicenter, retrospective study from five institutions reported the results of retinal reattachment surgery in 10 patients with coexisting posterior uveal melanomas,1 but only two of these patients underwent pars plana vitrectomy. Neither patient developed dissemination after approximately 4 years of follow-up. In this study, we provide further information on the safety of pars plana vitrectomy in eyes containing a treated posterior uveal melanoma by presenting the ocular and systemic findings in nine patients from a single institution.
ELSEVIER INC. ALL
RIGHTS RESERVED.
471
TABLE 1. Initial Patient Features and Pretreatment Tumor Characteristics
Patient
Age
Sex
Race
Eye
Basal Tumor Dimensions (mm)
Tumor Thickness (mm)
Tumor Location
Tumor Treatment
1 2 3 4 5 6 7 8 9
64 52 61 54 57 68 73 71 84
M M M M M F M M M
W W W W W W W W W
Right Left Left Right Left Left Right Right Left
7⫻5 15 ⫻ 12 10 ⫻ 9 9⫻8 10 ⫻ 7 9⫻7 12 ⫻ 10 12 ⫻ 11 7⫻6
2.9 4.3 8.0 10.9 2.5 2.5 5.7 4.0 3.5
Posterior Posterior Posterior Anterior Posterior Posterior Anterior Anterior Anterior
Plaque Plaque ⫹ TTT Plaque Plaque TTT TTT Plaque Plaque Plaque
F ⫽ female; M ⫽ male; TTT ⫽ transpupillary thermotherapy; W ⫽ white.
dimensions, and thickness at vitrectomy. Results of cytologic examination of vitreous specimen were recorded in the seven cases in which this was performed (in the other two cases the surgeon elected not to obtain cytologic analysis). Follow-up data recorded included date of last follow-up and ocular and systemic status at last follow-up. Results of histopathologic evaluation of enucleated globes were recorded when available.
METHODS THIS STUDY WAS PERFORMED WITH APPROVAL OF THE
Institutional Review Board of Washington University. Patients were identified by computer search of the Barnes Retina Institute medical records for patients with a diagnosis of uveal melanoma who underwent pars plana vitrectomy. Patients were included if they underwent treatment for a posterior segment tumor that was typical for uveal melanoma based on clinical and ultrasonographic examination7 and if they underwent a three-port pars plana vitrectomy for a vitreoretinal disorder. Patients were excluded if the vitrectomy was performed in conjunction with local surgical resection of an iridociliary melanoma. Nine patients who underwent vitrectomy between October 1990 and May 2002 met these criteria. Patient 4 was previously reported1 but subsequently developed further findings that are described herein. Plaque radiotherapy was performed with standard Collaborative Ocular Melanoma Study (COMS) style plaques and 125I radioactive seeds calculated to deliver 85 to 90 Gy to the tumor apex over 4 days. Transpupillary thermotherapy was performed in three treatment sessions, with subsequent tumor flattening, as previously described.8 Each patient underwent a standard three-port vitrectomy. In most cases, the vitrectomy cassette containing the vitreous aspirate was sent for cytologic examination, which was performed by spinning down the aspirate, smearing it onto slides, and staining with Papanicolou or DiffQuick stain. Baseline data recorded for each patient included patient age, sex, race, involved eye, tumor basal dimensions (estimated by indirect ophthalmoscopy), and thickness (estimated by A-scan ultrasonography) at diagnosis, tumor location (posterior vs anterior to the equator), tumor treatment, and date. Data related to the vitrectomy that were recorded included date of surgery, indication for vitrectomy, surgical procedures performed, tumor basal 472
AMERICAN JOURNAL
RESULTS NINE PATIENTS MET INCLUSION CRITERIA FOR THE STUDY.
Patient data and initial tumor characteristics are summarized in Table 1. In all nine patients, the tumor was treated before vitrectomy. 125I plaque radiotherapy was used as the only treatment in six patients, transpupillary thermotherapy (TTT) alone was used in two patients, and 125I plaque radiotherapy combined with TTT was used in one patient. The mean largest basal tumor dimension before tumor treatment was 10.1 mm (median, 10 mm; range, 7–15 mm), and the mean tumor thickness was 5.0 mm (median, 9 mm; range, 2.5 –10.9 mm). The tumor location was juxtapapillary in three cases, anterior to the equator in three cases, between the temporal vascular arcades and the equator in two cases, and macular in one case. All patients underwent a three-port pars plana vitrectomy (Table 2). The mean time from initial tumor treatment (either plaque implantation or initial TTT) to vitrectomy was 24.7 months (median, 27; range, 7– 47 months). The indications for vitrectomy included vitreous hemorrhage (five patients), macular pucker (two patients), macular hole (one patient), and rhegmatogenous retinal detachment (one patient). Vitreous hemorrhage occurred before tumor treatment in one case (Patient 4) and after tumor treatment in four patients. Tumor treatment caused a decrease in tumor thickness before vitrectomy in all OF
OPHTHALMOLOGY
SEPTEMBER 2003
TABLE 2. Summary of Pars Plana Vitrectomy Data
Patient
Reason for Vitrectomy
Interval From Tumor Treatment to Vitrectomy (months)
Tumor Thickness at Vitrectomy (mm)
Additional Procedures at Vitrectomy
Melanoma Cells by Vitreous Cytology
Follow-up After Vitrectomy (months)
Intraocular Tumor Dissemination at Final Follow-up
Metastasis at Final Follow-up
1 2 3 4 5 6 7 8 9
VH VH VH VH Pucker Pucker RhRD VH MH
47 30 32 19 22 27 7 29 13
2.0 0 3.8 2.4 1.4 0 3.0 0 2.2
MP,EL EL EL CE MP MP FAE,EL CE MP, FAE
ND Negative Negative Positive Negative Negative Inconclusive ND Negative
5 17 2 56 12 5 63 53 3
No No No Yes No No No No No
No No No No No No No No No
CE ⫽ cataract extraction and implantation of posterior chamber lens; EL ⫽ endolaser photocoagulation; FAE ⫽ fluid–air exchange; MH ⫽ macular hole; MP ⫽ membrane peeling; ND ⫽ not done; RhRD ⫽ rhegmatogenous retinal detachment; VH ⫽ vitreous hemorrhage.
cases. The mean tumor thickness at the time of vitrectomy was 1.6 mm (median, 1.37 mm; range, 0 –3.8 mm). In no case was there evidence at the time of vitrectomy that the surgical procedure disrupted the tumor or dispersed tumor cells. Additional surgical procedures during vitrectomy included endolaser photocoagulation (four patients), membrane peeling (four patients), fluid–air exchange (two patients), cataract extraction (two patients), and internal drainage of subretinal fluid (one patient). The vitreous aspirate was sent for cytologic examination in seven patients. Scant cells were found in the vitreous aspirates and were interpreted as negative for malignant cells in five patients, suspicious for malignant cells in one patient (Patient 7), and positive for malignant cells in one patient (Patient 4) based on cytomorphologic features (for example, prominent nucleolus, nuclear atypia). After mean follow-up of 24 months (range, 3– 63 months), eight of the nine patients showed no evidence of intra- or extraocular tumor dissemination. Patient 4 developed intraocular tumor dissemination and is described in detail below. None of the nine patients developed metastatic disease.
CASE REPORT A 54-YEAR-OLD MAN PRESENTED WITH FLOATERS AND
blurred vision in the right eye. His visual acuity in that eye was 20/25, and he was found to have a rhegmatogenous retinal detachment in the nasal periphery associated with three retinal breaks. In the superonasal quadrant of the same eye, an elevated mushroom-shaped, brown tumor was present. The tumor basal dimensions were 9 ⫻ 8 mm, and the ultrasound thickness was 10.9 mm. The internal reflectivity was consistent with a uveal melanoma. The VOL. 136, NO. 3
VITRECTOMY
IN
retinal detachment was repaired by pneumatic retinopexy with 0.3 ml perfluorethane (C2F6) gas and indirect laser photocoagulation to the retinal breaks, which was followed by prompt retinal reattachment. However, 1 week later the patient complained of sudden vision loss and was found to have a moderate vitreous hemorrhage. Blood appeared to be emanating from the tumor surface. Three months were allowed for the vitreous hemorrhage to clear, and the patient was then treated with 125I plaque radiotherapy. The tumor subsequently responded with marked shrinkage. Nevertheless, recurrent vitreous hemorrhages persisted. Two years later, the tumor had regressed to a thickness of 2.4 mm, but visual acuity had dropped to 20/200 owing to persistent vitreous hemorrhage and cataract. The patient underwent pars plana vitrectomy combined with phacoemulsification and posterior capsulotomy. The vitreous aspirate was sent to cytology and was found to contain definite melanoma cells. The postoperative visual acuity improved to 20/50. The tumor continued to regress with no evidence of local recurrence. Neovascular glaucoma was treated with panretinal photocoagulation. The vision eventually declined to light perception owing to radiation retinopathy. Two years later (76 months after initial presentation and 54 months after vitrectomy), the right eye became painful and was noted to have elevated intraocular pressure and multiple, lightly pigmented nodules in the anterior chamber angle and iris stroma, consistent with intraocular tumor dissemination (Figure 1). The eye was enucleated, and histopathologic examination revealed diffuse seeding of melanoma cells along the vitreous base and on the posterior and anterior iris surfaces, with multiple tumor nodules invading the anterior chamber angle and iris stroma. Only a flat scar of the primary tumor remained. Eight months later, the patient remained free of metastatic disease.
EYES CONTAINING
A
UVEAL MELANOMA
473
FIGURE 1. Intraocular dissemination of posterior uveal melanoma 54 months after combined pars plana vitrectomy, cataract extraction, and posterior capsulotomy (Patient 4). (A) Slit-lamp photograph of right eye, showing light brown tumor nodules in the anterior chamber angle at 6:00 and 9:00 o’clock. (B) Goniophotograph showing same tumor nodules reversed in mirror (white arrows). Note prominent iris vessels feeding the tumor nodules (black arrow). (C) Digitally enhanced composite photomicrograph showing tumor nodule invading the anterior chamber angle and a smaller tumor nodule on the iris stroma near the pupil. (D, E) Melanoma cells in the vitreous base overlying the pars plana and migrating anteriorly along the ciliary processes (arrow). (F) The primary uveal melanoma, treated with plaque radiotherapy 19 months before vitrectomy, demonstrating complete regression to a flat scar. Note residual pigment-laden macrophages and overlying atrophic retina (hematoxylin-eosin staining, original magnification ⴛ40).
474
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
SEPTEMBER 2003
DISCUSSION IN THIS ARTICLE, WE PRESENT NINE PATIENTS WHO UNDER-
went pars plana vitrectomy for repair of a vitreoretinal abnormality after treatment of a posterior uveal melanoma in the same eye. This represents the largest series reported to date. Seven tumors were treated with plaque radiotherapy and two with transpupillary thermotherapy. In all cases, the tumor regressed substantially before vitrectomy. In six of seven patients in whom the vitreous aspirate was examined for cytology, malignant cells were not found, and in eight of the nine patients, no intraocular or extraocular tumor dissemination occurred. In one patient, intratumoral bleeding and vitreous hemorrhage occurred before plaque radiotherapy, melanoma cells were found in the vitreous aspirate, and diffuse intraocular tumor dissemination occurred subsequent to combined cataract extraction/ vitrectomy. We conclude that pars plana vitrectomy in an eye containing a previously treated uveal melanoma is probably safe in most patients, as long as the tumor has responded to treatment before vitrectomy and there was no vitreous hemorrhage before tumor treatment to indicate possible retinal perforation and tumor dispersion. Patient 4 developed intraocular tumor dissemination 54 months after vitrectomy. Several factors may have contributed to the poor outcome in this case. First, vitreous hemorrhage from intratumoral bleeding occurred before treatment of the tumor and may have allowed viable tumor cells to disseminate into the vitreous. Vitreous hemorrhage is encountered not infrequently with in eyes containing a posterior uveal melanoma, but intraocular tumor dissemination in such patients (most of whom have not undergone subsequent vitrectomy) is extremely rare, suggesting that the vitrectomy in this patient may have contributed to the tumor dissemination. A similar case was reported by Laqua and associates9 in which vitrectomy was performed for vitreous hemorrhage in an eye that was found intraoperatively to contain a melanoma. Intraocular tumor dissemination subsequently developed, with histopathologic findings similar to our patient. In contrast, tumor dissemination occurred in none of the four patients who developed a vitreous hemorrhage after tumor treatment. The nonclearing vitreous hemorrhage in this case also prevented adequate clinical examination of the posterior segment, which may have revealed the disseminated tumor. In addition, cataract extraction and posterior capsulotomy at the time of vitrectomy potentially may have provided access for tumor cells into the anterior chamber. Shortcomings of this study include the small number of patients and limited follow-up. Thus, firm conclusions cannot be drawn from this study. However, this is the largest series reported so far, and provisional guidelines from this study may be useful until a larger study is conducted. First, treatment of the melanoma, preferably with irradiation, should be probably be performed before vitrectomy (or other intraocular surgery) whenever possiVOL. 136, NO. 3
VITRECTOMY
IN
ble. Several reports have documented intraocular tumor recurrence in patients who underwent vitrectomy without tumor irradiation.9 –11 The lack of tumor dissemination in eight of our nine patients (89%), all of whom underwent tumor treatment before vitrectomy, suggests that tumor dissemination may be less likely if the tumor has been previously treated. Plaque radiotherapy destroys the reproductive integrity of melanoma cells by inducing DNA damage.12 Therefore, any melanoma cells that might disseminate into the vitreous during vitrectomy would be expected to be nonviable. It is not clear how long one should wait to perform vitrectomy after tumor treatment. The shortest time interval in our series was 7 months, and no tumor dissemination developed in this patient. Although this study does not allow us to conclude the shortest safe interval between tumor treatment and vitrectomy, a definite treatment response (for example, resolution of exudative retinal detachment, decreasing tumor thickness, change in tumor color) should probably be observed before vitrectomy. Second, the vitreous aspirate should be sent for cytologic examination at the time of vitrectomy. The presence of malignant cells in the vitreous should alert the physician to follow the patient carefully for the possibility of tumor dissemination and possibly to refrain from further intraocular surgery. Third, vitrectomy should be approached with caution in the presence of a vitreous hemorrhage, particularly if the hemorrhage occurred before tumor treatment. Vitreous hemorrhage can be due to intratumoral bleeding, which may seed tumor cells into the vitreous cavity, especially if the tumor has broken through the Bruch membrane and ruptured the overlying retina. Conservatively, one should consider the possibility that any vitreous hemorrhage associated with a melanoma could potentially contain malignant cells, even if other explanations are possible. In conclusion, we found no evidence of intraocular tumor dissemination as a result of pars plana vitrectomy in eight treated posterior uveal melanomas that were not associated with vitreous hemorrhage before tumor treatment. Larger studies with longer follow-up will be needed to determine the safety of vitrectomy in this setting.
REFERENCES 1. Haimovici R, Mukai S, Schachat AP, Thomas MA, Meredith TA, Gragoudas ES. Rhegmatogenous retinal detachment in eyes with uveal melanoma. Retina 1996;16: 488 –496. 2. Wachtlin J, Bechrakis NE, Schueler AO, Helbig H, Bornfeld N, Foerster MH. Phacoemulsification following treatment of choroidal melanoma. Graefes Arch Clin Exp Ophthalmol 2000;238:942–948. 3. Gragoudas ES, Egan KM, Arrigg PG, Seddon JM, Glynn RJ, Munzenrider JE. Cataract extraction after proton beam
EYES CONTAINING
A
UVEAL MELANOMA
475
4. 5.
6. 7.
8. Oosterhuis JA, Journee-de Korver HG, Kakebeeke-Kemme HM, Bleeker JC. Transpupillary thermotherapy in choroidal melanomas. Arch Ophthalmol 1995;113:315–321. 9. Laqua H, Vo¨ lcker H. Pars plana vitrectomy in eyes with malignant melanoma. Graefe Arch Exp Clin Ophthalmol 1983;220:279 –284. 10. Bechrakis NE, Foerster MH, Bornfeld N. Biopsy in indeterminate intraocular tumors. Ophthalmology 2002;109:235– 242. 11. Damato B, Wong D, Green FD, Mackenzie JM. Intrascleral recurrence of uveal melanoma after transretinal “endoresection.” Br J Ophthalmol 2001;85:114 –115. 12. Brantley MA Jr, Worley L, Harbour JW. Altered expression of Rb and p53 in uveal melanomas following plaque radiotherapy. Am J Ophthalmol 2002;133:242–248.
irradiation for malignant melanoma of the eye. Arch Ophthalmol 1992;110:475–479. Fish GE, Jost BF, Snyder WI, Fuller DG, Birch DG. Cataract extraction after brachytherapy for malignant melanoma of the choroid. Ophthalmology 1991;98:619 –622. Boniuk M, Zimmerman L. Occurrence and behavior of choroidal melanomas in eyes subjected to operations for retinal detachment. Trans Am Acad Ophthalmol Otolaryngol 1962;66:642–657. Radtke N, Augsburger J, Schmitt T. Management of exudative retinal detachment after plaque therapy for intraocular melanoma. Am J Ophthalmol 1991;112:92–94. Harbour JW, Murray TG. In: Albert D, editor. Ophthalmic surgery. Principles and techniques. Malden, MA: Blackwell, 1998:682–705.
476
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
SEPTEMBER 2003
