- Email: [email protected]
Current Results of Proton Beam Irradiation of Uveal Melanomas
Current Results of Proton Beam Irradiation of Uve al Mel ano ma s EVANGELOS S. GRAGOUDAS, MD,* JOHANNA SEDDON, MD,* MICHAEL GOITEIN, PhD,t LYNN VERHEY, PhD,t JOHN MUNZENRIDER, MD,t MARSHA URIE, PhD, t HERMAN D. SUIT, MD, t PETER BLITZER, MD, t ANDREAS KOEHLER, AM*
Abstract: Proton beam irradiation has been used for the treatment of 241 uveal melanomas over the past 7~ years. Twelve melanomas (5%) were small, 99 (41%) medium, 103 (43%) large and 27 (1%) extra-large melanomas. The mean length of follow-up was 21 months and the median 15 months. Ninety-four percent of the treated lesions with a follow-up more than two years and 65% of tumors with shorter follow-up showed regression. The most recent visual acuity was 20/40 or better in 47% and 20/100 or better in 66%. Ten eyes were enucleated because of complications (9) or continued tumor growth (1 ). Thirteen patients developed metastases from 4 to 50 months of treatment. Our data indicate that proton irradiation can be used to treat melanomas of various sizes and in a variety of locations, and preliminary results suggest that proton therapy has no deleterious effect on the likelihood of the development of metastases. [Key words: irradiation, melanoma, proton beam, treatment, uvea.] Ophthalmology 92:284-291, 1985
Enucleation has long been the usual treatment for posterior malignant uveal melanoma. The advantage of this method in terms of increased life expectancy has been recently challenged 1 and alternative forms of treatment have been used with increased frequency during the past few years. 2- 8 Photocoagulation2 and local resection3 have been used in selected instances, but radiotherapy remains the most widely used method. 4 - 8 There are two major radiotherapeutic techniques for the treatment of uveal melanomas. Radioactive plaques can be sutured on the sclera over the area of the tumor4 - 6 and particulate radiations such as protons8 and helium ions7 can be used. The advantages of the latter
From the Retina Service, Massachusetts Eye and Ear Infirmary Department of Ophthalmology, Harvard Medical School,* Department of Radiation Medicine, Massachusetts General Hospital,t Boston, and Harvard Cy· clotron Laboratory:j: Cambridge. Presented at the Eighty-eighth Annual Meeting of the American Academy of Ophthalmology, Chicago, Illinois, October 3D-November 3, 1983. Reprint requests to Evangelos S. Gragoudas, MD, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston. MA 02114.
284
modalities are premised on better dose distributions between tumor and normal tissues. The dose delivered with protons or helium ions can be localized to the tumor more accurately, excluding from the radiation field all or some of the radiosensitive intraocular structures.7·8 Hopefully, this would improve the therapeutic ratio of local control versus complications. Proton beam irradiation has been used for the treatment of uveal melanomas at the Harvard Cyclotron since 1975, and its application has increased in frequency during the past few years. 8- 11 We report here the current results of treatment of 240 patients with uveal melanoma (241 eyes).
MATERIALS AND METHODS All patients were treated between July 1975 and December 1982. Follow-up data were analyzed up to June 1983. The 122 men and 118 women treated ranged in age from 14 to 83 years (mean and median age, 56 years). Most of the patients were in the sixth and seventh decades of life. All patients were white. The right eye
GRAGOUDAS, et al
was treated in 126 patients, the left in 113 and both eyes in one patient. The diagnosis was established in all cases through ( 1) clinical examination by multiple experienced consultants using indirect ophthalmoscopy and slit lamp biomicroscopy with contact lens, and (2) assessment of the results of ancillary tests, such as standard and wide angle fundus photography, fluorescein angiography and ultrasonography (both immersion A-scan and B-scan). Radioactive phosphorus (P32) uptake measurements were done routinely in the initial stages of the study but presently are performed only rarely, in cases where the diagnosis is still in doubt. All the patients received a complete systemic examination by an internist to exclude metastasis or other primary malignancy. Chest roentgenography and liver function studies were performed routinely; a liver scan was performed in selected cases. Patients with uveal melanomas up to 24 mm in largest diameter and 14 mm in height were treated. Largest diameters up to the axial length of the eye were observed in ciliochoroidal melanomas extending circumferentially. In general, larger lesions could be treated when located at the peripheral fundus, because the beam could enter the eye directly in the area of the tumor; thus the irradiated ocular volume can be kept smaller than in cases of posteriorly located tumors in which the beam enters the side of the globe opposite the lesion. Documentation of growth was required for small lesions (10 mm or less in diameter and 2 mm or less in height) in all but three cases which were treated early in this series. One hundred seventy-one patients presented with symptoms caused by their tumor. Those were in order of frequency: visual loss, photopsias, floaters, pain and/ or inflammation. Six patients had bilateral melanomas. Four had enucleation of one eye previously, one received proton irradiation to both eyes, and one had treatment of one eye with a Cobalt plaque two years prior to proton irradiation of his other eye. Ten patients presented with vitreous hemorrhage and 158 with retinal detachment at the time of initial examination. Sixty-nine tumors demonstrated evidence of growth before treatment. The technical aspects of the operation for the localization of melanomas, the treatment planning and the techniques of proton irradiation have been described in detail previously8- 11 and will be briefly summarized. One to two weeks prior to irradiation treatments the patients are admitted to the Massachusetts Eye and Ear Infirmary for the suturing of four 2.5 mm diameter tantalum rings which serve as reference markers for radiographic alignment of the tumor with the proton beam. The conjunctiva is incised, and the tumor is localized by transillumination and/or indirect ophthalmoscopy. The edges of the tumor are marked with a blue pencil, and the four tantalum rings are sutured to the sclera outlining the tumor. No operation is necessary for lesions confined to the ciliary body. In these tumors a light beam coaxial with the central axis of the proton beam is used to position the tumor relative to the beam. A three-dimensional treatment planning computer pro-
•
PROTON IRRADIATION
gram 12 facilitates the selection of the appropriate fixation angle, which is chosen to minimize irradiation of the lens, optic disc and fovea. (Figs 1-3) Patient immobilization is achieved with a bite block and individually made contoured plastic masks mounted into a frame on the headholder. Patients are set up radiographically and monitored during treatment with a video camera to assure that they are fixating on the predetermined point. A fluoroscopic system provides intratreatment verification of the position without reentering the room. Currently each patient receives a total dose of 70 GyE (7000 rad) a quantity equal to the dose (in Gray) multiplied by a factor of 1.1, which is our estimate of the radiobiological effectiveness of the proton beam relative to Cobalt 60 irradiation. The dose is delivered in five equal treatments during a period of eight to ten days. For purposes of analysis, we divided the melanomas into small, medium, large, and extra-large categories according to the diameters and heights of the tumors. Tumors with diameters of 10 mm or less and height of 2 mm or less were considered small. Medium tumors were larger in either diameter or height than the small melanomas and not more than 15 mm in diameter and 5 mm in height. Large tumors were larger in either diameter or height than the medium melanomas and not more than 20 mm in diameter and 10 mm height; and extra-large tumors were either larger than 20 mm in diameter or more than 10 mm in height. The diameters of the tumors were measured with calipers on the sclera after the tumor margins were localized for the placement of the tantalum rings, and the height was measured ultrasonographically. Only 12 eyes (5%) had small tumors, 99 (41%) eyes had medium size tumors and 130 (54%) had large and extra-large melanomas. Of the tumors that grew under observation, 9 were small, 35 medium, 18 large and 5 extra-large at the time of treatment. In 60 eyes the anterior margin of the tumor involved the ciliary body, in 67 it was anterior to the equator but without involvement of the ciliary body, and in 114 it was posterior to the equator. More large tumors were located anteriorly (P < 0.0001 ). In all but five patients the choroid was involved. Four involved only the ciliary body and one involved only the iris and the filtration angle. In six eyes, definite extrascleral extension of the tumor was present. In three ciliary body melanomas the extension was seen on clinical examination, and in three choroidal melanomas extension was found during the operation for tumor localization. Follow-up examinations were performed six weeks after treatment and then every three to four months during the first year. The patients were then reexamined every six months to a year. Fundus photography, fluorescein angiography and ultrasonography were performed at varying intervals for documentation of tumor regression. Chest roentgenography and liver function tests were recommended annually. The mean length of followup in this series was 21 months and the median 15 months. For purposes of analysis the follow-up period 285
OPHTHALMOLOGY
•
FEBRUARY 1985
•
VOLUME 92
•
NUMBER 2
Fig I. "Beam's eye view" of the eye of a patient with a 13 mm diameter tumor whose most posterior extent was 2 1h disc diameters from the macula. The magenta oval is the outline of the treatment aperture made to define the beam edges. The lens (in blue) is totally excluded from the aperture, as is the optic disc and nerve (yellow). The diagonal magenta line is the plane (seen edge on) of the dose distribution shown in Figure 2. Fig 2. Lines of constant dose (iso-dose contours) in a plane through the eye (defined in Fig I) which passes through the tumor and upper lens. The red line corresponds to 99% of the full dose (69 GyE); green to 90% (63 GyE); yellow to 50% (35 GyE) and cyan to 10% (7 GyE). The part of the lens shown in this figure is the portion which lies on the viewing side of the plane of calculation, and in that plane the lens is outside the 10% iso-dose contour. Fig 3. Iso-dose contours for the same treatment, shown on the retinal surface, as in a wide angle fundus picture.
was broken into two intervals: "short" in 164 eyes who were followed up to two years, and "long" in 77 eyes observed for longer than two years.
RESULTS TUMOR REGRESSION
Seventy-two of 77 (94%) tumors with "long" followup of more than two years and 107 of 164 tumors (65%) from the "short" follow-up group have decreased in size on clinical, photographic, and/or ultrasonographic evaluations (Figs 4, 5). The average height of tumors measured by ultrasound before and after treatment according to their size and duration of follow-up is shown in Figure 6. The differences in tumor height before and 286
after treatment for the short follow-up group were statistically significant for the medium size group (P < 0.001) the large size group (P < 0.001) and the extralarge group (P < 0.01 ). For the long follow-up group statistically significant differences in tumor height before and after treatment were found in medium (P < 0.001) and large tumors (P < 0.001). For the group of small tumors with short follow-up and groups of small and extra large tumors with long follow-up, differences in tumor height were present although not statistically significant; this could be due to the small number of patients in these groups (5 small with short follow-up, 4 small with long follow-up, and 3 extra large with long follow-up). In addition, the Wilcoxon rank sum test was performed in order to determine if there was a difference in change of tumor height between short and long periods of
GRAGOUDAS, et at
•
PROTON IRRADIATION
Fig 4. Large peripheral ciliochoroidal melanoma. Top row, before treatment (left, 30° fundus picture, right, wide-angle photograph of the same tumor). Second row, !6 months after proton irradiation (left, 30° fundus picture, right, wide-angle photograph of same tumor.) Visual acuity remains 20/30 32 months after treatment. Fig 5. Large posterior choroidal melanoma partially involving the fovea. Third row, before treatment (left, 30° fundus picture, right, wide-angle photograph). Bottom row, same tumor four months after proton irradiation (left, 30° fundus picture, right, wide-angle photograph). Visual acuity improved from C.F. to 20/100.
follow-up. According to this test, the change in tumor height before and after treatment was significantly different between short and long follow-up periods for patients with medium and large size tumors (P < 0.05 and P < 0.001, respectively). Again, few numbers of
patients in the small and extra-large tumor size groups may have precluded statistical significance. In the majority of the patients, the tumor regressed between four months and one year (range, 1 month-2 years). Disappearance of the lesion or formation of a flat scar was
287
OPHTHALMOLOGY
•
FEBRUARY 1985
observed in a small number of eyes. Continuous regression has been observed even four years after treatment but at a slower rate after the first year post-irradiation. Tumor growth post-irradiation was observed in two patients. In one the eye was enucleated and in the other the marginal increase was treated with argon laser photocoagulation; 6 months after photocoagulation and 13 months after proton treatment, this tumor continues to regress. Resolution of the secondary serous retinal detachments was usually the earliest finding. The earliest onset of resolution was one week and the latest two years. In two eyes the retina is still totally detached more than two years post-treatment. Detachments can occasionally increase in size during the first few months after treatment, but eventually resolve. Major fluoroangiographic changes after irradiation were destruction of choroidal and/or tumor vessels and cessation of fluorescein leakage.
•
VOLUME 92
•
NUMBER 2
->2yrs follow-up follow-up
---~2yrs
Extralarge ......
..c
·-:r:
Large
0\ Q)
\..
0
E
12
Medium
Q)
0\ 0
Small
\.. Q)
t---
:::.
VISUAL ACUITY
=--..:
"(
The pre-treatment and most recent post-irradiation visual acuities of the treated eyes are shown in Table 1. In 19 eyes the most recent visual acuity was not known. (These patients were not seen by us after treatment and the visual acuity was not included in the follow-up information furnished by the referring physician). The level of visual acuity improved in 28 eyes, remained the same in 119 and deteriorated in 75 eyes. The changes in visual acuity in relation to the size of the tumors are shown in Table 2. Worse post-treatment visual acuity was associated with larger tumors. Extra-large tumors were more likely to have worse post-treatment visual acuity compared to small and medium size tumors. The relationship between visual acuity status and the duration of follow-up time is shown in Table 3. No statistically significant difference was observed between the two groups, but patients with shorter follow-up were more likely to have better post-treatment vision ( 16%) than patients with longer follow-up (7% ). Changes in visual acuity in relation to the location of the tumors are shown in Table 4. Twenty-two patients with tumors 3 mm or less from the fovea and/or the optic nerve showed improvement of visual acuity. Only three of
0
**
After Before Treatment
Fig 6. Average height of tumors before and after treatment, according to tumor size and duration of follow-up.
these cases had follow-up more than two years. Five were followed from one to two years and 14 less than a year. For purposes of analysis we also divided the treated eyes into two groups: ( 1) eyes with most recent visual acuity of 20/200 or better and (2) eyes with most recent visual acuity worse than 20/200. Worse than 20/200 visual acuity was associated with larger tumors and tumors close to the disc and fovea (P < 0.0 1): however, no significant difference was observed between visual acuity status of the two groups and duration of followup time. COMPLICATIONS
Surgical complications from the suturing of the tantalum rings have been relatively few. Transient diplopia
Table 1. Visual Acuities Before and After Proton Irradiation Post-treatment Visual Acuity 20/15-20/40
HM-NLP
20/200-CF
20/50-20/100
Total
Initial Visual Acuity
N
(%)
N
(%)
N
(%)
N
(%)
N
(%)
20/15-20/40 20/50-20/100 20/200-CF HM-NLP
88 17 0 0
(84) (16) (0) (0)
16 15 10 0
(39) (37) (24) (0)
14 8 15 1
(37) (21) (39) (3)
10 15 12 1
(26) (39) (32) (3)
128 55 37 2
(57) (25) (17) (1)
105
(100)
41
(100)
38
(100)
38
(100)
222
(100)
Total
CF = count fingers; HM = hand motions; NLP = no light perception.
288
GRAGOUDAS,
et al •
PROTON IRRADIATION
Table 2. Post-treatment Visual Acuity and Size of Tumors Size Medium
Small
Large
Extra-large
Total
Visual Acuity
N
(%)
N
(%)
N
(%)
N
(%)
N
(%)
Better Same Worse
2 10 0
(17) (83) (0)
16 57 19
(17) (62) (21)
7 50 39
(7) (52) (41)
3 2 17
(14) (9) (77)
28 119 75
(13) (53) (34)
Total
12
(100)
92
(100)
96
(100)
22
(100)
222
(100)
has been observed in a small number of patients with tumors close to the fovea. In these cases part of the inferior oblique muscle was disinserted for accurate placement of the rings. One patient developed orbital cellulitis postoperatively and was treated successfully with antibiotics. Lid epitheliitis, epilation, and epiphora from punctum occlusion are observed in patients whose eyelids could not be retracted completely from the irradiation field. In one patient a Jones tube was inserted in the canaliculus for control of epiphora. Epithelial keratopathy has developed in a small number of cases and usually responds to artificial tears. One patient needed to use a soft contact lens. Lens changes from irradiation have developed in 22 patients with large choroidal or ciliary body melanomas in which the lens received substantial irradiation. Successful cataract extraction was performed in five eyes. Rubeosis iridis and neovascular glaucoma are the most serious complications and were observed in eight patients with large or extra-large tumors. One of these eyes became phthisical after cyclocryotherapy and subsequent cataract extraction. Vitreous hemorrhage was observed in eight eyes with large or extra-large tumors and cleared in all but two. Twenty-nine eyes developed radiation retinopathy involving the fovea with capillary closure, telangiectasia, microaneurysm formation, hemorrhages, exudates, cystoid edema and vascular sheathing. No association was found between development of radiation retinopathy and length of follow-up (P = 0.5).
Ten eyes were enucleated following proton beam irradiation. Six because of secondary glaucoma (four from angle neovascularization, and two from angle closure due to total retinal detachment), one because of progression of a preexisting retinal detachment, (this latter surgical procedure was premature since we have observed cases which showed resolution of serous detachments after initial progression), and one because of total loss of vision from optic atrophy and the presence of uveitis. The remaining eye was enucleated because of continued tumor growth. The light and electron microscopic findings in four of these cases have been previously published. 13•14
Table 3. Most Recent Visual Acuity and Follow-up
Table 4. Most Recent Visual Acuity and Location of Tumors
METASTASES
Thirteen patients developed metastases. Ten died from their disease and three remain alive and are being treated with chemotherapy. The development of metastases ranged from 4 to 50 months after treatment and in six patients occurred less than two years from irradiation. The liver was involved in all but one patient who developed skin and lung metastases. The diagnosis was confirmed by biopsy or autopsy in nine patients and by clinical tests in four. Seven patients had large, five had extra-large, and one had a medium size melanoma. All but the medium size melanoma had their anterior margin anterior to the equator. Two of the patients with metastases had bilateral melanomas. In one patient enucleation followed by exenteration of one eye was
Location of Tumor Edge
Follow-up Short*
Total
Longt
Visual Acuity
N
(%)
N
(%)
N
(%)
Better Same Worse
23 79 45
(16) (54) (30)
5 40 30
(7) (53) (40)
28 119 75
(13) (53) (34)
147
(100)
75
(100)
222
(100)
Total
* Eyes followed up to two years. tEyes followed longer than two years.
Visual Acuity Better Same Worse Total
:s:3 mm from Disc and/or Fovea
>3 mm from Disc and/or Fovea
N
(%)
N
(%)
22 59 53
(16) (44) (40)
6 60 22
(7) (68) (25)
28 119 75
(13) (53) (34)
134
(100)
88
(100)
222
(100)
Total N
(%)
289
OPHTHALMOLOGY
•
FEBRUARY 1985
performed before proton irradiation of his fellow eye. The other patient had proton therapy of both eyes. One other patient who developed metastases had extrascleral extension before irradiation and another had abnormal liver function tests and liver scan but negative liver biopsy before treatment.
DISCUSSION The theoretical advantages of proton beam irradiation compared with other conservative methods for the management of uveal melanomas have been discussed in detail previously. 8- 11 The physical characteristics of charged particles such as protons or helium ionsnamely minimal scatter, tissue sparing at the entry site, increased dose at the end of range (Bragg peak of ionization), and the sharp fall off of dose at the end of the beam-offer the possibility of highly localized dose distributions. These superior dose distributions provided by charged particles offer two major clinical advantages in the field of ocular radiotherapy. First, large tumors can be treated because the overall irradiated volume is reduced and therefore the tolerance of the eye to treatment is increased. The second important clinical advantage is the ability to treat lesions close to critical structures like the macula and the optic nerve. In these locations the sharp reduction of dose outside the target volume with the use of protons permits the delivery of potentially tumoricidal doses to the lesion, with sparing of the vital and radiosensitive nearby normal structures. It is the Bragg peak, characteristic of heavy charged particles such as protons and helium ions, which allows a very uniform dose distribution throughout the tumor, as well as the margin of normal tissue around the lesion that is included for adequate coverage of the melanoma borders. For very posterior lesions, part of uninvolved ocular tissues receive some irradiation as the beam enters the eye; this is kept to a minimum by choosing an appropriate fixation angle during treatments, which brings the tumor close to the entrance of the beam. Our present study with three times the number of treated patients of our last report8 and longer follow-up continues to show that proton therapy is an attractive alternative treatment to enucleation for the management of uveal melanomas. More than half of our treated patients had large or extra-large tumors, and we now believe that close to 80% of all diagnosed melanomas can be treated with charged particle radiotherapy. The regression patterns of uveal melanomas after proton irradiation consist of rather slow shrinkage, and in the majority of tumors response to radiation can be observed as long as two years after treatment. We believe it probable that such regressing tumors have been "sterilized" (that is, have lost their capacity for reproduction), since we haven't observed any tumor that showed growth after initial shrinkage. Thus, this slow regression is advantageous, since acute necrosis of especially large tumors would likely generate a destructive inflammatory 290
•
VOLUME 92
•
NUMBER 2
response. The possible mechanism of tumor regression after irradiation is a complex phenomenon that most likely involves multiple factors. Although a direct cytotoxic effect is probably the main mechanism of regression, destruction of the vascular supply to the tumor with gradual cell death could play an important role. 14 One hundred five of the 222 treated eyes (47%) had most recent visual acuity between 20/15 and 20/40. Visual acuity in eyes with tumors close to the macula and the optic nerve showed significant improvement in some cases, and significant deterioration in others. Resolution of serous retinal detachments and radiation retinopathy were the respective reasons for these changes. The data in Table 3 suggest that the visual acuity may become worse with longer follow-up, but these differences were not statistically significant. Longer follow-up and a multivariate analysis is needed in order to determine the independent effect of multiple variables on visual acuity outcome. Complications from proton beam irradiation continue to be few. The three-dimensional treatment planning computer program that has been employed during the last four years of this study has enabled us to minimize irradiation of the lens, the optic disc and the macula. In our series, ten eyes have been enucleated; nine because of complications and one for continued tumor growth. All eyes that developed complications necessitating enucleation had large or extra large tumors. Although the pathogenesis of rubeosis and neovascular glaucoma is unclear, angiogenetic factors elaborated by the intraocular tumors or the ischemic retina are possible stimuli to iris neovascularization. 1s-I? Radiation retinopathy in areas away from the tumor has been markedly reduced with the use of protons. This is entirely in accordance with expectation since the volume of normal tissues irradiated with the use of proton irradiation is minimized and the treatment volume almost coincides with the target volume. 18 Treatment of melanomas close to the macula and the optic disc resulted in higher incidence of visually destructive radiation vasculopathy. Cataract formation is an expected complication in some eyes since irradiation of a significant portion of the lens is accepted in some treatments particularly of large, anteriorly located lesions, in order to achieve adequate tumor coverage. In our series there have been five cataract extractions without any surgical complications. While preservation of the eye is the obvious advantage of proton beam irradiation compared with enucleation, the question of the possible influence of treatment on metastatic spread of disease remains unanswered. In our series of 240 patients with a 15-month median followup, 13 patients have developed metastases. A trial with patients randomly allocated to proton irradiation or enucleation is the ideal way to accurately determine the effect of proton irradiation versus enucleation on tumorrelated mortality, although some doubt the feasibility or propriety of such a study. We are currently comparing the survival of a subset of proton treated patients to
GRAGOUDAS, et al
groups of patients who underwent enucleation. Preliminary results suggest that proton irradiation has no deleterious effect on metastases. A number of other questions still remain unanswered regarding the use of proton beam radiotherapy for uveal melanomas. One of these issues is the optimum fractionation scheme and dose level for treatment. We presently use 70 GyE (equivalent to 7000 rad of Co60 radiation) delivered in five fractions over a period of eight days. Fractionation of dose is important in optimizing the relative response of tumor and normal tissue to irradiation and allows hypoxic cells to become oxygenated between fractions, thus rendering them more radiosensitive. We have chosen large fractions because clinical" studies in the radiotherapy of patients with skin melanomas have shown favorable results with a small number of relatively large dose fractions. 19 •20 We started with 50 GyE (5000 rad equivalent) in five fractions which was based on clinical data on treatment of superficial skin lesions. However, this dose was rapidly increased as our experience confirmed what our animal studies had predicted, 18 namely that the human eye tolerates quite large doses to small portions of the eye. We increased the dose, up to 100 GyE (10,000 rad equivalent) in some cases with the expectation that the probability of tumor control would increase with the increased dose. However, when we observed equally good results in the follow-up of patients treated with lower doses we returned to 70 GyE rad. We randomized a small number of patients between 70 to 90 GyE but no significant differences have been found with respect to tumor regression patterns. Consideration is presently given to a more highly fractionated regimen or lower dose for patients with very large tumors or tumors located adjacent to macula or optic disc. The tolerance of the optic disc to irradiation is not known. Although we inform all patients with lesions closer than 3 mm to the optic disc about the likelihood of optic atrophy, follow-up of a small number of eyes that received the full dose to the optic disc and observed for more than a year has shown no evidence of optic nerve damage. However, the number of treated eyes is small and the follow-up too short to draw definitive conclusions. In summary, proton irradiation of uveal melanomas appears to be feasible with the methods used and the achieved dose distributions are particularly advantageous for the treatment of these tumors. Although the long term results of this treatment are unknown, present knowledge indicates that proton irradiation has no deleterious effect on the likelihood of the development of metastases.
•
PROTON IRRADIATION
REFERENCES 1. Zimmerman LE, Mclean IW. An evaluation of enucleation in the management of uveal melanomas. Am J Ophthalmol1979; 87:74160. 2. Meyer-Schwickerath G, Vogel M. Treatment of malignant melanomas of the choroid by photocoagulation. Trans Ophthalmol Soc UK 1977; 97:416-20. 3. Peyman GA, Raichand M. Full-thickness eye wall resection of choroidal neoplasms. Ophthalmology 1979; 86:1024-36. 4. Shields JA, Augsburger JJ, Brady LW, Day JL. Cobalt plaque therapy of posterior uveal melanomas. Ophthalmology 1982; 89:1201-7. 5. Packer S, Rotman M. Radiotherapy of choroidal melanoma with iodine-125. Ophthalmology 1980; 87:582-90. 6. Lommatzsch PK. /)-irradiation of choroidal melanoma with 106Ru/ 06 Rh applicators; 16 years' experience. Arch Ophthalmol 1983; 101: 713-7. 7. Char DH, Castro JR, et al. Helium ion therapy for choroidal melanoma. Arch Ophthalmol1982; 100:935-8. 8. Gragoudas ES, Goitein M, Verhey L, et al. Proton beam irradiation of uveal melanomas; results of a 5v2-year study. Arch Ophthalmol 1982; 100:928-34. 9. Gragoudas ES, Goitein M, Koehler AM, et al. Proton irradiation of small choroidal melanomas. Am J Ophthalmol 1977; 83:665-73. 10. Gragoudas ES, Goitein M, Koehler AM, et al. Proton irradiation of choroidal melanomas; preliminary results. Arch Ophthalmol 1978; 96:1583-91. 11. Gragoudas ES, Goitein M, Verhey L, et al. Proton beam irradiation: an alternative to enucleation for intraocular melanomas. Ophthalmology 1980; 87:571-81. 12. Goitein M, Miller T. Planning proton therapy of the eye. Med Phys 1983; 10:275-83. 13. Zinn KM, Stein-Pokomy K, Jakobiec FA, et al. Proton-beam irradiated epithelioid cell melanoma of the ciliary body. Ophthalmology 1981; 88:1315-21. 14. Seddon JM, Gragoudas ES, Albert DM. Ciliary body and choroidal melanomas treated by proton beam irradiation; histopathologic study of eyes. Arch Ophthalmol 1983; 101:1402-8. 15. Gartner S, Henkind P. Neovascularization of the iris (rubeosis iridis). Surv Ophthalmol 1978; 22:291-312. 16. Gimbrone MA Jr, Leapman SB, Cotran RS, Folkman J. Tumor angiogenesis: iris neovascularization at a distance from experimental intraocular tumors. J Natl Cancer lnst 1973; 50:219-28. 17. Cappin JM. Malignant melanoma and rubeosis iridis; histopathological and statistical study. Br J Ophthalmol 1973; 57:815-24. 18. Gragoudas ES, Zakov NZ, Albert DM, Constable IJ. Long-term observations of proton-irradiated monkey eyes. Arch Ophthalmol 1979; 97:2184-91. 19. Habermalz HJ, Fischer JJ. Radiation therapy of malignant melanoma: experience with high individual treatment doses. Cancer 1976; 38: 2258-62. 20. Doss LL, Memula N. The radioresponsiveness of melanoma. lnt J Radial Oncol Bioi Phys 1982; 8:1131-4.
291
Abstract: Proton beam irradiation has been used for the treatment of 241 uveal melanomas over the past 7~ years. Twelve melanomas (5%) were small, 99 (41%) medium, 103 (43%) large and 27 (1%) extra-large melanomas. The mean length of follow-up was 21 months and the median 15 months. Ninety-four percent of the treated lesions with a follow-up more than two years and 65% of tumors with shorter follow-up showed regression. The most recent visual acuity was 20/40 or better in 47% and 20/100 or better in 66%. Ten eyes were enucleated because of complications (9) or continued tumor growth (1 ). Thirteen patients developed metastases from 4 to 50 months of treatment. Our data indicate that proton irradiation can be used to treat melanomas of various sizes and in a variety of locations, and preliminary results suggest that proton therapy has no deleterious effect on the likelihood of the development of metastases. [Key words: irradiation, melanoma, proton beam, treatment, uvea.] Ophthalmology 92:284-291, 1985
Enucleation has long been the usual treatment for posterior malignant uveal melanoma. The advantage of this method in terms of increased life expectancy has been recently challenged 1 and alternative forms of treatment have been used with increased frequency during the past few years. 2- 8 Photocoagulation2 and local resection3 have been used in selected instances, but radiotherapy remains the most widely used method. 4 - 8 There are two major radiotherapeutic techniques for the treatment of uveal melanomas. Radioactive plaques can be sutured on the sclera over the area of the tumor4 - 6 and particulate radiations such as protons8 and helium ions7 can be used. The advantages of the latter
From the Retina Service, Massachusetts Eye and Ear Infirmary Department of Ophthalmology, Harvard Medical School,* Department of Radiation Medicine, Massachusetts General Hospital,t Boston, and Harvard Cy· clotron Laboratory:j: Cambridge. Presented at the Eighty-eighth Annual Meeting of the American Academy of Ophthalmology, Chicago, Illinois, October 3D-November 3, 1983. Reprint requests to Evangelos S. Gragoudas, MD, Massachusetts Eye and Ear Infirmary, 243 Charles Street, Boston. MA 02114.
284
modalities are premised on better dose distributions between tumor and normal tissues. The dose delivered with protons or helium ions can be localized to the tumor more accurately, excluding from the radiation field all or some of the radiosensitive intraocular structures.7·8 Hopefully, this would improve the therapeutic ratio of local control versus complications. Proton beam irradiation has been used for the treatment of uveal melanomas at the Harvard Cyclotron since 1975, and its application has increased in frequency during the past few years. 8- 11 We report here the current results of treatment of 240 patients with uveal melanoma (241 eyes).
MATERIALS AND METHODS All patients were treated between July 1975 and December 1982. Follow-up data were analyzed up to June 1983. The 122 men and 118 women treated ranged in age from 14 to 83 years (mean and median age, 56 years). Most of the patients were in the sixth and seventh decades of life. All patients were white. The right eye
GRAGOUDAS, et al
was treated in 126 patients, the left in 113 and both eyes in one patient. The diagnosis was established in all cases through ( 1) clinical examination by multiple experienced consultants using indirect ophthalmoscopy and slit lamp biomicroscopy with contact lens, and (2) assessment of the results of ancillary tests, such as standard and wide angle fundus photography, fluorescein angiography and ultrasonography (both immersion A-scan and B-scan). Radioactive phosphorus (P32) uptake measurements were done routinely in the initial stages of the study but presently are performed only rarely, in cases where the diagnosis is still in doubt. All the patients received a complete systemic examination by an internist to exclude metastasis or other primary malignancy. Chest roentgenography and liver function studies were performed routinely; a liver scan was performed in selected cases. Patients with uveal melanomas up to 24 mm in largest diameter and 14 mm in height were treated. Largest diameters up to the axial length of the eye were observed in ciliochoroidal melanomas extending circumferentially. In general, larger lesions could be treated when located at the peripheral fundus, because the beam could enter the eye directly in the area of the tumor; thus the irradiated ocular volume can be kept smaller than in cases of posteriorly located tumors in which the beam enters the side of the globe opposite the lesion. Documentation of growth was required for small lesions (10 mm or less in diameter and 2 mm or less in height) in all but three cases which were treated early in this series. One hundred seventy-one patients presented with symptoms caused by their tumor. Those were in order of frequency: visual loss, photopsias, floaters, pain and/ or inflammation. Six patients had bilateral melanomas. Four had enucleation of one eye previously, one received proton irradiation to both eyes, and one had treatment of one eye with a Cobalt plaque two years prior to proton irradiation of his other eye. Ten patients presented with vitreous hemorrhage and 158 with retinal detachment at the time of initial examination. Sixty-nine tumors demonstrated evidence of growth before treatment. The technical aspects of the operation for the localization of melanomas, the treatment planning and the techniques of proton irradiation have been described in detail previously8- 11 and will be briefly summarized. One to two weeks prior to irradiation treatments the patients are admitted to the Massachusetts Eye and Ear Infirmary for the suturing of four 2.5 mm diameter tantalum rings which serve as reference markers for radiographic alignment of the tumor with the proton beam. The conjunctiva is incised, and the tumor is localized by transillumination and/or indirect ophthalmoscopy. The edges of the tumor are marked with a blue pencil, and the four tantalum rings are sutured to the sclera outlining the tumor. No operation is necessary for lesions confined to the ciliary body. In these tumors a light beam coaxial with the central axis of the proton beam is used to position the tumor relative to the beam. A three-dimensional treatment planning computer pro-
•
PROTON IRRADIATION
gram 12 facilitates the selection of the appropriate fixation angle, which is chosen to minimize irradiation of the lens, optic disc and fovea. (Figs 1-3) Patient immobilization is achieved with a bite block and individually made contoured plastic masks mounted into a frame on the headholder. Patients are set up radiographically and monitored during treatment with a video camera to assure that they are fixating on the predetermined point. A fluoroscopic system provides intratreatment verification of the position without reentering the room. Currently each patient receives a total dose of 70 GyE (7000 rad) a quantity equal to the dose (in Gray) multiplied by a factor of 1.1, which is our estimate of the radiobiological effectiveness of the proton beam relative to Cobalt 60 irradiation. The dose is delivered in five equal treatments during a period of eight to ten days. For purposes of analysis, we divided the melanomas into small, medium, large, and extra-large categories according to the diameters and heights of the tumors. Tumors with diameters of 10 mm or less and height of 2 mm or less were considered small. Medium tumors were larger in either diameter or height than the small melanomas and not more than 15 mm in diameter and 5 mm in height. Large tumors were larger in either diameter or height than the medium melanomas and not more than 20 mm in diameter and 10 mm height; and extra-large tumors were either larger than 20 mm in diameter or more than 10 mm in height. The diameters of the tumors were measured with calipers on the sclera after the tumor margins were localized for the placement of the tantalum rings, and the height was measured ultrasonographically. Only 12 eyes (5%) had small tumors, 99 (41%) eyes had medium size tumors and 130 (54%) had large and extra-large melanomas. Of the tumors that grew under observation, 9 were small, 35 medium, 18 large and 5 extra-large at the time of treatment. In 60 eyes the anterior margin of the tumor involved the ciliary body, in 67 it was anterior to the equator but without involvement of the ciliary body, and in 114 it was posterior to the equator. More large tumors were located anteriorly (P < 0.0001 ). In all but five patients the choroid was involved. Four involved only the ciliary body and one involved only the iris and the filtration angle. In six eyes, definite extrascleral extension of the tumor was present. In three ciliary body melanomas the extension was seen on clinical examination, and in three choroidal melanomas extension was found during the operation for tumor localization. Follow-up examinations were performed six weeks after treatment and then every three to four months during the first year. The patients were then reexamined every six months to a year. Fundus photography, fluorescein angiography and ultrasonography were performed at varying intervals for documentation of tumor regression. Chest roentgenography and liver function tests were recommended annually. The mean length of followup in this series was 21 months and the median 15 months. For purposes of analysis the follow-up period 285
OPHTHALMOLOGY
•
FEBRUARY 1985
•
VOLUME 92
•
NUMBER 2
Fig I. "Beam's eye view" of the eye of a patient with a 13 mm diameter tumor whose most posterior extent was 2 1h disc diameters from the macula. The magenta oval is the outline of the treatment aperture made to define the beam edges. The lens (in blue) is totally excluded from the aperture, as is the optic disc and nerve (yellow). The diagonal magenta line is the plane (seen edge on) of the dose distribution shown in Figure 2. Fig 2. Lines of constant dose (iso-dose contours) in a plane through the eye (defined in Fig I) which passes through the tumor and upper lens. The red line corresponds to 99% of the full dose (69 GyE); green to 90% (63 GyE); yellow to 50% (35 GyE) and cyan to 10% (7 GyE). The part of the lens shown in this figure is the portion which lies on the viewing side of the plane of calculation, and in that plane the lens is outside the 10% iso-dose contour. Fig 3. Iso-dose contours for the same treatment, shown on the retinal surface, as in a wide angle fundus picture.
was broken into two intervals: "short" in 164 eyes who were followed up to two years, and "long" in 77 eyes observed for longer than two years.
RESULTS TUMOR REGRESSION
Seventy-two of 77 (94%) tumors with "long" followup of more than two years and 107 of 164 tumors (65%) from the "short" follow-up group have decreased in size on clinical, photographic, and/or ultrasonographic evaluations (Figs 4, 5). The average height of tumors measured by ultrasound before and after treatment according to their size and duration of follow-up is shown in Figure 6. The differences in tumor height before and 286
after treatment for the short follow-up group were statistically significant for the medium size group (P < 0.001) the large size group (P < 0.001) and the extralarge group (P < 0.01 ). For the long follow-up group statistically significant differences in tumor height before and after treatment were found in medium (P < 0.001) and large tumors (P < 0.001). For the group of small tumors with short follow-up and groups of small and extra large tumors with long follow-up, differences in tumor height were present although not statistically significant; this could be due to the small number of patients in these groups (5 small with short follow-up, 4 small with long follow-up, and 3 extra large with long follow-up). In addition, the Wilcoxon rank sum test was performed in order to determine if there was a difference in change of tumor height between short and long periods of
GRAGOUDAS, et at
•
PROTON IRRADIATION
Fig 4. Large peripheral ciliochoroidal melanoma. Top row, before treatment (left, 30° fundus picture, right, wide-angle photograph of the same tumor). Second row, !6 months after proton irradiation (left, 30° fundus picture, right, wide-angle photograph of same tumor.) Visual acuity remains 20/30 32 months after treatment. Fig 5. Large posterior choroidal melanoma partially involving the fovea. Third row, before treatment (left, 30° fundus picture, right, wide-angle photograph). Bottom row, same tumor four months after proton irradiation (left, 30° fundus picture, right, wide-angle photograph). Visual acuity improved from C.F. to 20/100.
follow-up. According to this test, the change in tumor height before and after treatment was significantly different between short and long follow-up periods for patients with medium and large size tumors (P < 0.05 and P < 0.001, respectively). Again, few numbers of
patients in the small and extra-large tumor size groups may have precluded statistical significance. In the majority of the patients, the tumor regressed between four months and one year (range, 1 month-2 years). Disappearance of the lesion or formation of a flat scar was
287
OPHTHALMOLOGY
•
FEBRUARY 1985
observed in a small number of eyes. Continuous regression has been observed even four years after treatment but at a slower rate after the first year post-irradiation. Tumor growth post-irradiation was observed in two patients. In one the eye was enucleated and in the other the marginal increase was treated with argon laser photocoagulation; 6 months after photocoagulation and 13 months after proton treatment, this tumor continues to regress. Resolution of the secondary serous retinal detachments was usually the earliest finding. The earliest onset of resolution was one week and the latest two years. In two eyes the retina is still totally detached more than two years post-treatment. Detachments can occasionally increase in size during the first few months after treatment, but eventually resolve. Major fluoroangiographic changes after irradiation were destruction of choroidal and/or tumor vessels and cessation of fluorescein leakage.
•
VOLUME 92
•
NUMBER 2
->2yrs follow-up follow-up
---~2yrs
Extralarge ......
..c
·-:r:
Large
0\ Q)
\..
0
E
12
Medium
Q)
0\ 0
Small
\.. Q)
t---
:::.
VISUAL ACUITY
=--..:
"(
The pre-treatment and most recent post-irradiation visual acuities of the treated eyes are shown in Table 1. In 19 eyes the most recent visual acuity was not known. (These patients were not seen by us after treatment and the visual acuity was not included in the follow-up information furnished by the referring physician). The level of visual acuity improved in 28 eyes, remained the same in 119 and deteriorated in 75 eyes. The changes in visual acuity in relation to the size of the tumors are shown in Table 2. Worse post-treatment visual acuity was associated with larger tumors. Extra-large tumors were more likely to have worse post-treatment visual acuity compared to small and medium size tumors. The relationship between visual acuity status and the duration of follow-up time is shown in Table 3. No statistically significant difference was observed between the two groups, but patients with shorter follow-up were more likely to have better post-treatment vision ( 16%) than patients with longer follow-up (7% ). Changes in visual acuity in relation to the location of the tumors are shown in Table 4. Twenty-two patients with tumors 3 mm or less from the fovea and/or the optic nerve showed improvement of visual acuity. Only three of
0
**
After Before Treatment
Fig 6. Average height of tumors before and after treatment, according to tumor size and duration of follow-up.
these cases had follow-up more than two years. Five were followed from one to two years and 14 less than a year. For purposes of analysis we also divided the treated eyes into two groups: ( 1) eyes with most recent visual acuity of 20/200 or better and (2) eyes with most recent visual acuity worse than 20/200. Worse than 20/200 visual acuity was associated with larger tumors and tumors close to the disc and fovea (P < 0.0 1): however, no significant difference was observed between visual acuity status of the two groups and duration of followup time. COMPLICATIONS
Surgical complications from the suturing of the tantalum rings have been relatively few. Transient diplopia
Table 1. Visual Acuities Before and After Proton Irradiation Post-treatment Visual Acuity 20/15-20/40
HM-NLP
20/200-CF
20/50-20/100
Total
Initial Visual Acuity
N
(%)
N
(%)
N
(%)
N
(%)
N
(%)
20/15-20/40 20/50-20/100 20/200-CF HM-NLP
88 17 0 0
(84) (16) (0) (0)
16 15 10 0
(39) (37) (24) (0)
14 8 15 1
(37) (21) (39) (3)
10 15 12 1
(26) (39) (32) (3)
128 55 37 2
(57) (25) (17) (1)
105
(100)
41
(100)
38
(100)
38
(100)
222
(100)
Total
CF = count fingers; HM = hand motions; NLP = no light perception.
288
GRAGOUDAS,
et al •
PROTON IRRADIATION
Table 2. Post-treatment Visual Acuity and Size of Tumors Size Medium
Small
Large
Extra-large
Total
Visual Acuity
N
(%)
N
(%)
N
(%)
N
(%)
N
(%)
Better Same Worse
2 10 0
(17) (83) (0)
16 57 19
(17) (62) (21)
7 50 39
(7) (52) (41)
3 2 17
(14) (9) (77)
28 119 75
(13) (53) (34)
Total
12
(100)
92
(100)
96
(100)
22
(100)
222
(100)
has been observed in a small number of patients with tumors close to the fovea. In these cases part of the inferior oblique muscle was disinserted for accurate placement of the rings. One patient developed orbital cellulitis postoperatively and was treated successfully with antibiotics. Lid epitheliitis, epilation, and epiphora from punctum occlusion are observed in patients whose eyelids could not be retracted completely from the irradiation field. In one patient a Jones tube was inserted in the canaliculus for control of epiphora. Epithelial keratopathy has developed in a small number of cases and usually responds to artificial tears. One patient needed to use a soft contact lens. Lens changes from irradiation have developed in 22 patients with large choroidal or ciliary body melanomas in which the lens received substantial irradiation. Successful cataract extraction was performed in five eyes. Rubeosis iridis and neovascular glaucoma are the most serious complications and were observed in eight patients with large or extra-large tumors. One of these eyes became phthisical after cyclocryotherapy and subsequent cataract extraction. Vitreous hemorrhage was observed in eight eyes with large or extra-large tumors and cleared in all but two. Twenty-nine eyes developed radiation retinopathy involving the fovea with capillary closure, telangiectasia, microaneurysm formation, hemorrhages, exudates, cystoid edema and vascular sheathing. No association was found between development of radiation retinopathy and length of follow-up (P = 0.5).
Ten eyes were enucleated following proton beam irradiation. Six because of secondary glaucoma (four from angle neovascularization, and two from angle closure due to total retinal detachment), one because of progression of a preexisting retinal detachment, (this latter surgical procedure was premature since we have observed cases which showed resolution of serous detachments after initial progression), and one because of total loss of vision from optic atrophy and the presence of uveitis. The remaining eye was enucleated because of continued tumor growth. The light and electron microscopic findings in four of these cases have been previously published. 13•14
Table 3. Most Recent Visual Acuity and Follow-up
Table 4. Most Recent Visual Acuity and Location of Tumors
METASTASES
Thirteen patients developed metastases. Ten died from their disease and three remain alive and are being treated with chemotherapy. The development of metastases ranged from 4 to 50 months after treatment and in six patients occurred less than two years from irradiation. The liver was involved in all but one patient who developed skin and lung metastases. The diagnosis was confirmed by biopsy or autopsy in nine patients and by clinical tests in four. Seven patients had large, five had extra-large, and one had a medium size melanoma. All but the medium size melanoma had their anterior margin anterior to the equator. Two of the patients with metastases had bilateral melanomas. In one patient enucleation followed by exenteration of one eye was
Location of Tumor Edge
Follow-up Short*
Total
Longt
Visual Acuity
N
(%)
N
(%)
N
(%)
Better Same Worse
23 79 45
(16) (54) (30)
5 40 30
(7) (53) (40)
28 119 75
(13) (53) (34)
147
(100)
75
(100)
222
(100)
Total
* Eyes followed up to two years. tEyes followed longer than two years.
Visual Acuity Better Same Worse Total
:s:3 mm from Disc and/or Fovea
>3 mm from Disc and/or Fovea
N
(%)
N
(%)
22 59 53
(16) (44) (40)
6 60 22
(7) (68) (25)
28 119 75
(13) (53) (34)
134
(100)
88
(100)
222
(100)
Total N
(%)
289
OPHTHALMOLOGY
•
FEBRUARY 1985
performed before proton irradiation of his fellow eye. The other patient had proton therapy of both eyes. One other patient who developed metastases had extrascleral extension before irradiation and another had abnormal liver function tests and liver scan but negative liver biopsy before treatment.
DISCUSSION The theoretical advantages of proton beam irradiation compared with other conservative methods for the management of uveal melanomas have been discussed in detail previously. 8- 11 The physical characteristics of charged particles such as protons or helium ionsnamely minimal scatter, tissue sparing at the entry site, increased dose at the end of range (Bragg peak of ionization), and the sharp fall off of dose at the end of the beam-offer the possibility of highly localized dose distributions. These superior dose distributions provided by charged particles offer two major clinical advantages in the field of ocular radiotherapy. First, large tumors can be treated because the overall irradiated volume is reduced and therefore the tolerance of the eye to treatment is increased. The second important clinical advantage is the ability to treat lesions close to critical structures like the macula and the optic nerve. In these locations the sharp reduction of dose outside the target volume with the use of protons permits the delivery of potentially tumoricidal doses to the lesion, with sparing of the vital and radiosensitive nearby normal structures. It is the Bragg peak, characteristic of heavy charged particles such as protons and helium ions, which allows a very uniform dose distribution throughout the tumor, as well as the margin of normal tissue around the lesion that is included for adequate coverage of the melanoma borders. For very posterior lesions, part of uninvolved ocular tissues receive some irradiation as the beam enters the eye; this is kept to a minimum by choosing an appropriate fixation angle during treatments, which brings the tumor close to the entrance of the beam. Our present study with three times the number of treated patients of our last report8 and longer follow-up continues to show that proton therapy is an attractive alternative treatment to enucleation for the management of uveal melanomas. More than half of our treated patients had large or extra-large tumors, and we now believe that close to 80% of all diagnosed melanomas can be treated with charged particle radiotherapy. The regression patterns of uveal melanomas after proton irradiation consist of rather slow shrinkage, and in the majority of tumors response to radiation can be observed as long as two years after treatment. We believe it probable that such regressing tumors have been "sterilized" (that is, have lost their capacity for reproduction), since we haven't observed any tumor that showed growth after initial shrinkage. Thus, this slow regression is advantageous, since acute necrosis of especially large tumors would likely generate a destructive inflammatory 290
•
VOLUME 92
•
NUMBER 2
response. The possible mechanism of tumor regression after irradiation is a complex phenomenon that most likely involves multiple factors. Although a direct cytotoxic effect is probably the main mechanism of regression, destruction of the vascular supply to the tumor with gradual cell death could play an important role. 14 One hundred five of the 222 treated eyes (47%) had most recent visual acuity between 20/15 and 20/40. Visual acuity in eyes with tumors close to the macula and the optic nerve showed significant improvement in some cases, and significant deterioration in others. Resolution of serous retinal detachments and radiation retinopathy were the respective reasons for these changes. The data in Table 3 suggest that the visual acuity may become worse with longer follow-up, but these differences were not statistically significant. Longer follow-up and a multivariate analysis is needed in order to determine the independent effect of multiple variables on visual acuity outcome. Complications from proton beam irradiation continue to be few. The three-dimensional treatment planning computer program that has been employed during the last four years of this study has enabled us to minimize irradiation of the lens, the optic disc and the macula. In our series, ten eyes have been enucleated; nine because of complications and one for continued tumor growth. All eyes that developed complications necessitating enucleation had large or extra large tumors. Although the pathogenesis of rubeosis and neovascular glaucoma is unclear, angiogenetic factors elaborated by the intraocular tumors or the ischemic retina are possible stimuli to iris neovascularization. 1s-I? Radiation retinopathy in areas away from the tumor has been markedly reduced with the use of protons. This is entirely in accordance with expectation since the volume of normal tissues irradiated with the use of proton irradiation is minimized and the treatment volume almost coincides with the target volume. 18 Treatment of melanomas close to the macula and the optic disc resulted in higher incidence of visually destructive radiation vasculopathy. Cataract formation is an expected complication in some eyes since irradiation of a significant portion of the lens is accepted in some treatments particularly of large, anteriorly located lesions, in order to achieve adequate tumor coverage. In our series there have been five cataract extractions without any surgical complications. While preservation of the eye is the obvious advantage of proton beam irradiation compared with enucleation, the question of the possible influence of treatment on metastatic spread of disease remains unanswered. In our series of 240 patients with a 15-month median followup, 13 patients have developed metastases. A trial with patients randomly allocated to proton irradiation or enucleation is the ideal way to accurately determine the effect of proton irradiation versus enucleation on tumorrelated mortality, although some doubt the feasibility or propriety of such a study. We are currently comparing the survival of a subset of proton treated patients to
GRAGOUDAS, et al
groups of patients who underwent enucleation. Preliminary results suggest that proton irradiation has no deleterious effect on metastases. A number of other questions still remain unanswered regarding the use of proton beam radiotherapy for uveal melanomas. One of these issues is the optimum fractionation scheme and dose level for treatment. We presently use 70 GyE (equivalent to 7000 rad of Co60 radiation) delivered in five fractions over a period of eight days. Fractionation of dose is important in optimizing the relative response of tumor and normal tissue to irradiation and allows hypoxic cells to become oxygenated between fractions, thus rendering them more radiosensitive. We have chosen large fractions because clinical" studies in the radiotherapy of patients with skin melanomas have shown favorable results with a small number of relatively large dose fractions. 19 •20 We started with 50 GyE (5000 rad equivalent) in five fractions which was based on clinical data on treatment of superficial skin lesions. However, this dose was rapidly increased as our experience confirmed what our animal studies had predicted, 18 namely that the human eye tolerates quite large doses to small portions of the eye. We increased the dose, up to 100 GyE (10,000 rad equivalent) in some cases with the expectation that the probability of tumor control would increase with the increased dose. However, when we observed equally good results in the follow-up of patients treated with lower doses we returned to 70 GyE rad. We randomized a small number of patients between 70 to 90 GyE but no significant differences have been found with respect to tumor regression patterns. Consideration is presently given to a more highly fractionated regimen or lower dose for patients with very large tumors or tumors located adjacent to macula or optic disc. The tolerance of the optic disc to irradiation is not known. Although we inform all patients with lesions closer than 3 mm to the optic disc about the likelihood of optic atrophy, follow-up of a small number of eyes that received the full dose to the optic disc and observed for more than a year has shown no evidence of optic nerve damage. However, the number of treated eyes is small and the follow-up too short to draw definitive conclusions. In summary, proton irradiation of uveal melanomas appears to be feasible with the methods used and the achieved dose distributions are particularly advantageous for the treatment of these tumors. Although the long term results of this treatment are unknown, present knowledge indicates that proton irradiation has no deleterious effect on the likelihood of the development of metastases.
•
PROTON IRRADIATION
REFERENCES 1. Zimmerman LE, Mclean IW. An evaluation of enucleation in the management of uveal melanomas. Am J Ophthalmol1979; 87:74160. 2. Meyer-Schwickerath G, Vogel M. Treatment of malignant melanomas of the choroid by photocoagulation. Trans Ophthalmol Soc UK 1977; 97:416-20. 3. Peyman GA, Raichand M. Full-thickness eye wall resection of choroidal neoplasms. Ophthalmology 1979; 86:1024-36. 4. Shields JA, Augsburger JJ, Brady LW, Day JL. Cobalt plaque therapy of posterior uveal melanomas. Ophthalmology 1982; 89:1201-7. 5. Packer S, Rotman M. Radiotherapy of choroidal melanoma with iodine-125. Ophthalmology 1980; 87:582-90. 6. Lommatzsch PK. /)-irradiation of choroidal melanoma with 106Ru/ 06 Rh applicators; 16 years' experience. Arch Ophthalmol 1983; 101: 713-7. 7. Char DH, Castro JR, et al. Helium ion therapy for choroidal melanoma. Arch Ophthalmol1982; 100:935-8. 8. Gragoudas ES, Goitein M, Verhey L, et al. Proton beam irradiation of uveal melanomas; results of a 5v2-year study. Arch Ophthalmol 1982; 100:928-34. 9. Gragoudas ES, Goitein M, Koehler AM, et al. Proton irradiation of small choroidal melanomas. Am J Ophthalmol 1977; 83:665-73. 10. Gragoudas ES, Goitein M, Koehler AM, et al. Proton irradiation of choroidal melanomas; preliminary results. Arch Ophthalmol 1978; 96:1583-91. 11. Gragoudas ES, Goitein M, Verhey L, et al. Proton beam irradiation: an alternative to enucleation for intraocular melanomas. Ophthalmology 1980; 87:571-81. 12. Goitein M, Miller T. Planning proton therapy of the eye. Med Phys 1983; 10:275-83. 13. Zinn KM, Stein-Pokomy K, Jakobiec FA, et al. Proton-beam irradiated epithelioid cell melanoma of the ciliary body. Ophthalmology 1981; 88:1315-21. 14. Seddon JM, Gragoudas ES, Albert DM. Ciliary body and choroidal melanomas treated by proton beam irradiation; histopathologic study of eyes. Arch Ophthalmol 1983; 101:1402-8. 15. Gartner S, Henkind P. Neovascularization of the iris (rubeosis iridis). Surv Ophthalmol 1978; 22:291-312. 16. Gimbrone MA Jr, Leapman SB, Cotran RS, Folkman J. Tumor angiogenesis: iris neovascularization at a distance from experimental intraocular tumors. J Natl Cancer lnst 1973; 50:219-28. 17. Cappin JM. Malignant melanoma and rubeosis iridis; histopathological and statistical study. Br J Ophthalmol 1973; 57:815-24. 18. Gragoudas ES, Zakov NZ, Albert DM, Constable IJ. Long-term observations of proton-irradiated monkey eyes. Arch Ophthalmol 1979; 97:2184-91. 19. Habermalz HJ, Fischer JJ. Radiation therapy of malignant melanoma: experience with high individual treatment doses. Cancer 1976; 38: 2258-62. 20. Doss LL, Memula N. The radioresponsiveness of melanoma. lnt J Radial Oncol Bioi Phys 1982; 8:1131-4.
291