BritishJournal ofOphthalmology, 1991, 75, 718-721

Phototherapy of posterior uveal melanomas

Ian Favilla, William R Barry, Andrew Gosbell, Peter Ellims, Fabian Burgess

AbstractThe results of phototherapy on a small seriesof 19 patients with uveal melanomas are pre-sented. Laser radiation at a wavelength of 620to 630 nm was used in conjunction with a 5 to7 5 mg/kg dose ofhaematoporphyrin derivativeadministered 24 hours before treatment, withtotal doses/treatment averaging 960 J/cm'.Eleven patients received two treatments, andone received a third. The longest duration oftumour control to 30 September 1990 was 31months. Of the 19 patients treated six hadcomplete regression of the tumour, whileanother five had minor to significant regres-sion. A strong correlation between degree ofregression and degree of tumour pigmentationwas found, the lighter tumours respondingmuch better than darker tumours. There werealso strong indications that if a tumour did notrespond to the initial phototherapy it was veryunlikely that subsequent phototherapy wouldprovide any further benefit.

Department ofOphthalmology, MonashMedical Centre,Melbourne, AustraliaI FavillaW R BarryA GosbellP EllimsF BurgessCorrespondence to:Dr I Favilla, 15 Collins Street,Melbourne 3000, Australia.Accepted for publication10 May 1991

For over a hundred years enucleation has beenthe traditional method of treatment for posterioruveal melanomas, particularly for large tumours.However, its singular lack of success in increas-ing life expectancy' has led to alternativemethods of treatment, with radiotherapy beingthe most effective and widely used.2" Sincerecent reports have shown that enucleation andradiotherapy have essentially equivalent survivalrates," new forms of therapy such as photo-therapy need to be evaluated.

Phototherapy (PT) has been shown to destroyimplanted melanomas and retinoblastomas inanimals, with few adverse effects.89 The opera-tion of PT is based on the fact that malignanttissue selectively retains a photosensitiser such as

haematoporphyrin derivative (HpD). If thesensitiser is subsequently activated with light ofan appropriate wavelength, it releases high-energy radicals which cause tumour cell destruc-tion.'0 Although the use of light-sensitive sub-stances in the treatment of disease was describedas early as 4000 BC,"I it was first used success-

fully to treat basal cell carcinomas in 1905.12Since then it has been used to treat manydifferent forms of cancer with varying degrees ofsuccess and with minimal side effects.'3 14

The few reports of phototherapy for posterioruveal melanomas in humans have beenencouraging, despite the small numbers, shortfollow-up, and varied response. In one seriesseven posterior uveal melanomas were treatedwith one complete response in an amelanoticmelanoma of 9-4 mm height at the posterior polefollowed up for four years, with varied to zero

response in the others.'5 In another report, of 23patients, the greatest tumour reduction occurredin melanomas of less than 1000 cm3 volume, with

no adverse systemic side effects from HpD.'6Marked necrosis extending to the base of themelanoma has been reported in enucleated speci-mens following PT. 17HpD has been the photosensitiser most widely

used. It binds tightlyto serum proteins, includinglipoprotein carriers, and there is a linear relation-ship between cell volume and uptake.'I Animalexperiments have shown that it localises invascular structures of the eye (choroid, iris,retina) within 48 hours of intravenous adminis-tration, with lower concentrations in the sclera,and with none located within the cornea andlens.'9 In experimental tumours it is concen-trated in the vascular stroma (macrophages, mastcells, and possibly fibroblasts), with a lowerconcentration in the cells, and the earliesthistological change following PT is bleeding oftumour cells adjacent to the microvasculature.2'The actual mechanism by which tumour cells aredestroyed is not clear, though ischaemia,2122destruction of cell membrane functions, intra-cellular enzyme inhibition, and DNA damageprobably initiated by singlet oxygen production"and other radicals have been reported.24 Oculartoxicity studies have shown that its acute effectsare confined to the treated area, and the chronicchanges are non-progressive. '9 Its major sideeffect is skin photosensitisation for up to 80 daysfollowing administration.

This report describes the effects ofPT on a small series of choroidal melanomas.

Materials and methodsUnder a management protocol approved by theResearch (Advisory and Ethics) Committee atPrince Henry's Hospital, Melbourne, and withinformed consent, 19 patients with posterioruveal melanomas diagnosed by ophthalmologicalexamination, fluorescein angiography, and ultra-sonography were treated with PT after growthhad been demonstrated by an increase in volumein consecutive measurements (17 patients), or atthe patient's request (patients 76 and 80).Tumour size was assessed by volume in accord-ance with previously reported methods by a'blind' observer.25 26 Location of the anteriormargin of the tumour was graded as posteriorto the equator, equatorial, or anterior to theequator, and tumour pigmentation was gradedas deeply, moderately, or lightly pigmented asdetermined by indirect ophthalmoscopy andcolour photographs. All patients were examinedby an oncologist and received whole body CTscans. A full blood examination and renal andliver function tests were performed prior totreatment. Exclusion criteria were the presenceof clinical porphyria, metastases, pregnancy,allergic reactions to HpD, and tumours morethan 10 mm in height.

Patients were treated with a pulse-injected

718

on Septem

ber 10, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.75.12.718 on 1 Decem

ber 1991. Dow

nloaded from

Phototherapy ofposterior uveal melanomas

Table I Degree ofpigmentation, position, pretreatment height and volume, PT dose, degree ofregression, and other details for 19 patients treated with PT

First treatment (months) Second treatment (months) Third treatment (months)Pretreatment size

Position Total To 30 Sept 90 Total To 30 Sept 90 Total To 30 Sept 90 Degree ofPatient Colour of of Height Volume dose or to 2nd Deg dose or to 3rd Deg dose or since 3rd Deg regression Enucleated eyeno. pigment tumour mm mm' J/cm2 treatment reg JIcm2 treatment reg J/cm2 treatment reg overall cell-type

NR 1150 12-9NR 1410 12-6NRNR 2200 31-3MR 674 2-4SRNR 720 13 7NR 1280 22-1NR 710 2-8NR 1180 27-0NR 770 31-3MRSRSR 740 16 4SR 992 6-0CRCRCRCR

NRNR 1300 6 5

NRNR

NRNRNRMRSR

CRCR

NR=no regression. MR-=minor regression. SR=significant regression. CR=complete regression. Deg reg=degree of regression. *Patient dead.

intravenous infusion of HpD obtained from theQueen Elizabeth Hospital, South Australia,dosage 5-7 5 mg/kg body mass. For all patients,monochromatic light of 620-630 nm wavelengthand 1000 [tm beam width from a coherent argon/dye laser was delivered through the pupil on tothe surface of the melanoma via a Rodenstockpanfunduscope lens and/or Goldman three-mirror lens, in two sessions 24 and 48 hours afteradministration of HpD. For three patients addi-tional light was delivered transclerally by a 1 mm

diameter fibreoptic bundle applied to the tumourbase. The total dose in J/cm2 received by each

tumour was calculated from its basal surface area

measurement, or best estimate, and is listed inTable 1.

Eight patients with posterior uveal melanomaswhich showed either zero or insufficient responsereceived a second treatment (mean 10 months,ranging from 5 to 21 months following the initialtreatment). One patient received three treat-ments. All patients received systemic prednis-olone 10 mg/day, topical atropine 1%, andprednisolone 1% for four weeks followingtreatment.

All patients wore protective clothing,

Figure I Fluoresceinangiogram prior to PT(patient 8): early, middle,late frames.

Figure 2 Fluoresceinangiogram followingPT(patient 8): early, middle,lateframes.

22258

l641259

11112

7*16

116257672778

80132

DarkDarkDarkDarkDarkDarkModerateModerateModerateModerateModerateModerateModerateModerateModerateModerateModerateLightLight

PostPostPostPostEquPostPostPostEquPostPostAntPostPostPostPostPostEquEqu

4-16-63.33-34-42-02-65.35-24.45.54 04-71-23.94-72-65*08-9

111255175129167297610019321164955

151139382111145NA

1490130010001309201190980110092078026059511203207901140960960760

7-813 323-69-29.55.4

21 45 19 07-86-26-717 89018-411-327-223 51.9

Spindle cell

Spindle cell

Spindle cellEpithelioid

NRNR NR

NRNRNRSRNRNRNRMRSRSRSRCRCRCRCRCRCR

719

on Septem

ber 10, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.75.12.718 on 1 Decem

ber 1991. Dow

nloaded from

Favilla, Barry, Gosbell, Ellims, Burgess

ultraviolet-block skin creams, and avoided directsunlight for a minimum period of 45 daysfollowing treatment. They were regularlyreviewed, with measurement of visual acuitywith and without pinhole, slit-lamp biomicro-scopy, clinical photography, and serial measure-ment of the melanoma volume.

ResultsTumour response to treatment was classified intofour grades:

(1) Complete regression (CR): no evidence oftumour macroscopically to end of the currentfollow-up period (30 September 1990).

(2) Significant regression (SR): more than50% reduction of original tumour size to the endof the current follow-up period.

(3) Minor regression (MR): up to 50% reduc-tion of original tumour size to the end of thecurrent follow-up period.

(4) No regression (NR): no change or con-

tinued growth after treatment.Apart from one patient who died, none was

lost to follow-up. The results are summarised inTable 1. To illustrate the relation between thedegree ofpigmentation and the level ofregressionthe table is sorted in order ofpigmentation (fromdark to light) as the primary sort condition, withregression (from none to complete) as a second-ary criterion.

After one session of PT a complete response

was seen in four of the 19 patients (2 1%), while a

further six showed a partial response (3 1%).After a second application of PT the number ofcases showing a complete response increased tosix (31%), while the number showing a partialresponse remained at six (31%). As of 30September 1991 there was no sustained regres-sion or continued growth of tumour in eightpatients (42%).The longest duration oftumour control was 31

months to 30 September 1990. The average PTdose per treatment was 960 J/cm2 for the 31treatments given. One patient (patient 7) whohad MR following treatment died as a result ofhepatic metastases three years after diagnosisand two years after PT.The posterior uveal melanomas which regres-

sed following PT exhibited fluorescein angio-graphic evidence of tumour and choroidalvascular occlusion and elimination of fluoresceinleakage (Figs 1, 2).Ofthe eight patients showing NR, six received

two treatments, and one received three. Nopatient experienced any adverse side effect as aresult of multiple treatments. For those tumoursthat showed either NR or MR as a result of PT,neither a second nor third treatment producedfurther improvement except in one patient(patient 16). We now believe that tumour regres-sion occurred in this patient following the firsttreatment but was masked by a large overlyingserous retinal detachment. Two patients with aninitial CR of melanomas at the posterior polereceived a second treatment resulting in CR forrecurrences arising from the tumour edge, andextending into the macular area which had beendeliberately avoided in the initial treatment.Four eyes which showed NR, ofwhich three had

Table 2 Spearman rank correlation analysis parameters:relationship ofdegrees ofregression ofchoroidal melanomaswith pre-PT treatment height, volume, initial dose,cumulative dose, and degree ofpigmentation

Parameter(correlated with four levels ofdegree of regression) Corr. coeff. Significance

Height (1 mm inc.) -0 03 0-91 (NS)Volume (100 mm' inc.) -0-18 0-48 (NS)Initial PT dose (500 J/cm2 inc.) -0 07 0-79 (NS)Cumulative PT dose (500 J/cm2 inc.) -0 59 0-006 (S)Degree of pigmentation (3 levels) 0-71 0 0005 (S)

inc. =increment. NS=not significant at 5% level. S=significant.

second treatments, were enucleated. Histo-pathology, including transmission electronmicroscopy, showed active tumour cell growthin all eyes with extrascleral extension in one(patient 9).

Table 2 lists Spearman's rank correlationcoefficients and significance levels for therelationship between degree of regression withtumour height, volume, initial and cumulativePT doses, and degree of pigmentation. Therewas no significant correlation with height,volume, or initial PT dose. Cumulative doseshowed a strong negative correlation with degreeofregression, which implies that ifa tumour doesnot respond to initial PT it probably will notrespond to further treatment. Very strongcorrelation was found between degree of regres-sion and degree of pigmentation; the less thepigmentation, the greater the degree of regression.

DiscussionDespite the small size of this series the short-term results following PT were encouraging.The lack of correlation between tumour responseand its physical attributes, apart from the degreeof pigmentation, highlights the unpredictabilityof the PT response owing to multiple factorsincluding photosensitiser type, dosimetry andtissue uptake, and the total dose of light energy.Although HpD is the most commonly usedsensitiser for PT, it is not ideal. It maximallyabsorbs light at 390 nm wavelength; but becausethis wavelength has poor tissue penetration, theminor absorption peak at 630 nm wavelength,which penetrates tissues more effectively, butrequires relatively more energy, is generallyused. The poor response of the six deeplypigmented melanomas suggests that they mayhave received a lower than calculated total lightenergy dose. Transcleral delivery of light energyin addition to transpupillary delivery may over-come the problems of light absorption by thesedeeply pigmented melanomas.

Unfortunately, there is no satisfactory quanti-tative method of determining in vivo the concen-tration and localisation ofHpD in tumour tissue.Porphyrins are usually measured by the level offluorescence emission; however HpD is notparticularly fluorescent.27 Experimental workhas shown that the concentration of HpD is ofthe order of several micrograms per gram oftumour tissue at dosages of 3 0 mg/kg at 48-72hours after injection. It may peak after a fewhours and remain essentially constant for days.Variation occurs between different tumours inone patient and between different patients with

720

on Septem

ber 10, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.75.12.718 on 1 Decem

ber 1991. Dow

nloaded from

Phototherapy ofposterior uveal melanomas

the same kind of tumour. Moreover it maydecrease with irradiation.28 The poor correlationbetween tumour response to PT and volume,height, and energy dose may be due to variableuptake of HpD by tumour tissue. Newer photo-sensitisers may prove to be more suitable andovercome some of the problems ofHpD.At present there is no reliable method of

calculating the most appropriate dose of lightand HpD to obtain maximal tumoricidal activitywith minimal damage to normal tissue. Theeffective absorbed dose depends on the lightenergy density at the dose point, the concentra-tion of the photosensitising drug, and the con-centration of molecular oxygen.28 There isexperimental evidence that the depth of tumourkill is directly related to the dose of light energy,rather than the concentration of HpD.29 In thisseries higher doses of HpD (5-7-5 mg/kg)than previously reported for posterior uvealmelanomas' were used, and all tumours weretreated initially 24 hours after administration, aslaboratory data indicated this to be the optimaltime.28 Because light energy densities above 200J/cm2 significantly increase the depth of tumourkill,30 and eradication of posterior uvealmelanomas by light energy densities in excess of1500 J/cm2 has been reported,'6 we aimed atabout 1000 J/cm2; total light energy levelsaveraged 960 J/cm2 in our series.The fluorescein angiographic evidence of

tumour and choroidal vascular occlusion andelimination of fluorescein leakage for thosemelanomas which showed regression have alsobeen observed following radiotherapy.3' Thissupports the concept that ischaemia due toendothelial cell damage of tumour blood vesselsis a possible mechanism of action of PT.32

Sunburn persisting for one or two days andpresenting as an erythematous swelling of theskin, mainly of the face and hands, has been themajor side effect in all patients. The commonestocular complication in 14 patients (74%) follow-ing PT was a transient serous retinal detachmentoverlying the tumour, nine of which resolved insix weeks. This probably reflects the inflamma-tory response to the high laser energy levels used,and, in the five patients (26%) in whom thelocalised retinal detachment persisted, overlyingvitreous traction was associated. Two patients(10%) developed anterior uveitis which resultedin pupil block glaucoma and subsequent cataractin one and an associated posterior scleritis inthe other. A transient vitreous haemorrhageimmediately following PT occurred in threepatients (16%), all of which resolved withoutcomplication. Four eyes (2 1%) which showed noresponse to PT were enucleated, though no eyewas enucleated for complications of PT. Thepost-treatment visual function was worse in 15patients (79%). However, in 13 patients (68%)the tumour was located posteriorly and involved

the macula, and in two patients (10%) thetumour was located in the temporal equatorialposition, and the macula was involved second-arily from the retinal and choroidal vascularocclusion which followed PT.At present no globe preserving treatment is

entirely successful. The obvious advantages ofPT are its non-invasiveness and repeatability. Atthis stage ofdevelopment phototherapy may be asatisfactory treatment for small and medium sizelight to moderately pigmented posterior uvealmelanomas, and as a supplementary treatmentfollowing recurrences with other methods.

1 Zimmerman LE, McLean 1W. An evaluation of enucleation in themanagement of uveal melanoma. AmJ7 Ophthalmol 1979; 87: 741-60.

2 Stallard HB. Radiotherapy for malignant melanoma of the choroid. BrJOphthalmol 1966; 50:147-55.

3 Shields JA, Augsberger JJ, Brady LW, Day JL. Cobalt plaque therapy ofposterior uveal melanoma. Ophthalmology 1982; 89: 1201-7.

4 Char DH, Castro JR. Helium ion therapy for choroidal melanoma. ArchOphthalmol 1982; 100: 935-8.

5 Augsberger JJ, Gamel JW, Lauritzen K, Brady LW. Cobalt-60 plaqueradiotherapy vs enucleation for posterior uveal melanomas. Am J

Ophthalmol 1990; 109: 585-92.6 Adams KS, Abramson DH, Ellsworth RM, et al. Cobalt plaque versus

enucleation for uveal melanoma. Comparison of survival rates. BrJ7Ophthalmol 1988; 72: 494-7.

7 Lommatzsch PK. Results after P-irradiation (10Ru/'1Rh) of choroidalmelanomas: 20 years' experience. BrJ Ophthalmol 1986; 70: 844-51.

8 Sery TW, Dougherty TJ. Photoradiation of rabbit ocular malignantmelanoma sensitised with haematoporphyrin. Curr Eye Res 1984; 3:229-37.

9 Dougherty TJ. In: Kessel D, ed. Methods in porphyrin sensitisation. NewYork: Plenum, 1985: 313-28.

10 Dougherty TJ, Kaufman JE, Goldfarb A, et al. Photoradiation therapyfor the treatment ofmalignant tumours. CancerRes 1978; 38: 2628-35.

11 Forrest JB, Forrest HJ. Case report. Malignant melanoma arising duringtherapy for vitiligo. J Surg Oncol 1980; 13: 337-40.

12 Jesionek A, Tappeiner HV. Zur Behandlung der Hautcarcinome mitfluorescierenden Stolle. Dtsch Arch KlinMed 1905; 82: 223-7.

13 Lipson RL, Blades EJ, Gray MJ. Haematoporphyrin derivative fordetection and management of cancer. Cancer 1967; 42: 623-9.

14 Forbes IJ, Ward AJ, Jacka AJ, et al. Multidisciplinary approach tophototherapy of human cancers. In: Andreoni A, Cubeddu R, eds.Porphyrin localization and treatment of tunours. New York: Plenum,1984.

15 Murphree AL, Cote M, Gomer CJ. The evolution of photodynamictherapy in the treatment of intraocular tumours. Photochem Photobiol1987; 46: 919-23.

16 Bruce RA. Evaluation of haematoporphyrin photoradiation therapy totreat choroidal melanomas. Lasers SurgMed 1984; 4: 59-64.

17 Tse DT, Dutton JJ, Weingeist TA, Hermsen VM, Kersten RC.Haematoporphyrin photoradiation therapy for intraocular and orbitalmalignant melanoma. Arch Ophthalmol 1984; 102: 833-8.

18 Bohmer RM, Morstyn G. Uptake of Hpd by normal and malignant cells:effects of serum pH, temperature and cell size. Cancer Res 1985; 45:5328-34.

19 Gomer CJ, Jester JV, Razum NJ, Szirth BC, Murphree AL. Photo-dynamic therapy of intraocular tumours. Examination of haemato-porphyrin derivative distribution and long term damage in rabbitocular tissue. CancerRes 1985; 45: 3718-25.

20 Bugelski PJ, PorterCW, Dougherty TJ. Autoradiographic distribution ofhaematoporphyrin derivative in normal and tumour tissue of themouse. CancerRes 1981; 41: 4606.

21 Packer A, Tse D, Hayreh SS, et al. Haematoporphyrin photoradiationtherapy for iris neovascularisation. Arch Ophthalmol 1984; 102: 1193-7.

22 Selman SH, Kreimer-Birnbaum MK, Klaunig JE, et al. Blood flow intransplantable bladder tumours treated with haematoporphyrinderivative and light. CancerRes 1984; 44: 1924-7.

23 Weishaupt KR, Gomer CJ, Dougherty TJ. Identification of singletoxygen as the cytotoxic agent in photoinactivation of a murine tumour.Cancer Res 1976; 36: 2326-9.

24 Oosterhuis JA. Haematoporphyrin derivative photoradiation ofmalignantmelanoma in the anterior chamber ofthe rabbit. In: Oosterhuis JA, ed.Ophthalmic tumours. Dordrecht: Junk, 1985.

25 Gosbell A, Barry W, Favilla I. Computer-aided volume measurement ofchoroidal melanomas. Aust NZJ Ophthalmol 1987; 15: 349-57.

26 Gosbell A, Barry W, Favilla I, Turner I. Computer-aided area measure-ment of ocular images. Aust NZJ7 Ophthalmol 1987; 15: 341-8.

27 Kessel D. In vivo fluorescence of tumours after treatment with derivativesof haematoporphyrin. Photochem Photobiol 1986; 44: 107-8.

28 Profio AE, Wuld LR, Sarnaik J. In: Kessel D, ed. Methods in porphyrinsensitization. New York: Plenum, 1985: 35-41.

29 Kaye AH, Morsteyn G, Brownbill D. Adjuvant high dose photoradiationtherapy in the treatment of cerebral glioma: a phase I-II study. J

Neurosurg 1987; 67: 500-5.30 Kaye AH, Hill JS. A review of photoradiation therapy in the management

of central nervous system tumours. AustNZ Surg 1988; 58: 767-80.31 Gragoudas ES, Goitein M, Verhey L, et al. Proton beam irradiation of

uveal melanomas: results of 51/2-year study. Arch Ophthalmol 1982;100:928-34.

32 Henderson BW, Dougherty TJ, Malone PB. Studies on the mechanism oftumour destruction by photoradiation therapy. In: Doiron DR,Gomer CJ, eds. Porphyrin localization and treatment of tumours. NewYork: Liss, 1984: 601-12.

721

on Septem

ber 10, 2021 by guest. Protected by copyright.

http://bjo.bmj.com

/B

r J Ophthalm

ol: first published as 10.1136/bjo.75.12.718 on 1 Decem

ber 1991. Dow

nloaded from