Willem A. Manschot¹, William R. Lee² & Roel van Strik^3
1Institute of Pathology; ³ Institute of Epidemiology and Biostatistics, Erasmus University, Rotterdam, The Netherlands; ² Departments of Ophthalmology and Pathology, University of Glasgow, Glasgow, Scotland

Accepted 30 November1995

Key words: uveal melanoma, doubling times, first dissemination, irradiation, long term visual acuity, post-therapeutic quality of life

Abstract

Published data on growth rates of uveal melanomas and effects of treatment modalities raise important considerations. Dissemination from uveal melanomas starts after the tumour is larger than 7 mm diameter; growth from 7 to 10 mm diameter increases the risk of metastases incrementally to approximately 16% Estimations of tumour doubling times indicate that metastatic death before 8 years is nearly always due to pre-therapeutic dissemination so that the impact on survival by therapy can only be assessed thereafter. Histopathology on irradiated melanomas reveals that reproductive activity has not been suppressed and the anticipated (and unfavourable) risk of metastases is not balanced by poor postirradiation visual acuity. Also the psychological wellbeing of a patient with a functional fellow eye is better after primary enucleation. Conservative management is most appropriate for: small melanomas, patients with a short life expectancy, melanomas in a single functioning eye, and patients refusing enucleation.

   Optimum treatment of any form of neoplastic disease by modern therapeutic modalities should ideally be based on an accurate knowledge of the biological behaviour of the specific neoplastic process. With regard to ocular melanomas, there are serious gaps in knowledge concerning growth rates in situ and in the documentation of the time intervals between the presentation of the primary tumour and the appearance of metastases. The rarity of the tumour necessitates careful accumulation of data, so that an adequate analysis can be made of the claimed success of cure rates. Equally important, the side effects and the complications of treatment should be the subject of careful scrutiny. In this communication the literature on the relevant aspects of melanoma biology and current treatment modalities is critically analysed.

Tumour biology

Tumour doubling time

   Information on tumour doubling time of uveal melanomas was introduced in 1980 [1] and updated reviews were provided in 1987 [2] and in 1992 [3]. From this data it was possible to estimate the risk of metastatic disease in relation to the dimensions of the tumour. The application of meta analysis provided by Diener-West and associates [4] and Markowitz et al. [5] has indicated that a radical reappraisal of the current philosophy of management is required.

When do uveal melanomas metastasize?

Table 1. Metastatic death after enucleation small melanomas. Davidorf et al. [8].   Diameter < 10 mm < 7 mm 7-10 mm   Number cases 50 18 32   Number 10 yr follow-up 18 + 3 tumour death ? ?   Lost to follow-up 29 ? ?   Tumour-related death 5 (10%) 0 5 (16%)

Table 2. Metastatic death after enucleation small melanomas. Thomas et al. [9].   Diameter < 10 mm < 7 mm 7-10 mm   Number cases 65 42 23   Number 10 yr follow-up 27 (0 tumour death) ? ?   Lost to follow-up 38 ? ?   Tumour-related death 6 (9%) 0 6 (26%)

   In 1984, an editorial in the American Journal of Ophthalmology stated that 'it is unclear at what stage in its course a uveal melanoma develops the capacity to metastasize' [6]. However an analysis of the data then available in the literature provided evidence which indicated that metastatic disease does not occur until the tumour dimensions are greater than a diameter of 6-7 mm [7]. This conclusion was based on two separate studies of survival from small melanomas treated by enucleation [8, 9]. In both series [8, 9], metastatic death had not occurred in the <7 mm diameter groups, but had occurred in 10% [8] and 9% [9] in the < 10 mm diameter groups; it is noteworthy that the < 10 mm group also included the <7 mm cases. Recalculation of this data [8, 9] revealed that the percentage metastatic death rate in the group tumours measuring 7-10 mm in diameter groups appeared to be 16% and 26%, against zero in the <7 mm diameter groups (Tables 1,2). There are exceptions, however: one patient with a melanoma <7 mm had metastases, but had not yet died. In the literature thereafter, only one more case of metastatic death after enucleation in a patient with a < 7 mm melanoma has been reported [10]. This data strongly suggests that choroidal melanomas rarely metastasize when the tumours have a diameter less than 6-7 mm at the time of detection.

Death rates in patients with melanomas less than 10 mm diameter

   The previously cited editorial [6] also stated that patients with melanomas < 10 mm diameter had 'an excellent prognosis regardless of treatment; the five-year tumour-related mortality in this group is less than 5%' This statement, clearly contradicted by the observation shown in Tables 1 and 2, can be seriously challenged by the application of meta-analysis, which provides proof that five-year mortality rate in patients with melanomas < 10 mm diameter is 16% [4]. This implies that prolonged clinical observation of melanomas as they enlarge from 7 mm diameter to 10 mm diameter could result in the death of 16% of these patients. However it is too simplistic to suggest that all such patients with small pigmented tumours of 6-7 mm or less should be treated by immediate enucleation since many such tumours could be nævi and the majority of melanomas are not brought to the attention of ophthalmologists at this stage. Nonetheless on theoretical grounds, metastatic death could have been averted by complete eradication of the malignant neoplasm at the 6-7 mm stage and this important postulate should not be neglected.

Failure to appreciate the consequences of 'doubling time'

   The new approach to uveal melanomas, published in 1980 [1], with regard to their biological behaviour, was based on the published data on skin melanomas, for which tumour doubling times (Td's) had been calculated. Information derived from this literature showed that the time intervals between dissemination and metastatic death were widely spread and varied between 3 and 30 years.
   Subsequently numerous reports on tumour doubling times in uveal melanomas have been provided [11-14]; in only three out of thirtynine patients were the doubling times estimated to be shorter than 69 days.
   On theoretical grounds it has been suggested that metastatic tumour death (after the primary tumour has released cells into the circulation) will occur at a time interval derived from multiplication of the tumour doubling time by 35-40 [15]. This postulate (which does not assume that all metastases grow exponentially from the first cell division to a clinically detectable size, because the tumour doubling time represents a mean value) has not been contradicted in later studies [see reference 3], which have repeated and extended the fundamental work of Collins [15].
   Taking these two assumptions together, it may, therefore, be accepted for practical therapeutic and prognostic considerations in individual patients, that death from metastatic death in unlikely to occur before 35 x 69 days or no less than six years after dissemination of the first viable embolus of tumour cells.
   Furthermore, since the tumour doubling times of uveal melanomas may exceed 299 days, the time interval between dissemination and metastatic death may vary widely and range from 6 to 60 years. Crowley [16] reported a mean disease-free interval for cutaneous melanomas of 14.3 years and compared this with 22.3 years for primary ocular lesions. It is noteworthy that in this series [16], five out of twelve ocular melanoma patients had a tumour-free interval of 47, 45, 30, 28 and 27 years, respectively.
   The minimal latent interval of six years implies that metastatic death within 7-8 years after local therapy on the uveal melanoma is nearly always due to pretherapeutic dissemination. This is the only reason why a statistically significant impact on survival has never been and never will be established by a detailed analytical comparison of survival rates within 8-10 years after various conservative interventions, after primary enucleation or after observation only. It is therefore more likely that differences in survival rates after any given form of treatment will reflect the selection criteria for patients subjected to that specific form of treatment. The validity of this postulate is supported by the agreement in all reports, that within the first 8-10 years no significant difference in survival rate has ever been noted when various treatment modalities are compared [17-21] or when survival in 'promptly treated patients' is compared with those in whom 'treatment has been delayed' [22].
   Therefore, the essential requirements for valid statistical comparisons are:
    - a 10-15 year survival rate of all treated patients, and
    - accurate assessment of visual function in all living patients after a 5-year or more follow-up, without any selection bias.
   Many ophthalmic oncology institutes have employed conservative management in uveal melanomas for 15-20 years, but rather surprisingly, results after follow-up periods of 10 years or more of all treated patients have not been made available. Neglect in providing this fundamental information has been highlighted by Markowitz et al. [5], who emphasized that researchers, clinicians, journal editors, and reviewers should be more attentive to the completeness of reports from clinical research studies; attention to basic information regarding design and methods was strongly recommended. Statements made by these authors [5] concerning the overall quality of reporting (which have not improved over the last two decades) should not be ignored. Publication of unverifiable data still continues e.g. visual acuities between 6/24 and zero are unspecified, and statements such as 'stable disease' defined as 'an elevated remnant, remaining unchanged for at least one year' are equally vague [23].

Tumour biology in relation to ionising radiation.

   A consideration of the histopathological findings in 231 eyes containing irradiated melanomas was provided in 1992 [3]. Thereafter, another 170 such pathological studies have been reported [24-31]. A substantial number of irradiated tumours are treated successfully, but the actual percentage of loss of proliferative capacity of all stem cells will never be known. To compound the difficulty in assessment, a sufficient follow-up period for all irradiated patients may never be achieved and by the very nature of the treatment, radiotherapy is applied to 'masses' which are not histologically classified.
   The well-established radio-insensitivity of uveal melanomas is reflected in the absence of radiation-induced necrosis in about 40% of the treated tumours, and in the identification of well preserved tumour cells in about 95% of the specimens [3]. Appearances suggesting continuing tumour cell activity was noted in ruthenium106-irradiated melanomas in 55 of 56 cases [24]. On the other hand, microscopically complete tumour necrosis was found in 10 out of 25 large melanomas, measuring > 7-8 mm in height after iodine1 25-brachytherapy [25]. In another report unequivocal mitoses were found in 36% of tumours, treated with various types of radioactive plaques: enucleation was performed at an average of 29 months after therapy [26]. Mitoses were also reported in 20% of 25 ruthenium1 06-treated melanomas [27]. Mitoses after proton beam irradiation have been described in 9 (27%) of 33 melanomas [28]. Similar data was obtained from five proton beam-treated tumours in five patients [29] and in four of seven cases [30]. The two latter reports noted intervals between irradiation and enucleation of 36, 49 and 66 months. Mitoses in a growing tumour and mitoses of more than two years after irradiation must be regarded as significant indicators of the presence of viable tumour stem cells. Their presence indicates a failure of irradiation to abolish the proliferative capacity of the stem cells within the tumour.

Clinical implications

The importance of long-term visual acuity data

   Preservation of vision is the principle justification for conservative management of ocular melanomas. It is unfortunate therefore, that some of the radiotherapeutic and surgical centres which publish regularly, have been reluctant to present verifiable, detailed data on the retained visual acuity of all treated patients. In general, follow-up periods of visual function have been shorter than those for other data relevant to the treatment and further management of the patients. An extreme example of a great potential for bias is presented in a report [32] published in 1993, on 163 patients who had been treated between 1972 and 1991. This study mentions only a 'one-year (!) post-operatively retained visual acuity of all patients'. Moreover, vision had been measured with a pinhole aperture. The reported visual acuity is, therefore, in no sense representative of the post-therapeutic quality of life of these patients.

Table 3. Reported visual acuities (VA) or adverse visual events (AVE)*.     No. Follow-up VA or AVE   Lommatzsch (1986)a 33 227 5 years First VA >0.3:31 (14%)3 Last VA <0.2:83 (37%)3   Tjha-Heslinga (1993)a 23 49 1-6 years First VA >=0.25:9 (19%) Last VA <=0.25:19 (39%)   Linstadt (1988)b 34 189 median 26 months First VA >=0.1:164 (87%) Last VA <=0.1:89 (48%)   Kindy-Degnan (1989)b 35 279 3 years First VA >=0.5:173 (62%) Last VA >=0.5:27 (10%) First VA >0.1:238 (85.3%) Last VA >0.1:35 (12.5%)   Char (1993)b 36 86 4-5 years First VA >=0.5:26 (30%) Last VA >=0.5:16 (19%) First VA <=0.1:6 (7%) Last VA <=0.1:58 (67.4%)   Char (1993)c 36 98 4-5 years First VA >=0.5:29 (29%) Last VA >=0.5:23 (23.5%) First VA <=0.1:9 (9.2%) Last VA <=0.1:66 (67%)   Packer (1992)c 37 63 mean 65 months Total AVE: 49 (77.8%) Last VA <=0.1:35 (55%)   Fontanesi (1993)c 38 144 median 46 months First VA <0.1.29%3 Last VA <0.1:59%3   Guyer (1992)d 39 218 mean 40 months First VA <0.1.22 (10%)3 Last VA <0.1:86 (39.4%)3 a Ruthenium106; b Helium ions; c Iodine125; d Proton beam. * Adverse visual event: acuity decreased to <= 6/30 or >= 2 lines. 3 Data unverifiable or concerning selected groups of patients.

   An appropriate method of analysing the published accounts of post-therapeutic visual function would be to arrange the data according to various irradiation modalitis (Table 3).
   The lack of uniformity in the published information prevented the use of one scale in this table, so that proper comparisons could not be made and the data not added. It would be highly desirable if the Collaborative Ocular Melanoma Study (COMS) and those large centres, which do not participate in the COMS, could reach an agreement to use identical designs and methods for their future follow-up reports. This would enable an appropriate long-term appraisal of the data for retention of useful visual function, particularly with reference to different treatment modalities.
   It may be noted that some numbers and percentages in Table 3 do not correspond with the data provided in the original reports. Several authors used percentages which did not refer to all treated patients, but to selected sub-groups. As far as possible, the recalculated data in Table 3 refer to all treated patients. The information provided confirms that visual acuity has decreased to <6/60 in about 50% of the eyes, three years after irradiation using any modality. A continuing decrease in visual acuity of about 10% per annum has also been noted [40]. A randomized study [36] to compare the results of helium ion irradiation versus 1125-brachytherapy showed that visual acuity had decreased by >= 4 lines in 69% of each group. Another study on visual acuity after 1125-therapy [37] revealed that 49 (78%) of 63 patients had an acuity decreased to <= 6/30, or an acuity which had decreased 2 lines or more. It was concluded that the eye encounters significant morbidity after 1125-irradiation. Data on post-therapeutic visual acuity after ruthenium106, cobalt60 and proton beam irradiation have either not been published, or have been reported in an incomplete fashion, with short follow-up periods, which prevent comparison with other therapeutic modalities.
   After five years, some 70% or more of irradiated eyes in which data were verifiable had retained a visual acuity of <= 6/60 (Table 3). A case report [41] mentions a maintained reading vision 10 years after proton beam irradiation of a histologically proven macular melanoma. The generally poor visual result however questions the authenticity of the 'fight for sight' justification for conservative treatment. The anticipated incremental ratio of the 'unfavourable risk' of post-therapeutic dissemination from retained viable tumour stem cells against the 'benefit' of the conservative procedure cannot be assessed before ten post-therapeutic years, due to our inability to control metastatic disease.

The future quality of life after treatment

   There is little consideration in the literature on conservative management of the post-therapeutic physical and psychological stresses imposed on the treated patients. Considerable anxiety arises from frequent re-examination, re-treatment of recurrence, and the treatment of the complications after irradiation [29, 30, 37, 42, 43] or intraocular surgery [32]. This contrasts with patients treated by primary enucleation, who, moreover, are unaware of the anxious and depressing burden of harbouring an intraocular malignancy, which might continually have the capacity to disseminate and to cause a much feared metastatic death.
   By contrast, the impact of enucleation on the post-surgical quality of life has not been found to be significant in four vision-dependent activities: viz. working; driving, reading and television viewing in 62 (87%) of 71 patients [44]. Fifteen years after enucleation, 18/20 (90%) of patients retained the ability to drive and 25/26 (96%) retained the ability to read. Another comparative study on the same four post-enucleation vision-dependent activities as against those of brachytherapy [45] revealed that 48/51 (94%) of the patients after enucleation and 46/51 (90%) of those after irradiation - mean follow-up of respectively 89 and 87 months - reported no vision-related change in any of these four activities. It appears questionable, therefore, to burden a patient with a normally functioning fellow-eye with the incremental risk of preventable metastatic death by post-therapeutic dissemination from a uveal melanoma.
   An exemplary 25-35 year follow-up report [46] on 302 between 1943-1952 primary enucleated patients with a posterior uveal melanoma, with a follow-up percentage of 99.8%, revealed that 148 (49%) patients had been saved from metastatic death. In the intervening decades the figure of 49% should, in theory, have been markedly improved. Present diagnostic techniques provide earlier detection rates in small melanomas and survival rates after enucleation depend largely on the volume of the tumour at the time of enucleation. Knowledge that a malignancy has been removed completely is the predominant factor in the future psychological well being of many patients.

Back to the index