Expert Opinions on Uveal Melanoma: Insights from the 58th Ophthalmic Oncology Group Meeting

The clinical management of uveal melanoma is complex, with decision-making influenced by numerous factors, including patient preferences and health status, tumor size and anatomic extent, genetic status and cytomorphology, current scientific evidence, local resources, and physician experience. There are also regional variations in practice and disease characteristics. For example, prognostic testing based on gene expression profiling is more prevalent in the USA, while plaque brachytherapy using ruthenium-106 is more commonly employed in Europe and Asia [1‒3].

Evidence from randomized clinical trials is not available for every situation, leaving physicians to make decisions based on limited evidence. In such cases, understanding the distribution of opinions among experts in the field can be invaluable. These reports can guide clinicians by reflecting the majority opinion in specific clinical situations and demonstrating that multiple options may be viable where expert views diverge.

In preparation for the 58th Ophthalmic Oncology Group (OOG) meeting in Stockholm, Sweden, held from June 6–9, 2024, a questionnaire was distributed to individuals on the OOG mailing list. The respondents, representing experts with varying levels of experience, provided their insights on ten current topics in uveal melanoma through anonymous responses. The aim of this article was to explore and report the distribution of opinions among European experts in ophthalmic oncology on these ten topics. It is important to note that this study does not intend to establish guidelines, medicolegally binding documents, or instructions endorsed by the OOG or any other entity.

A questionnaire was developed, designed to be answered on a computer or smartphone. A link to this questionnaire was sent to the OOG email list on May 17, 2024, with a reminder sent on May 30, 2024. The questionnaire was closed on June 7, 2024, coinciding with the presentation of the results by the first author at the OOG meeting in Stockholm, Sweden. A description of how the questionnaire was developed and a complete list of the questions, including all response options, as well as the raw data from the questionnaire responses, are available in the online supplementary material (for all online suppl. material, see https://doi.org/10.1159/000541016). The questionnaire was fully anonymized, with no collection of name, sex, age, location, identification number, email address, other contact details, photos, or any other data about the respondents, and participation was entirely voluntary.

Statistical significance was defined as p < 0.05, and all p values were two-sided. Questionnaire responses were summarized as absolute numbers, proportions of the total number of respondents, or means with standard deviations (SDs). For categorical variables, we used the χ² test to analyze contingency tables. For comparisons of ordinal or continuous variables between two groups, we employed the Mann-Whitney U test. To control for the family-wise error rate across multiple comparisons within each question category, p values were adjusted using the Holm-Bonferroni method. All statistical analyses were conducted using GraphPad Prism (version 10.2.3 for macOS) and the tidyr package in R (R Core Team, Vienna, Austria, 2022).

Between May 17, 2024, and June 7, 2024, 34 respondents completed the questionnaire, which was distributed via the OOG email list. The email list consisted of 309 addresses, of which 5 were undeliverable and 8 were duplicates, resulting in 296 valid recipients. This yielded an overall response rate of 11% (34 out of 296). However, the list included professionals not directly responsible for the clinical management of uveal melanoma patients (e.g., physicians primarily treating retinoblastoma, ocular surface and orbit tumors, general oncologists, geneticists, preclinical researchers, biomedical scientists, nurses, and students), who were unlikely to report diagnosing or treating more than one new case of uveal melanoma annually. It is estimated that approximately 125 clinically active respondents involved in the primary treatment of uveal melanoma patients received the questionnaire, leading to an estimated response rate of 27%.

Among the 34 respondents, 9 (26%) had 0–5 years of experience in ophthalmic oncology postresidency; 7 (21%) had 6–10 years of experience; 5 (15%) had 11–20 years of experience; and 13 (38%) had more than 20 years of experience. Regarding the annual number of new patients with uveal melanoma that the respondents diagnose and/or treat, 8 (24%) respondents saw approximately 1–25 new patients, 5 (15%) saw 26–50, 5 (15%) saw 51–75, 4 (12%) saw 76–100, and 12 (35%) saw more than 100. The mean time to complete the questionnaire was 7 min and 31 s (SD 11 min and 18 s).

Question: “For a typical patient with a dome- or mushroom-shaped melanoma located away from the optic disc and with a diameter less than 16 mm, what is the maximum thickness at which plaque brachytherapy using ruthenium-106 is advisable? Please select one option.”

The mean value for maximum recommendable tumor thickness for ruthenium-106 plaque brachytherapy was 5.4 mm (SD 1.7, median 6, min 0, max 8). Excluding two respondents who responded “0 mm,” the mean value was 5.7 mm (SD 1.1, median 6, min 3, max 8). There was no significant difference in the maximum recommendable tumor thickness between respondents with ≤20 years of experience and those with >20 years of experience (Mann-Whitney U test, p = 0.48, Fig. 1a). Similarly, no significant difference was observed between those treating ≤50 versus >50 new patients annually (Mann-Whitney U test, p = 0.66, Fig. 1b).

Questions: “Does your institution employ a minimum scleral dose for plaque brachytherapy? Please select one option.” And “If your institution uses a minimum scleral dose for plaque brachytherapy, what is the specified dose level? (in gray)” Seven respondents (21%) indicated that they always use a minimum scleral dose for plaque brachytherapy, fifteen respondents (44%) reported that they never use it, and twelve respondents (35%) stated that they use it selectively, depending on specific situations (e.g., when using ruthenium-106 but not iodine-125, or for peripheral rather than central tumors).

Among the nineteen respondents who reported using a minimum scleral dose either always or selectively, the mean minimum scleral dose was 307 Gy (SD 139, median 350, min 20, max 700). There was no significant difference in the reported minimum scleral dose between respondents with ≤20 years of experience and those with >20 years of experience (Mann-Whitney U test, p = 0.24, Fig. 2a), nor between those treating ≤50 versus >50 patients annually (Mann-Whitney U test, p = 0.14, Fig. 2b).

Question: “Patients may be excluded from radiological surveillance for metastasis based on one or more of the following criteria. Please select all applicable responses (multiple choice questions).”

Regarding criteria for excluding patients from radiological surveillance for metastases, twenty-three respondents (68%) indicated that surveillance should be conducted for all patients, irrespective of primary tumor characteristics. Nine respondents (26%) suggested that it may be excluded for patients with disomy 3; six respondents (18%) for those with an EIF1AX mutation; and six respondents (18%) for patients in AJCC stage I. Interestingly, three respondents (9%) stated that surveillance may be excluded for patients with BAP1-mutated tumors. More than one choice was possible. The experience level of respondents (≤20 years vs. >20 years) did not significantly influence their responses regarding radiological surveillance (χ² test, p = 0.77). Similarly, no significant difference was found between those managing ≤50 versus >50 new cases per year (χ² test, p = 0.94).

Question: “For a newly identified choroidal melanocytic lesion measuring 6 mm in diameter and 1.5 mm in thickness, under which conditions is immediate treatment advised without waiting for evidence of growth? Please select one option.”

Respondents were asked under which conditions immediate treatment is advised for a hypothetical, newly identified choroidal melanocytic lesion measuring 6 mm in diameter and 1.5 mm in thickness, without waiting for evidence of growth. Twenty-five respondents (74%) indicated that treatment may be recommended based on the presence of at least one of the following risk factors: presence of orange pigment/hyperautofluorescence, subretinal fluid, a dome-shaped appearance, ultrasound evidence of hollowness, or symptomatic vision loss. One respondent (3%) suggested that treatment may be recommended when any single risk factor is present. Seven respondents (21%) recommended treatment when two risk factors are present, eleven respondents (32%) when three are present, and six respondents (18%) when four or five risk factors are present. Nine respondents (26%) did not recommend treatment based solely on the number of risk factors, advising that treatment should only be considered upon observation of lesion growth. Respondents could select only one option.

The experience level of respondents did not significantly influence their recommendations regarding conditions for immediate treatment of a small melanocytic lesion without waiting for evidence of growth (χ² test, p = 0.28). Similarly, there was no significant difference between respondents managing ≤50 versus >50 new cases per year (χ² test, p = 0.24).

When responses were treated as ordinal, with “treatment is not recommended based solely on the number of risk factors present. Treatment should only be considered upon observation of lesion growth” ranked as the first group, followed by recommendations based on 1, 2, 3, or 4+ risk factors, there was no significant difference between respondents with ≤20 years or >20 years of experience (Mann-Whitney U test, p = 0.14, Fig. 3a), nor between those managing ≤50 or >50 new cases per year (Mann-Whitney U test, p = 0.31, Fig. 3b).

Question: “After a diagnosis of uveal melanoma is confirmed, approximately how long does it typically take for definitive treatment (e.g., plaque brachytherapy, enucleation, proton beam irradiation) to begin at your institution? Please select one option.”

Thirteen respondents (38%) indicated that, at their institution, the time between diagnosis and treatment (e.g., plaque brachytherapy, enucleation, proton beam irradiation) of uveal melanoma was typically within 2 weeks. Nineteen respondents (56%) reported a timeframe of 2 weeks to a month, two respondents (6%) indicated one to 2 months, and no respondents reported a duration of more than 2 months. There were no significant differences in the reported delays between diagnosis and treatment based on the experience level of respondents (Mann-Whitney U test, p = 0.06, Fig. 4a) or the number of new cases managed annually (Mann-Whitney U test, p = 0.37, Fig. 4b).

Questions: “At what stage of nonmetastatic disease do you recommend drawing blood for the analysis of circulating tumor DNA (ctDNA) or circulating tumor cells (CTCs)? Please select one option.” and “At what stage of metastatic disease do you recommend drawing blood for the analysis of ctDNA or CTCs? Please select one option.”

Twenty-three respondents (68%) indicated that sampling of ctDNA or CTCs from peripheral blood was not recommended for any patient with nonmetastatic disease. However, five respondents (15%) believed it was advisable for high-risk patients (e.g., those with monosomy 3, gain of 8q, BAP1 mutation, SF3B1 mutation, large tumors), and six respondents (18%) thought it was recommendable for all patients when available.

In cases of metastatic disease, twenty respondents (59%) stated that sampling of ctDNA or CTC was not recommended for any patient. Five respondents (15%) recommended it only for patients undergoing immunotherapy, while nine respondents (26%) believed it was advisable for all patients when available.

The experience level of respondents (≤20 years vs. >20 years) did not significantly influence the recommendations for sampling ctDNA or CTC in either nonmetastatic or metastatic patients (χ² test, p = 0.60 and p > 0.99, respectively). Similarly, there were no significant differences between those managing ≤50 versus >50 new cases per year (for nonmetastatic patients, p = 0.48; for metastatic patients, p = 0.38).

In this article, we report the outcomes of a questionnaire distributed to experts in the field of ophthalmic oncology. Among the 34 respondents, over half (18 of 34, 53%) had 11 or more years of postresidency experience, and a similar proportion (21 of 34, 62%) were involved in the diagnosis or treatment of 51 or more new patients with uveal melanoma per year. This indicates that the responses were largely representative of true experts in the field.

Unequivocal evidence for the appropriate course of action in every aspect of care is rare, leading to reliance on personal experience, tradition, manufacturer recommendations, or field standards that have developed over many years through discussions, clinical observations, arbitrary decisions, and varying levels of consensus. The questions in the survey were designed to address areas where evidence, particularly from randomized clinical trials, is lacking. As such, it is not possible to evaluate the correctness of the responses against objective standards but rather to compare them with existing practices described in the literature. The maximum recommendable apical tumor thickness for ruthenium-106 plaque brachytherapy, reported as 5.7 mm, is consistent with recommendations from multiple authors across European and American centers [4, 5]. The American Brachytherapy Society’s 2014 guidelines recommend excluding tumors with T4e extraocular extension, basal diameters exceeding the limits of brachytherapy, blind painful eyes, and those with no light perception vision from plaque therapy, though they do not specify a tumor thickness threshold for choosing between ruthenium and iodine [6]. While tumor thickness has been established as an independent predictor of local tumor recurrence following ruthenium plaque brachytherapy, a previous multivariate analysis, which included tumor thickness and diameter as covariates, did not reveal a significantly increased hazard for tumors ≥5.5 mm treated with ruthenium-106 compared to iodine-125 (p = 0.08) [7, 8]. Kaiserman et al. [9] demonstrated that ruthenium-106 brachytherapy is an effective and safe alternative to enucleation for treating posterior uveal melanomas ≥8 mm in thickness, though they did not compare outcomes with other radioisotopes. Conversely, Shields and colleagues identified ruthenium-106 plaque brachytherapy as an independent risk factor for local tumor recurrence in tumors ≥8 mm in thickness [10].

There was a notable disparity in the use of a minimum scleral dose among the respondents. Brualla et al. [11] recommended avoiding ruthenium-106 brachytherapy for tumors exceeding 5 mm in height, citing the potential for very high doses to the sclera and surrounding healthy structures at risk. However, it is important to note that high doses to the sclera or the tumor base do not necessarily result in high doses to tissues lateral to the plaque margin, particularly with ruthenium plaques [12]. Some authors advocate for a minimum scleral dose of 350 Gy, arguing that choroidal atrophy should become visible within 6 months of posttreatment. This visible atrophy serves as ophthalmoscopic evidence of adequate plaque placement and ensures that tumoricidal doses reach all margins of tumors with low apical thickness [13]. Others use a minimum scleral dose of 700 Gy. [12]. There is no consensus on a maximum scleral dose, with doses as high as 1,500 Gy having been administered using ruthenium-106 without resulting in scleral necrosis [14].

There is considerable variability in recommendations for systemic surveillance for metastases across institutions. Some studies advocate for early interventions, such as hepatic metastasectomy or immunotherapy, as strategies to prolong survival, while others question the actual benefit of such surveillance, arguing that the observed survival extensions may result from lead-time bias, with surveillance detecting metastases before symptoms appear [15‒19]. In 2015, Singh reported on an anonymized poll conducted among ocular oncologists at the 16th Congress of the International Society of Ocular Oncology in Cleveland, OH, USA. The poll revealed that if prognostic testing indicated a low risk of metastasis, most respondents did not alter their surveillance protocols, whereas if a high risk was indicated, the majority increased the frequency of examinations [20]. There is now increasing recognition of the importance of early metastasis detection to facilitate therapeutic interventions, optimize patient eligibility for clinical trials, maintain good performance status, and limit metastatic burden [21‒23]. The 2021 National Comprehensive Cancer Network (NCCN) guidelines reflect this shift, recommending surveillance frequencies based on metastatic risk: low-risk patients should undergo examinations every 12 months; medium-risk patients every 6–12 months for 10 years; and high-risk patients every 3–6 months for 5 years, then every 6–12 months through year 10 [24].

Most respondents to our questionnaire recommended radiological surveillance for all patients. However, the rationale behind excluding surveillance for patients with BAP1-mutated tumors, as suggested by three respondents, remains unclear.

The treatment of small choroidal melanocytic lesions, with or without previous observation of growth, is a highly relevant and debated topic. In the Collaborative Ocular Melanoma Study (COMS), involving 188 small choroidal melanomas (1–3 mm in apical thickness and 5–16 mm in largest basal dimension) that were not treated at the time of enrollment, the Kaplan-Meier estimates of progression to a lesion eligible for treatment were 21% at 2 years and 31% at 5 years [25]. While it is uncommon for an ocular oncologist to refrain from treating a lesion they are convinced is a melanoma, regardless of its size, the real challenge lies in accurately diagnosing whether a lesion is indeed a melanoma before there is any evidence of growth. This distinction is particularly challenging in small lesions, where limited growth data may complicate the diagnosis. The critical question, therefore, is not whether to treat a melanoma of any size, but whether a small lesion that shows no growth over time is truly a melanoma. Extended observation without signs of growth naturally raises questions about the lesion’s classification. This ties into the debates on the metastatic potential of small melanomas and the prognostic importance of delays in the treatment [26‒28]. A clear majority (25 of 34 respondents, 74%) would treat a lesion with a largest basal diameter of 6 mm and a thickness of 1.5 mm without waiting for evidence of growth if risk factors were present. If three or more risk factors such as orange pigment, subretinal fluid, dome-shaped appearance, ultrasound hollowness, or symptomatic vision loss were present, 19 respondents (56%) would opt for treatment without observation for growth. Delays in treatment appear to be rare, with only two respondents (6%) reporting delays of more than 1 month between diagnosis and treatment.

Several recent studies have suggested that CTCs and ctDNA may be used to ascertain chromosomal copy number variations and for detection of metastases, respectively [29‒33]. However, these methods are not yet clinically implemented in most centers. Most experts did not consider this recommendable for patients with either nonmetastatic or metastatic disease. This likely reflects the need for more data on the utility of such sampling for patient management and their impact on survival before they can be considered clinically recommendable.

This study has several limitations. One significant limitation is the inherent bias that can arise from the perspectives of those who design the questionnaire. The way questions are phrased can subtly influence respondents, leading them to answer in a manner that aligns with the underlying assumptions or views of the questionnaire’s author. Additionally, the selection of which questions are included – and perhaps more importantly, which are excluded – can introduce another layer of bias, shaping the scope of responses and potentially steering the outcomes of the study. These factors can result in a survey that reflects not only the opinions of the respondents but also the biases of those who crafted the questions, ultimately influencing the data collected.

Second, the response rate was rather poor, estimated at 27% of clinically active ocular oncologists on the OOG mailing list, and the questionnaire was anonymized, and we could not ensure that each respondent had the desired clinical expertise to answer the questions accurately. The anonymity also means that we do not know how the respondents were distributed across different institutions, which could result in a skewed sample with overrepresentation from certain institutions, potentially biasing the results toward the views prevalent in those specific settings.

Third, the questionnaire format, which consisted primarily of prespecified answers (except for the minimum scleral dose question), did not allow us to collect the deeper reasoning, rationale, and causes behind the given responses. For instance, when addressing systemic surveillance, the questionnaire did not differentiate between various radiological modalities (e.g., ultrasonography, MRI). Additionally, there is a risk that some respondents may have misunderstood certain questions, which could have influenced the results. For example, three respondents recommended excluding surveillance for patients with BAP1-mutated tumors, and two respondents indicated that the maximum recommendable tumor thickness for ruthenium-106 plaque brachytherapy was “0 mm.” These responses could be attributed to various factors, such as a misunderstanding of the question (e.g., interpreting it as asking which patients should undergo systemic surveillance rather than the opposite), a belief that ruthenium brachytherapy is not recommended for any tumor, a lack of opinion on the appropriate threshold, or the unavailability of plaque brachytherapy at their institution.

Moreover, a larger sample of experts with a higher response rate would have been desirable. This would enable us to conclude with more confidence that the responses were representative of the broader field, illustrating the distribution of opinions, consensus points, and areas of disagreement.

This article reports the opinions of 34 European experts in ophthalmic oncology on various topics, including the maximum advisable tumor thickness for ruthenium plaque brachytherapy, criteria for exclusion from radiological surveillance, treatment of small choroidal melanocytic lesions without evidence of growth, and the sampling of ctDNA and CTC. The responses illustrate both the agreements and disparities in opinions among experts.

This study adhered to the tenets of the Declaration of Helsinki. According to the Swedish Ethical Review Act (2003:460), approval from the Swedish Ethical Review Authority and written informed consent were not required, as this study did not include any analyses of sensitive data; did not involve any treatments, examinations, or other interventions with patients or questionnaire respondents; did not affect subjects physically or psychologically; did not involve biological samples; and did not include any data or information that could be traced back to any individual.

The authors have no conflicts of interest to declare.

This study was not supported by any sponsor or funder.

G.S.: conceptualization, methodology, software, formal analysis, investigation, data curation, and writing – original draft. V.V.-D. and R.M.V.: writing – review and editing. A.M.: conceptualization and writing – review and editing.

The data generated from the questionnaire are fully presented in this article and its supplementary material. Further inquiries can be directed to the corresponding author.