Advances in Positron Emission Tomography/Computed Tomography for Diagnosing and Managing Primary Posterior Uveal Melanoma

Introduction

Posterior uveal melanoma is the most common type of primary intra-ocular malignancy in adults. It is a highly aggressive tumor with a high risk of distant metastasis. The AJCC classification of posterior uveal melanoma (choroidal and ciliary body) has an essential role in establishing prognosis and staging of melanoma [1]. Early detection of posterior uveal melanoma and implementation of appropriate treatments are critical for patient survival. Given the intra-ocular location of uveal melanoma, its detection on routine ophthalmological examination can sometimes be challenging, making it difficult to distinguish it from an easily detectable cutaneous melanoma. This places particular emphasis on various imaging techniques, both those specific to ophthalmology and general radiological methods, for detecting choroidal melanomas. The array of methods that have been used in ophthalmology for years, including ultrasonography and fluorescein and indocyanine green angiography, has recently expanded into relatively novel and minimally invasive techniques, such as optical coherence tomography (OCT) and optical coherence tomography angiography (OCTA). However, conventional computed tomography (CT) and magnetic resonance imaging (MRI) still play important roles in evaluating choroidal melanoma, particularly in cases of extra-ocular extension [2].

Although positron emission tomography/computed tomography (PET/CT) plays a vital role in oncology, it is rarely used to detect or monitor ocular neoplasms. However, with technological progress and the growing availability of PET/CT, this method has been increasingly used in ophthalmic oncology [3–5]. PET/CT appears to be particularly useful for determining the tumor stage before treatment, establishing prognosis, and monitoring treatment outcomes.

This article reviews the role of PET/CT in diagnosis, staging, and monitoring treatment outcomes in patients with primary posterior uveal (choroidal and ciliary body) melanoma.

Posterior Uveal Melanoma

POSTERIOR UVEAL MELANOMA – INCIDENCE AND CLINICAL CHARACTERISTICS:

Posterior uveal melanoma is a highly aggressive intra-ocular neoplasm found in adults, and it is derived from choroidal melanocytes. The choroid is a highly vascularized tissue that forms the intermediate layer of the eye and is primarily responsible for nourishing the underlying retina. Uveal melanoma accounts for 3–5% of all melanomas [6], with 98% of cases diagnosed in Whites. The prevalence of uveal melanoma is estimated at 5.1–6 cases per million in the United States and 6–7.5 cases per million in Europe [7–9].

Although uveal melanoma can develop in adults of any age, it is most frequently diagnosed in patients aged 58–60 years, and slightly more often in men than in women [10,11]. Only 1% of uveal melanomas are detected in patients aged <20 years, with the remaining 99% found in those aged ≥21 years. Uveal melanoma in children is very rare [12,13].

Although the exact underlying mechanism of uveal melanoma development is unknown, the role of oxidative damage to the pigmented tissue of the eye has been emphasized, with the latter being modulated by the type and degree of pigmentation. Although uveal melanoma can develop within a pre-existing benign mass or choroidal nevus, recent evidence suggests that the risk of such a malignant transformation is low [6,14]. Recently, the role of genetics has been highlighted in the development of uveal melanoma. Eyes with uveal melanoma have been shown to harbor mutations in chromosomes 1,3,6, and 8 [15,16]. Chromosome 3 monosomy has been observed in 50–60% of aggressive uveal melanomas. Risk factors for uveal melanoma include choroidal nevi, freckles, light eye color, abnormal skin nevi, ota nevus, ocular melanocytosis, ultraviolet exposure, sun exposure, and outdoor activities [14].

The management of uveal melanomas depends on their location in the eye. Tumors are located at the periphery of the fundus outside the visual axis, and small lesions are usually asymptomatic in the early stages. Melanomas located closer to the macula and posterior pole of the eye, as well as large tumors and those associated with exudative retinal detachment, manifest with blurry vision, flashes, floaters, metamorphopsia, and visual field defects. Retinal detachment is more often observed in uveal melanomas >4 mm [14].

Uveal melanoma presents as a dome-shaped (75%), mushroom-shaped (20%), or diffuse mass (5%) [7]. While most uveal melanomas are pigmented (55%), amelanotic (15%) and mixed lesions (30%) have been reported [10]. Depending on cellular architecture, uveal melanomas are classified as epithelioid cells (>90% epithelioid cells), spindle cells (>90% spindle cells), or mixed lesions (>10% epithelioid cells and <90% spindle cells) [17].

At the time of diagnosis of uveal melanoma, the average basal tumor diameter is 11.3 mm, with a mean thickness of 4*–5 mm [7,18]. Depending on their diameters, uveal melanomas are classified as small (4–8 mm diameter and/or 1–2.4 mm thickness), medium (6–16 mm diameter and/or 2.5–10 mm thickness), or large (>16 mm diameter and/or >10 mm thickness) [14]. Tumors with large basal diameters have a high risk of metastatic spread, and smaller tumors are associated with improved survival rate and increasing incidence rate [19].

POSTERIOR UVEAL MELANOMA STAGING – AMERICAN JOINT COMMITTEE ON CANCER CLASSIFICATION:

The classification published by the American Joint Committee on Cancer (AJCC) includes 4 stages of posterior uveal melanoma (choroidal and ciliary body), from the least advanced (T1) to the most advanced (T4). Each stage is further classified into T category subclassification depending on the involvement of the ciliary body and extra-ocular spread or lack thereof [1]. Posterior uveal melanomas are most often diagnosed as T1 lesions, with T4 tumors being the rarest. Melanomas involving the ciliary body and showing extra-scleral extensions have a high risk of rapid metastasis. The AJCC classification of posterior uveal melanomas plays an essential role in clinical practice as a determinant of prognosis and melanoma staging (Tables 1–5).

POSTERIOR UVEAL MELANOMA – PROGNOSTIC FACTORS:

The prognosis of uveal melanoma is directly related to its cellular architecture and histological features. Epithelioid cell or mixed-cell melanomas, with high microvasculature density, extravascular matrix patterns such as closed loops, high mitotic index, and high proliferation index, are generally considered to have a poor prognosis [20]. The prognosis of spindle cell melanomas is generally better [16,21,22]. Another determinant of prognosis is the cytogenetic status of the tumor [23]. Melanomas harboring a mutation in the SF3B1 gene are more prone to metastatic spread, whereas the presence of EIF1AX mutation is associated with a relatively favorable prognosis [24]. In turn, the loss of the tumor suppressor gene BAP1 has been demonstrated to be associated with a rapid spread and markedly shorter survival [25]. Depending on the gene expression profile, uveal melanomas are classified into group I (low risk) and group II (high risk) lesions [26]. The co-existence of some chromosomal abnormalities, such as chromosome 1p deletion, monosomy 3, and chromosome 8q gain, is considered an unfavorable prognostic factor [27,28]. According to a large study, only 30% of patients with uveal melanomas had normal chromosomes 3 and 8, and the number of tumors with normal genetic status decreased significantly, with increasing AJCC stage and increasing tumor diameter as independent predictors [23]. AJCC stage III and abnormal copy numbers of chromosomes 3 and 8 were identified as significant predictors of poor prognosis. However, the prognostic value of the AJCC staging system decreased in tumors with normal genetic status.

Obviously, the lower the stage of the disease, the more favorable the prognosis. The risk of metastatic spread over 10 years increases with disease stage, from merely 15% for stage T1 to 63% for stage T4 [29]. The primary location of distant metastases is the liver (95%), followed by the lungs (28%), bones (18%), and skin (12%) [30]. The risk of metastatic spread is the lowest in the case of small melanomas up to 1 mm in thickness, which places particular emphasis on early detection of the disease.

POSTERIOR UVEAL MELANOMA – DIAGNOSTICS:

The first steps in evaluating patients with suspected uveal melanoma include routine slit-lamp biomicroscopy, indirect fundus ophthalmoscopy, and pupil dilation. Careful inspection of the distant periphery of the fundus is of utmost importance, as lesions located there are easily overlooked. Uveal melanomas are more likely to develop in the eye periphery than at the posterior pole. If an intra-ocular tumor with potentially malignant characteristics is detected, routine biopsy and histopathological examination are not recommended, given the risk of cancer cell spread and retinal injury. Moreover, uveal melanoma biopsy is frequently associated with bleeding, both around the tumor and within the corpus vitreum [20]. Bleeding can hinder diagnostic imaging and prevent accurate placement of a radioactive plaque during brachytherapy. However, this does not mean that a uveal melanoma biopsy should not be performed. In line with the most recent approach, a biopsy can be performed while establishing a primary diagnosis; however, it aims to obtain material for further genetic testing rather than to confirm the diagnosis. Genetic tests are conducted to determine the patient’s specific risk of metastasis and to plan an individualized treatment strategy [31].

If a lesion is detected in the fundus, the diagnostic process needs to be expanded to confirm or exclude the possibility that the mass is malignant. Additional diagnostic procedures may include both non-invasive methods, such as ultrasonography (B-scan ultrasonography, color and power Doppler), OCT, and OCTA, and invasive procedures with the use of contrast agents, such as fluorescein angiography (FA) and indocyanine green angiography (IGA) [2,32]. Methods for visualizing vasculature and perfusion within tissues are particularly helpful in evaluating uveal melanoma, as this tumor has a dense vascular network. CT and MRI are rarely used to detect intra-ocular lesions. Their role is limited primarily to confirming an extra-ocular expansion and distinguishing contrast-enhancing uveal melanomas from non-enhancing hemorrhagic choroidal detachment and highly calcified choroidal osteoma [8]. The role of CT and MRI increases, however, if the fundus is not accessible for examination because of optic center opacity, such as corneal leukoma, presence of blood in the anterior chamber, mature cataract, or ciliary hemorrhage. Under such circumstances, ophthalmological examination using methods other than ultrasonography is impossible.

Considering all the diagnostic options mentioned above, other methods are needed to examine patients with suspected posterior uveal melanoma. Modern oncological diagnostics cannot be performed without PET/CT. Thus, the question arises as to whether this method could be helpful in the evaluation of posterior uveal melanoma and its role therein.

Positron Emission Tomography

POSITRON EMISSION TOMOGRAPHY – THE IDEA AND TECHNIQUE OF EXAMINATION:

Positron emission tomography (PET) is an imaging method from nuclear medicine, using radioactive substances that emit positrons (anti-electrons). During the examination, a small amount of a radiotracer, which is a radionuclide-labeled molecule with a short half-life that is rapidly metabolized in human tissues, is introduced into the human body.

Radiotracers are obtained by the fusion of 2 components: a ligand, which is a chemical that can accumulate within a given tissue or organ (eg, glucose, methionine, or choline), and a radionuclide (eg, 18F or 11C) that emits radiation at a detectable level. The ligand, which acts as a carrier for the radionuclide, is chosen based on the function of the given tissue. Radiotracers are also known as positron markers. Because of their short half-lives, radiotracers used in PET undergo rapid radioactive decay after reaching the target tissue or organ [33].

PET is based on annihilation, which refers to a reaction between a particle and its anti-particle, in the case of PET, positron, and electron, respectively. A positron is a positively charged (+1) particle with a charge and mass equal to that of a negatively charged electron (−1). During annihilation, the masses of both particles are converted into energy emitted as gamma quanta (photons). During PET, positrons emitted by a radiotracer undergo an annihilation reaction, colliding with electrons present within the tissue, which results in the formation of a pair of 511 keV gamma photons moving in opposite directions (at a 180° angle). The photons are detected using a concentric detector ring suitable for registering both coincidently released energy quanta and accurately locating their sources. The number of detector rings around the patient’s body can vary from 6 to 32. Each ring contains 4000–25 000 crystals. Images from the detectors are transmitted to a computer, where they are converted into 2D and 3D images. Therefore, PET is referred to as scanning. The radiation emitted during the examination is registered by a reader, which can estimate the metabolic activity of cells within the human body based on the rate of radioactive decay [33,34]. Because the examination involves radionuclides with short half-lives, most of the radiation is generated during the examination, thereby limiting the risk of radiation tissue injury. The sources of positrons are radioactive substances, such as carbon-11 (11C), gallium-68 (68Ga), fluorine-18 (18F), nitrogen-13 (13N), oxygen-15 (15O), and technetium-99 (99Tc). The ultrashort or short half-life of radiotracer isotopes is a strength of the examination, as it guarantees patient safety, but also a weakness, as the cyclotron producing the isotopes must be located in close proximity to a tomography examination room to facilitate fast delivery of the isotope when its activity is high. The half-life of the most commonly used isotope, 18F, is 110 min, whereas the half-life of 11C is 20 min [35]. Currently, the most popular radiotracer, most commonly used during PET, is fluorodeoxyglucose (FDG), which is a glucose analog labeled with fluorine-18 isotope, 18F. Thus, PET allows visualization of the distribution and metabolism of FDG as a marker of glucose metabolism in the human body. FDG absorbed by cells is transformed into 18F-deoxyglucose-6-phosphate in a process catalyzed by hexokinase. This newly formed compound does not undergo further metabolic processes and it cannot be released from the cells, a phenomenon referred to as a metabolic trap. Thus, 18F-deoxyglucose-6-phosphate accumulates extensively in all the glucose-metabolizing organs. Glucose is the primary source of energy in the human body. In addition to the heart and brain, other areas of extensive glucose metabolism in the human body include inflammatory foci and malignant lesions. It is estimated that up to 90% of early-stage neoplasms can be detected using PET [36]. In addition to glucose metabolism, PET radiotracers can be used to evaluate other metabolic pathways that are disrupted in cancer cells. Considering the heterogeneity of human tissues in terms of their metabolic profiles, the evaluation of neoplastic lesions during PET can involve many radiotracers other than FDG. The most commonly used FDG alternatives include 68Ga-PSMA (prostate-specific membrane antigen), 11C-methionine, choline, and acetate labeled with 11C and 18F [34].

POSITRON EMISSION TOMOGRAPHY – TECHNOLOGY WITH COMPUTED TOMOGRAPHY:

A recent breakthrough has been the introduction of hybrid total-body PET scanners combining PET technology with multi-slice spiral computed tomography or magnetic resonance imaging (PET/CT and PET/MRI systems, respectively). These hybrid scanners can visualize the anatomy of examined tissues and provide functional data on the site of PET radiotracer accumulation. The first PET/CT scanner was developed in 1998 and has become widely available in healthcare units since 2001.

Polyethylene terephthalate/CT is performed as follows: Approximately 1 hour before the examination, a radiotracer, for example, ¹⁸F-FDG, is administered to the patient intravenously. During the first stage of PET, a short (9-second) whole-body scan, the so-called topogram, is obtained to verify correct patient placement. Computed tomography (CT) is performed to obtain a radiographic scan of the body, constituting a reference map for PET images. Finally, PET/CT begins, during which the detectors circle the long axis of the patient and proceed alongside the body. Scanning of the human body lasts approximately 20 minutes and consists of 6 or 7 stages of 3 minutes each. The number of stages depends on a patient body height. After completion of scanning, data from the detectors are transmitted to a computer and analyzed.

First, a qualitative analysis is conducted to verify whether a lesion visible on PET/CT accumulated the radiotracer. Quantitative analysis is then performed; digital images obtained during the body scan accurately depict the actual concentration of the tracer in a given organ, but not with 100% precision. In clinical practice, the quantitative kinetics of PET radiotracers are determined based on the standardized uptake value (SUV), defined as the ratio of tracer concentration within the tissue of interest (in MBq/ml) to the mean concentration of the tracer across the body (ie, administered activity per body weight [MBq/body weight in kg]) [37]. The SUV is calculated using the following formula:

SUV values greater than 2.5–3 correspond to a malignant character of the lesion, although no cut-off value exists to accurately distinguish neoplasms from normal tissues. Interpretation of results also depends on the SUV, showing the metabolic activity of normal tissues, which varies from organ to organ. The heart and brain have high metabolic rates, whereas metabolism within the lungs and bones is relatively low. Hence, the SUVs of normal tissues differ. For instance, SUV values for FDG vary from 0.7 for normal lung tissue, 1.0 for bone marrow, 0.5 for breasts, 2.5 for the liver, and 5.0 for the brain [37]. During the final interpretation of PET results for a suspected lesion, the background uptake of the tracer (ie, the uptake by the adjacent normal tissue) should also be considered. Background uptake is expressed as the tumor-to-background ratio (TBR), defined as the ratio of the SUVmax of the lesion to the SUVmean of healthy tissues. SUV and TBR values are more critical during the monitoring of treatment outcomes than are isolated parameters.

Posterior uveal melanoma is suitable for PET/CT evaluation, similar to other highly aggressive malignant tumors, and is characterized by a higher glucose consumption and, consequently, a higher 18F-FDG uptake, which is reflected by high SUVmax values [38,39]. SUVmax values >2.5 are generally considered a marker of high metabolic activity of the tumor, but some studies have used 2 cut-off values, 2.5 and/or 4.0.

Compared with the SUV, the metabolic rate of glucose (MRglu, in milliliters per minute per 100 g) is a better measure of glucose consumption by a tumor. MRglu is an outcome of the 18F-FDG clearance rate and blood glucose concentration. Thus, MRglu is used during the qualitative analysis of the PET images, whereas SUV is considered a semiquantitative parameter. SUVs are used more frequently in clinical practice because they are easier to calculate. However, the SUV is only a surrogate for qualitative parameters, as it is an estimate of glucose consumption by a tumor. More accurate results can be obtained through a qualitative analysis based on MRglu, which measures actual glucose consumption. However, given that the calculation of MRglu is more complex and time-consuming than SUVmax and SUVmean, this parameter is not routinely used when interpreting PET/CT results [40].

Role of PET/CT in the Detection of Posterior Uveal Melanoma

RELATIONSHIP BETWEEN PET/CT IMAGING OF UVEAL MELANOMA AND TUMOR SIZE:

The first reports on the application of PET/CT for detecting uveal melanomas date back to the first years when this technique was utilized. Reddy et al [41] examined 50 untreated choroidal melanomas (AJCC stages T1 to T3) and found that 14 of them – 8 (33%) medium (T2) tumors and 6 (75%) large (T3) tumors – had uptake and metabolism of a tracer, 18F-FDG (SUV >2.5). PET/CT did not reveal any small uveal melanomas. The smallest detectable tumor had a base diameter of 3×5.9 mm and a thickness of 2.9 mm. The mean resolution of PET/CT images is 3–4 mm; hence, smaller lesions can easily be overlooked. The smallest uveal melanoma detected by PET/CT in a study conducted by Spraul et al [42] was 3×7 mm, confirming that PET/CT is not superior to routine ophthalmological evaluation for detection of smaller tumors. Finger et al [38] found an increased SUV (>2.5) in 11 of 14 choroidal melanomas (T2 and T3). The higher detection rate of melanomas in the study conducted by these authors was associated with the lack of small tumors in their material. Interestingly, the SUVmax for tumors diagnosed more than 10 years before PET/CT exceeded 4.0. Tumors with the highest SUVmax values were primarily composed of epithelioid cells, contained large-diameter (>150 μm) blood vessels, were more often located anteriorly to the equator of the eye, and had a large base. The maximum SUVmax for the largest tumors was 9. McCannelet et al [43] observed SUV >2.5 solely in larger melanomas with a mean diameter of 14.6 mm and a mean thickness of 7.4 mm. Metabolically inactive melanomas (SUV <2.5) had a mean diameter of 10 mm and a mean thickness of 5.0 mm. Matsuo et al [22] also found a positive correlation between SUVmax and tumor diameter. According to these authors, all PET/CT-detectable tumors were dome-shaped and clearly extruded into the vitreous cavity. Lesions that were not detected on PET/CT were flat, diffuse, and/or formed small mushroom-like protrusions. Singh et al [44] analyzed the relationship between the detectability of uveal melanomas on PET/CT and tumor size. The analysis included tumors with no evidence of distant metastasis. The melanoma detection rate on PET was 60%, and it was correlated with tumor size. Detectable tumors had diameters of at least 13 mm, thickness of 4.5 mm, 565 mm3 volume, and SUVmax values of 3.5–8.6. Their study confirmed that small uveal melanomas are not visible on PET/CT and would remain undetected if this method was used as the only diagnostic option. They postulated that the lack of melanoma detection on PET/CT might be associated not only with the small diameter of the tumor but also with the effect of other factors, such as concomitant inflammation, necrosis or lack thereof, intratumoral hemorrhage, and tumor vascularization on glucose metabolism [40,44,45]. According to Calcagni et al [40], the PET/CT detectability of medium and large melanomas qualified for surgical treatment reached 88.5%. Surgical treatment is dedicated primarily to large lesions, and the results of the study mentioned above confirm that the larger the examined tumor, the more likely it is to be visualized on PET/CT. In line with these findings, Leško et al [46] found hypermetabolism on PET/CT scans only in tumors with volumes greater than 565 mm3.

RELATIONSHIP BETWEEN PET/CT IMAGING OF UVEAL MELANOMA AND HISTOLOGICAL CHARACTERISTICS:

Analysis of the correlation between FDG uptake on PET/CT and the histopathological characteristics of posterior uveal melanoma is possible only in tumors treated by resection or enucleation. A study analyzing glucose consumption, a parameter that is considerably elevated in malignant and aggressive tumors, demonstrated that epithelioid cell melanomas presented with significantly higher MRglu values (mean 27.8 ml/min/100 g) than spindle cell melanomas (mean 13.8 ml/min/100 g). MRglu values for mixed tumors were moderate, with a mean value of 19.73 ml/min/100 g [40]. These findings are in agreement with the observation that epithelioid and mixed-cell melanomas exhibit aggressive phenotypes. Although the same study did not show significant cellular structure-related differences in SUVmax values, more aggressive melanomas (epithelioid and mixed-cell tumors) had higher values of this parameter (4.6 and 4.1, respectively) than spindle cell melanomas (2.9) [40]. In a multivariate analysis analyzing a link between histopathological features corresponding to the aggressive phenotype of uveal melanoma (tumor size, mitotic index, proliferative index, density of microvasculature, epithelioid or mixed cells, Ki67) and qualitative (MRglu) and semiquantitative (SUV) PET/CT indices, the cut-off values that most accurately identified high-risk tumors were SUVmax >4.16 and MRglu >21.96 ml/min/100 g. According to Calcagni et al [40], the application of these 2 parameters for the identification of high-risk choroidal melanomas may result in the development of a reliable diagnostic test with high sensitivity (few false negatives) and high positive predictive value (few false positives). Using the cut-off values presented above, the authors obtained 71% sensitivity and 92% positive predictive value for identifying high-risk melanomas.

RELATIONSHIP BETWEEN PET/CT IMAGING OF UVEAL MELANOMA AND CHROMOSOME 3 LOSS:

Loss of chromosome 3 is a risk factor for uveal melanoma development and it is usually associated with a more aggressive tumor phenotype. McCannel et al [43] found a significant association between monosomy 3 and enhanced glucose metabolism within the tumor, which manifested as higher SUVmax values. The results showed a sensitivity of 54% and a specificity of 100%. Papastefanou et al [27] also showed a positive correlation between the metabolic activity of uveal melanoma, monosomy 3, and other prognostic factors, including larger tumor size and TNM prognosis. Up to 92% of the examined melanomas had an SUVmax of >2.5, and in 67%, the SUVmax exceeded 4.0. However, the material analyzed by these authors did not include small melanomas and large tumors constituted up to 84% of the sample. Up to 94% of melanomas with monosomy 3 had an SUVmax >2.5, and up to 80% had an SUVmax greater than 4.0. The authors of that study did not observe a considerable increase in SUVmax in uveal melanomas with chromosome 8q gain, which is another anomaly considered an unfavorable prognostic factor. However, melanomas with monosomy 3 coexisting with chromosome 8q gain had higher SUVmax values than those with monosomy 3 as an isolated chromosomal abnormality. The same study found no significant association between an increase in SUVmax and the predominance of epithelioid cells in tumor architecture, necrosis, inflammatory infiltration, or hemorrhage. These findings are in contrast to the results published by other authors, who observed higher SUVmax values in epithelioid and mixed-cell tumors [38,47] and in tumors with concomitant necrosis and local inflammation [45,47]. According to Yamada et al [39], increased FDG uptake by uveal melanomas may be associated with a high proliferation index and high viability of tumor cells.

Role of PET/CT Imaging in Distinguishing Between Posterior Uveal Melanomas, Benign Lesions and Other Primary or Metastatic Intra-Ocular Malignancies

CHOROIDAL NEVUS:

Distinguishing between uveal melanoma and choroidal nevus is one of the greatest challenges in differential diagnosis. Choroidal nevus is a benign intra-ocular mass that is typically found in Whites, and according to the Blue Mountains Eye Study [48], it has a prevalence of 7%. A nevus usually presents as a well-demarcated, flat, round, or longitudinal area of increased choroidal pigmentation with a dark color. However, amelanotic nevi have also been reported previously. In 91% of cases, choroidal nevi are located posterior to the equator and have a similar frequency, regardless of the eye quadrant [49]. The mean diameter of the choroidal nevus is estimated to be 5 mm, with a mean thickness of 1–1.5 mm. Frequently, the nevus is surrounded by drusen and an altered retinal pigment epithelium [50].

Each identified choroidal nevus should be regularly monitored for potential malignant transformation. Features that may be associated with an increased risk of transformation to choroidal melanoma include thickness >2.0 mm, diameter >5 mm, accumulation of an orange pigment and/or subretinal fluid, <3.0 mm distance from the fovea or optic disc, ultrasonographic acoustic hollowness, floaters, flashes, and blurred vision [50,51]. The presence of at least 3 of these features is associated with a 50% risk of malignant transformation within the next 5 years [52].

Although observing skin nevi for potential malignant transformation is relatively simple, follow-up of choroidal nevi is complicated. This has stimulated research on new alternative techniques to specialist ophthalmological examinations (FA, IGA, OCT, and OCTA) that could replace the direct histological analysis of the lesion, which is the only accurate diagnostic method currently available. Choroidal nevi typically need to be differentiated from small, thinner melanomas. Mathematical analyses of tumor doubling and the results of The Collaborative Ocular Melanoma Study (COMS) suggest that metastatic spread of choroidal melanoma can occur even at the very early stages when a small tumor is typically described as a “suspicious choroidal nevus” [53,54]. Theoretically, differential diagnosis can be facilitated by analyzing the metabolic activity within the lesion, assuming that malignant tumors are more metabolically active than nevi. However, small choroidal melanomas cannot be detected using PET/CT. Detecting choroidal nevi can also be challenging, given their thinness. The group of patients examined by Spraul et al [42] included 6 with choroidal or ciliary body nevi, and none of these lesions were detected on PET/CT. In view of these findings, the role of PET/CT in distinguishing choroidal melanomas from choroidal nevi remains limited [2].

BENIGN INTRA-OCULAR LESIONS:

Benign intra-ocular lesions such as hemangiomas of the choroid and extrafoveal choroidal neovascularization, which may need to be differentiated from choroidal melanoma, also yield negative results on PET/CT [42]. This suggests that PET/CT is not suitable for evaluating metabolically inactive benign lesions. The lack of metabolic activity on PET/CT could mean that the lesion is not malignant. Unfortunately, under this assumption, small uveal melanomas that are undetectable on PET/CT can be misdiagnosed as benign. However, PET/CT is not sufficiently sensitive to detect such lesions.

OTHER PRIMARY OR METASTATIC INTRA-OCULAR MALIGNANCIES:

PET/CT can be helpful in the differential diagnosis of intra-ocular tumors, given that some malignancies, including retinoblastoma and low-grade mucosa-associated lymphoid tissue lymphoma, have very low FDG uptake [3], which differentiates them from uveal melanomas. However, ophthalmic lymphoma derives from the orbital tissues, eyelids, and cornea instead of being a primary intra-ocular tumor, and retinoblastoma occurs primarily in children; therefore, these tumors are rarely considered as differential diagnoses during the evaluation of suspected uveal melanoma.

As a richly vascularized tissue, the choroid is often a site of metastatic spread for malignancies located in other organs, such as the breast, lung, or prostate. Although both ocular metastases and primary tumors can be detected using FDG-PET/CT, this method is not superior to other methods for differentiating between intra-ocular metastases and primary uveal melanoma. Available evidence suggests that these 2 tumor types do not differ in their metabolic activity and thus cannot be differentiated on PET/CT [4].

Role of PET/CT in Staging of Metastatic Uveal Melanoma

METASTATIC UVEAL MELANOMA:

Neoplasm staging is essential for planning therapeutic strategies. Despite treatment, approximately 50% of patients with uveal melanoma eventually develop distant metastases [6,55]. The route of spread is associated with the eye anatomy. Since the eye does not have lymphatic vessels (present solely in the conjunctiva), uveal melanoma spreads through the blood, and its distant metastases are found primarily in the liver [14]. Other markedly less frequent locations of uveal melanoma metastases include the lungs, skin, abdominal cavity, pelvis, and brain [18]. Treatment options for metastatic uveal melanoma are quite limited. Whenever possible, the primary lesion is surgically removed, which, in some cases, requires enucleation. Surgery is combined with systemic chemotherapy. Other treatment options include chemoembolization or radioembolization and targeted hepatic chemotherapy [14]. Some patients with metastatic uveal melanoma can also qualify for immunotherapy with small-molecule kinase inhibitors and immune modulators [56].

Although evidence of metastatic spread is found in less than 1% of patients at the time of uveal melanoma diagnosis, the risk of metastasis increases substantially with disease duration. PET/CT can also be used to monitor patients with uveal melanoma for potential metastatic spread [46]. According to Freton et al [57], PET/CT detected metastases in 4% of the patients with newly diagnosed melanoma. It must be stressed that even if primary uveal melanoma can be controlled through local treatment, the patient may still die because of subclinical micrometastases present at the time of the initial diagnosis. In line with the UK National Guidelines for managing uveal melanoma, tumor staging is an essential component [58], as accurate determination of the disease stage can optimize further therapeutic approaches.

APPLICATION OF PET/CT TO EVALUATION OF LIVER METASTASES FROM UVEAL MELANOMA:

As the liver is the primary location of distant metastases in uveal melanoma, most published studies have dealt with the application of PET/CT to evaluate this organ. Liver lesions were interpreted as metastases if their SUVmax on PET/CT was higher than that of the adjacent tissues. The SUVmax for normal liver tissues did not exceed 2.5. The metabolic activity of a lesion could also be estimated based on its TBR, with higher values of this parameter corresponding to more metabolically active and, hence, more aggressive masses. Del Carpio et al [59] estimated the mean SUVmax for the liver metastases of uveal melanoma is 8.5 (range, 2.6–42.2). Lesions with low SUVmax values, similar to normal liver uptake, and TBR <1.1 constituted 19.6% of all identified metastatic tumors. The largest of the metastases with the lowest metastatic activity (TBR <1.1) had a mean diameter of 10.5 mm (range, 8–50 mm), whereas the mean diameter of metabolically active (TBR >1.1) metastatic lesions was 24 mm (range, 10–178). All tumors with diameters >50 mm were metabolically active, whereas all metastases with diameters <10 mm were iso-metabolic [59].

Although PET/CT is not routinely used to detect uveal melanoma metastases to the liver, it has become an important examination in patients with normal CT results and normal liver enzyme levels [60]. Several studies have demonstrated that liver function tests have poor sensitivity for detecting liver metastases [61]. In turn, the results of the COMS group study suggest that liver function tests provide high specificity and predictive values, but low sensitivity in identifying choroidal melanoma metastases to the liver [54]. PET/CT also plays an essential role in the detection of metastatic uveal melanoma in patients with contraindications to MRI. Some evidence suggests that PET/CT can detect liver metastases that remain invisible on abdominal ultrasound [56,57]. Nevertheless, ultrasonography is an excellent tool for detecting benign liver lesions, which, unlike malignant masses, have no metabolic activity and cannot be visualized using PET/CT. Cohen et al [56] examined 108 uveal melanomas for potential liver spread by using PET/CT and ultrasonography. Although malignant lesions could be visualized effectively on PET/CT, benign liver lesions in 18% of patients could be detected solely on ultrasound, with PET/CT yielding a negative result. These findings confirm that PET/CT and ultrasound are complementary methods, and should be used together in the staging of uveal melanoma.

In some studies, PET/CT yielded 100% sensitivity and 100% specificity in detecting uveal melanoma metastases to the liver [61,62]. However, according to Strobel et al [63], the sensitivity of PET/CT in detecting liver metastases was only 59%. Kurle et al [61] demonstrated that the sensitivity of liver function tests for detecting liver metastases was only 12.5%, which is much lower than the 100% sensitivity of PET/CT. Francken et al [64] highlighted the role of FDG-PET/CT in the detection of isolated and potentially resectable liver metastases, as this method provided 100% sensitivity and 67% specificity, along with positive and negative predictive values of 88% and 100%, respectively. According to Servious et al [65], the sensitivities of MRI and PET/CT for staging liver metastases were 67% and 41%, respectively, with positive predictive values of 95% and 100%, respectively.

APPLICATION OF PET/CT TO EVALUATION OF OTHER METASTASES FROM UVEAL MELANOMA:

Finger et al [66] confirmed that PET/CT is useful in detecting metastases not only to the liver, but also to the bones and lymphatic nodes. Additionally, PET/CT detected benign lesions in the bones, lungs, lymph nodes, large intestine, and rectum in 13.4% of patients with choroidal melanoma. However, the study group also included 5.7% of patients in whom the presence of suspected malignant lesions on PET/CT scans was not confirmed on histopathological examination [66]. In a study by Kurli et al [61], concomitant benign lesions in the bones and lymph nodes (inflammation and old fractures) were detected by PET/CT in 15% of patients. These lesions could be distinguished from choroidal melanoma metastases on PET/CT scans by comparing glucose uptake by the mass and adjacent normal tissues and quantitatively comparing the SUVmax values.

One advantage of PET/CT as a screening tool for the detection of distant metastases of uveal melanoma is that it can visualize malignant lesions in multiple organs during a single whole-body scan [57]. This eliminates the need to obtain separate scans for individual body parts. Although the liver is the primary location of uveal melanoma metastases, in 20–30% of patients, concomitant metastases are also be present in other organs. Interestingly, 8.4% of patients with choroidal melanoma present with concomitant primary tumors in other locations [57]. Cohen et al [56] confirmed the role of whole-body PET/CT in detecting such concomitant malignancies in patients with uveal melanoma. In a study conducted by these authors, concomitant malignancies in the lungs, colon, breasts, adrenal glands, and thyroid were found in 9.3% of patients with choroidal melanoma. The early detection of other asymptomatic primary malignancies has a fundamental effect on patient survival. The risk of developing concomitant primary malignancies in cancer patients is twice as high as that in sex- and age-matched cancer-free controls [67]. The concomitant occurrence of some malignancies is genetically determined. For example, in patients with hereditary BAP1 cancer predisposition syndrome, uveal melanoma can co-occur with other neoplasms [68]. Thus, early detection of choroidal melanoma should not exempt physicians from evaluating patients for concomitant primary malignancies in other locations [57]. Whole-body PET/CT scans can be particularly useful in such cases.

Five-year overall survival rates of uveal melanoma vary between 50% and 65% [69]. Glucose metabolism in liver metastases, expressed as SUVmax and TBR, is considered an independent prognostic factor for survival. In patients with SUVmax ≥8.5, mean overall survival was estimated at 9.4 months, as compared with up to 38.4 months in those with SUVmax <8.5 [59]. Similarly, the mean overall survival in patients with liver metastases and TBR <1.86 was 38.5 months versus merely 13 months in those with TBR ≥1.86.

Strobel et al [63] compared PET/CT results in patients with liver metastases of choroidal and cutaneous melanomas. The mean SUVmax for uveal melanoma liver metastases was significantly lower than that for cutaneous melanoma metastases to the liver (3.5 vs 6.6). However, histological analysis revealed no differences in glucose transporter-1 (GLUT-1) expression between choroidal and cutaneous melanoma metastases. The authors of the study did not speculate on the mechanism by which FDG uptake in uveal melanoma metastases was significantly lower than that in cutaneous melanoma metastases. Interestingly, in the same study, PET/CT detected extra-hepatic metastases in 36% of patients with uveal melanoma and in 86% of patients with cutaneous melanoma [63].

Monitoring of Treatment Outcomes in Posterior Uveal Melanoma with PET/CT

METHOD OF UVEAL MELANOMA TREATMENT:

The choice of treatment method for uveal melanoma depends on the tumor size, location within the eye, especially in relation to the macula and optic disc, concomitant complications, visual acuity in the affected eye, and the condition of the fellow eye. Additionally, patient age and general condition are important factors that should be considered. Currently available treatment options for uveal melanoma include radiotherapy (brachytherapy and irradiation with charged elementary particles), transpupillary thermotherapy, tumor resection (endo-resection or exo-resection with trans-scleral excision of the tumor), enucleation, and orbital exenteration [70].

Brachytherapy is the most commonly used treatment for this condition. In patients treated with this method, a radioactive plaque containing radionuclides is sewn onto the surface of the sclera based on the tumor, leaving a 2-mm margin of normal tissue. Plaques often contain iodine-125 (I-125) or ruthenium-106 (Ru-106) isotopes. Ruthenium undergoes beta-radioactive decay and is used to treat tumors with a thickness of no greater than 5–6 mm. Iodine emits gamma radiation and has applications in the treatment of thick tumors. The effect of brachytherapy is associated with the destruction of DNA in cancer cells, which eventually leads to the annihilation of cancer cells, with concomitant fibrosis and obliteration of intrinsic tumor vessels [71].

Another currently available treatment option is the irradiation of melanoma with a significant amount of energy emitted by charged elementary particles, protons, or helium ions. Tumor cells can also be destroyed by a diode laser with an 810-nm wavelength during a procedure referred to as transpupillary thermotherapy. The laser beam heats the tumor tissues to 45–60°C, which leads to direct injury of the cells and blood vessels and disruption of cellular enzymatic systems. Because transpupillary thermotherapy causes necrosis of tissues 3–4 mm in depth in the tumor, this method is used to treat thinner melanomas. In thicker tumors, transpupillary thermotherapy produces optimal effects when combined with brachytherapy in the so-called “sandwich method”. Photodynamic therapy, with indocyanine green as a photosensitizing agent can be used to treat smaller melanomas or as an adjuvant treatment after brachytherapy [70].

Tumors that are not eligible for conservative management are qualified for surgical treatment using local resection. Surgical treatment of advanced uveal melanomas may require enucleation, and orbital exenteration is the only option for cases with extra-orbital spread.

Monitoring therapeutic outcomes and tumor regression play key roles in assessing the effectiveness of treatment. Radiotherapy, the most common treatment modality, results in tumor shrinkage, scarification, microvasculature reduction, and retinochoroidal atrophy around the tumor base.

MONITORING OF CHEMOTHERAPY IN UVEAL MELANOMA WITH PET/CT:

According to Orcurto et al [72], the ratio of lesion SUVmax to liver SUVmax decreased considerably during chemotherapy for metastatic uveal melanomas with a stable diameter of metastatic foci on MRI, but the growing size of the metastatic foci on MRI was associated with an increase in this ratio. These findings suggest that 18F-FDG-PET/CT may be helpful for early monitoring of treatment responses. Response to the treatment is assessed with the Response Evaluation Criteria in Solid Tumor (RECIST) criteria [73]. In line with these criteria, therapeutic response is defined as the complete disappearance of metastases or at least a 30% decrease in the sum of the longest diameters of metastatic lesions. Patients who developed new metastases and those in whom the sum of the longest lesion diameters decreased by less than 30% were considered non-responders. On PET/CT scans, a lack of response to therapy is defined as an increase in FDG uptake in pre-existing metastases or new metastases, or a lack of a significant decrease in FDG uptake in pre-existing lesions. Response to chemotherapy is defined as a decrease in the size and FDG uptake of the lesions by more than 30% or complete disappearance of pathological FDG uptake [74]. The correct identification of responders and non-responders might contribute to optimizing the therapeutic approach, primarily regarding adverse effects and treatment costs. According to Strobel et al [74], a clear relationship exists between overall survival (OS) and progression-free survival (PFS) in patients with stage IV uveal melanoma and PET/CT evidence of treatment response or lack thereof. The OS of PET/CT responders to chemotherapy was markedly longer, with the proportion of those surviving for 1 year being 80% versus 40% in PET/CT non-responders. The mean OS in PET/CT responders was 18 months (range, 14–28) versus 11 months in PET/CT non-responders, whereas the mean PFS was 9 and 3 months, respectively. Importantly, neither OS nor PFS correlated with the levels of the S-100B marker, which suggests that PET/CT is superior as a measure of response to chemotherapy. Additionally, serum S-100B yielded one-third of the false negatives as a marker of the overall treatment response in metastatic uveal melanoma [75]. The analysis of mean pre- and post-chemotherapy SUVmax values in patients with stage IV uveal melanoma decreased from 6.8 at baseline to 5.5 after 3 treatment cycles [74].

MONITORING OF BRACHYTHERAPY IN UVEAL MELANOMA WITH PET/CT:

PET/CT has also been shown to play a role in monitoring of brachytherapy outcomes. Eyes with uveal melanoma were demonstrated to have a significantly thicker choroid than non-affected eyes, with mean values of 293 μm and 242 μm, respectively [76]. Increasing choroid thickness in melanoma eyes is associated with increased perfusion, providing blood for highly metabolically active tumor cells undergoing tumorigenesis. Another explanation for the higher choroidal thickness in eyes with melanoma might be the overexpression of vascular endothelial growth factor and the resultant increase in choroid permeability, leading to fluid accumulation in this tissue. Among eyes with melanoma, those with metabolically active tumors were shown to have a substantially thicker choroid than those with metabolically inactive lesions (mean 348 μm and 280 μm, respectively). PET/CT performed 6 months after Ru-106 brachytherapy with adjuvant transpupillary thermotherapy demonstrated a statistically significant decrease in choroid thickness, by 40 μm on average [76]. The post-treatment decrease in choroidal thickness results from reduced choroidal perfusion associated with post-irradiation changes in the tumor vasculature, such as degeneration, hyalinization, necrosis, and occlusion of the vessels [77]. The decrease in choroidal perfusion and thickness is also associated with the lower metabolic activity of the tumor.

Lee et al [78] analyzed the use of PET/CT to evaluate uveal melanoma regression after Ru-106 brachytherapy combined with thermotherapy. Before treatment, the mean thicknesses of the metabolically active and inactive tumors were 8.8 mm and 5.0 mm, respectively. The mean tumor thickness at 3, 6, and 12 months after treatment was lower than that at baseline, with reductions of 88%, 78%, and 64%, respectively, for metabolically active melanomas, and by 95%, 89%, and 81%, respectively, for metabolically inactive lesions. The degree of tumor regression was substantially higher in metabolically active melanomas than in metabolically inactive melanomas. This suggests that metabolically active uveal melanomas are more prone to PET/CT-detectable regression in response to combined brachytherapy and thermotherapy. Thus, this imaging method appears to have value as a prognostic factor for treatment response. Finger and Chin [79] analyzed the time required to reduce the metabolic activity of choroidal melanoma to an SUV of 0 after brachytherapy. In the study group with a mean pretreatment SUVmax of 3.7, the mean time to tumor regression or lack of glucose uptake (SUV=0) was estimated to be 8 months (range, 6–18 months).

Limitations of PET/CT Imaging

The high false-negative rate, corresponding to the failure to detect uveal melanoma in patients presenting with malignancy, is a serious limitation of PET/CT. False-negative results are obtained particularly often in the case of small uveal melanomas, which are, among others, associated with the low resolution of the method (3–4 mm). When examining the liver for potential metastases, one should remember that FDG uptake in that organ also reflects the physiological glucose metabolic flux in the normal liver. A normal mottled background can reduce the sensitivity of PET/CT in detecting liver pathologies [56]. Thus, even larger liver lesions may remain undetected if the physiological processes in that organ are not considered when interpreting the PET/CT results.

Another weakness of PET/CT in the detection of uveal melanoma metastases is the inability to visualize small or very small liver lesions, and all metastases of uveal melanoma to the abdominal and thoracic cavity, regardless of their diameter, can be detected by MRI [60,66,72]. The mean diameter of liver lesions detectable on MRI but invisible on PET/CT is 0.6 cm, whereas the mean diameter of lesions detected with both methods is 2.1 cm [72]. PET/CT can detect 79% of liver metastases with diameters equal to or greater than 1.2 cm. Metastases with diameters below 1.2 cm have significantly lower SUVmax values than those with diameters ≥1.2 cm. Although a positive correlation exists between the size of the liver lesion on MRI and SUVmax, no association was found between the SUV and the number of metastases [72]. Metabolic activity was particularly high in metastatic tumors that developed shortly after the diagnosis of uveal melanoma [57].

Whether irradiation during whole-body PET/CT poses a risk to patients remains debatable. The effective radiation dose of a single 18F-FDG-PET/CT scan is 18–25 mSv, but it can increase to 30 mSv in the case of multiphasic abdominal and pelvic scans [80]. Such a dose is associated with a lifetime cancer risk of 0.6% [80]. Thus, one should consider whether the benefits associated with detecting metastases or concomitant asymptomatic primary malignancies in other organs and the early implementation of appropriate therapy outweigh the risk of increased radiation exposure.

Future Directions

The future of PET/CT in ophthalmology is an interesting topic. PET/CT is inaccurate in detecting small uveal melanomas, small metastatic lesions, and tumors with low metabolic activity, yielding a large proportion of false-negative results. However, PET/CT may have applications in ophthalmological oncology if new and more specific radiotracers are introduced. Such new radiotracers targeting melanin, including radiolabeled antibodies, benzamide, and benzamide analogs, are the subject of ongoing research [2]. Although these studies dealt with cutaneous melanoma, the new radiotracers might also find applications in evaluating uveal melanoma. One new melanin-targeted radiotracer that has already been introduced to clinical practice is 5-bromo-N-(2-[diethylamino]ethyl) picolinamide labeled with 18F (18F-5-FPN) [81]. Studies in mice have demonstrated that this radiotracer has a high potential for effectively detecting primary tumors and metastatic lesions with diameters of 1–2 mm. One potential drawback of 18F-5-FPN is that it is metabolized in the liver, which may limit its application in the detection of metastases to this organ. The TBR value on 18F-5-FPN PET was significantly higher than that on 18F-FDG-PET [2]. A modified 18F-5-FPN, in the form of 18F-PEG3-FPN, had lower liver uptake but an equally high affinity to melanin. Clinical trials of this radiotracer in patients with suspected or confirmed uveal melanoma are planned at the end of 2023 after recruiting the study group [2]. Another substance being assessed in ongoing research is 18F-6-fluoro-N-[2-(diethylamino)ethyl] pyridine-3-carboxamide (18F-MEL050) [82]. This radiotracer was shown to distinguish between melanoma and normal adjacent tissues effectively and has a higher TBR than 18F-FDG.

Another future direction in ophthalmological oncology might be PET/MRI, an examination that, to the best of our knowledge, has not yet been used in ophthalmology. High-contrast imaging of soft tissues, availability of various imaging sequences, simultaneous image acquisition, and low radiation doses might constitute arguments for applying PET/MRI to evaluate ocular tumors.

As technology continues to evolve, the role of PET/CT in the comprehensive management of eye cancer is likely to expand, offering enhanced insights into tumor characteristics and treatment outcomes. Regarding the impact of technological developments in the PET/CT field, the application of long-axial field-of-view PET/CT scanners in melanoma patients should be considered in the future [83].

Conclusions

Although PET/CT has limited diagnostic value in the detection of posterior uveal melanomas, it is useful for detecting medium and large, dome-shaped, or mushroom-shaped melanomas. PET/CT poorly visualizes flat or diffuse lesions. PET/CT is a useful imaging technique for staging patients with choroidal melanomas.

The metabolic activity of uveal melanoma on PET/CT, expressed as the maximum standardized uptake value (SUVmax), correlates strongly with the tumor size. SUVmax may be considered an independent prognostic factor. High metabolic activity of the tumor or metastasis on PET/CT is associated with a poor prognosis.

PET/CT is a useful surveillance tool for the simultaneous detection of distant metastases of uveal melanoma in multiple organs. In patients with posterior uveal melanoma, PET/CT can play a role in monitoring treatment responses and determining whether systemic chemotherapy or local brachytherapy should be combined with thermotherapy.