Purpose

Historically, enucleation has been the standard treatment for primary uveal melanoma. The collaborative ocular melanoma study (COMS) demonstrated equivalent overall survival following episcleral brachytherapy compared with enucleation, while allowing globe preservation and, in selected cases, retention of useful vision [1-3]. Subsequent COMS confirmed these results with long-term follow-up [4].

Currently, interventional radiotherapy (IRT; brachytherapy) is considered the treatment of choice for small- and medium-sized uveal melanomas, achieving 5-year local control rates of approximately 85-95% with acceptable toxicity [2, 5-7]. A recent pooled analysis by Buonanno et al. confirmed high tumor control rates across different isotopes and institutions [8]. Among the isotopes used worldwide, ruthenium-106 (¹⁰⁶Ru) remains the most frequently applied in Europe [7, 9, 10].

Despite the favorable outcomes, local recurrence after plaque brachytherapy still remains a clinically significant problem, reported in 2-21% of cases depending on tumor characteristics, isotope selection, plaque positioning, and duration of follow-up [5, 9, 11, 12]. Management of recurrent disease is challenging, as treatment options must ba- lance oncologic control against cumulative ocular toxicity.

Several globe-sparing salvage strategies have been described. Transpupillary thermotherapy (TTT) is commonly used for small marginal recurrences and may achieve local tumor regression in selected cases; however, its effectiveness is limited in thicker or amelanotic tumors [13, 14]. For this reason, TTT is frequently applied as an initial, minimally invasive salvage attempt, while patients with persistent or progressive disease are subsequently considered for more definitive treatments, including re-irradiation or enucleation. In our institutional practice, TTT is routinely applied as a first-line salvage therapy, and patients with persistent or progressive tumor growth are subsequently offered fractionated stereotactic radiotherapy (fSRT) as a non-invasive, organ-preserving alternative to enucleation.

Fractionated stereotactic radiotherapy enables highly conformal dose delivery and precise sparing of critical ocular structures, and has been used both as primary treatment and salvage therapy for uveal melanoma. Nevertheless, re-irradiation raises concerns regarding cumulative toxicity, including radiation retinopathy, optic neuropathy, cataract, vitreous hemorrhage, secondary glaucoma, scleral necrosis, and the potential need for secondary enucleation [7, 9, 15-19].

Given the limited data on fSRT as a salvage option following plaque brachytherapy and unsuccessful TTT, we conducted a retrospective case series to describe the feasibility, technical aspects, and short-to-mid-term clinical outcomes of fSRT in patients with locally recurrent uveal melanoma.

Material and methods

A retrospective chart review was performed on patients treated with fSRT for locally recurrent uveal melanoma after primary ¹⁰⁶Ru plaque brachytherapy. Initial plaque implantations were performed at a tertiary ophthalmology center, in collaboration with a radiation oncology department.

Local recurrence was defined as documented tumor progression after plaque brachytherapy, characterized by one or more of the following: increase in tumor thickness by more than 20% on ultrasound, lateral growth beyond the original plaque margins, or new tumor activity on co-lor fundus photography or fundus autofluorescence.

All recurrences were confirmed using a combination of color fundus photography, fundus autofluorescence, and ultrasonography.

In accordance with our institutional practice, all patients initially underwent transpupillary thermotherapy as the first-line salvage therapy. TTT was applied as an initial salvage attempt, whereas fSRT was offered only in cases of insufficient response or further progression, and therefore served as the definitive treatment for persistent disease. Due to this sequential approach, oncologic outcome reflected combined salvage management, with fSRT serving as the definitive treatment for persistent disease.

Fractionated stereotactic radiotherapy was delivered using the CyberKnife^® system (Accuray Inc., Sunnyvale, CA, USA), a robotic linear accelerator allowing non-isocentric, non-coplanar beam delivery. Eye immobilization was achieved using a vision-fixation device secured to the treatment couch.

Treatment planning was based on non-contrast computed tomography with thermoplastic mask immobilization. Since 2020, contrast-enhanced magnetic resonance imaging has been additionally acquired to improve target delineation. Gross tumor volume (GTV) encompassed the visible recurrent lesion, and planning target volume margins were minimized to account for setup uncertainty and eye motion.

Dose and fractionation schedules varied over the study period, and were individualized based on tumor size, location relative to critical ocular structures, and prior radiation exposure. Each fraction was prescribed to a 80% isodose line. The detailed fractionation schemes are summarized in Table 1.

Follow-up examinations were scheduled every 3-6 months, and included assessment of best-corrected visual acuity, intra-ocular pressure, slit-lamp and fundoscopic examination, and ultrasonography. Additionally, liver ultrasound was performed every 6 months as part of routine metastatic surveillance.

Local tumor control was defined as the absence of further progressive tumor growth requiring additional oncologic intervention, and globe preservation was described as the avoidance of enucleation during follow-up. Treatment-related toxicities were assessed clinically, but formal grading using standardized toxicity scales was not performed.

Results

Nine patients (4 female, 5 male) with a mean age of 63 years (range, 42-84 years) were included in the study. All patients had previously undergone ¹⁰⁶Ru plaque brachytherapy between 2013 and 2022, with a mean prescribed apex dose of 94 Gy. All recurrent tumors were located in the posterior segment and/or close to the optic disc. The median interval between primary brachytherapy and diagnosis of recurrence was 15.1 months (range, 5.2-40.3 months).

All patients received TTT as the initial salvage therapy before fSRT. The median interval between initial TTT and fSRT was 3.2 months (range, 1.5-6.0 months), and variations in tumor thickness and location likely influenced individual responses to salvage therapy. The fractionated stereotactic radiotherapy regimens and target volumes are summarized in Table 1. The median planning target volume was 1.3 cm³ (range, 0.9-4.4 cm³).

The median follow-up after fSRT was 45.2 months (range, 13.9-121.4 months). Initial post-treatment tumor shrinkage was observed in all cases. During follow-up, globe preservation was achieved in seven eyes (78%), while two eyes (22%) required secondary enucleation at a mean interval of 41.5 months after fSRT (range, 19-64 months). One enucleation was performed due to tumor extension involving circumferentially the ciliary body, and the other due to suspected tumor re-growth and bleeding during long-term follow-up. Histopathological examination in both cases revealed partially necrotic melanoma with positive BAP1 staining.

The most frequently observed treatment-related complication was cataract formation, occurring in five patients (55.6%), and no cases of secondary glaucoma were recorded. Other late radiation-related complications were not systematically assessed, representing a limitation of this study. Final visual acuity ranged from severe visual impairment to near or no light perception (logMAR, 1.3-3.0). Baseline visual acuity prior to fSRT varied widely, and was influenced by prior treatments.

At the time of last follow-up, no distant metastasis were detected on routine imaging, although metastatic surveillance relied primarily on liver ultrasound, which may underestimate the true incidence of distant disease.

Discussion

Local recurrence of uveal melanoma after plaque brachytherapy presents a significant therapeutic challenge. While enucleation remains a definitive option, globe-sparing salvage strategies are increasingly sought to preserve ocular integrity and patient quality of life. Repeat plaque brachytherapy has been shown to be feasible in selected patients but associated with increased cumulative radiation exposure and risk of late toxicity [20, 21].

Other re-irradiation approaches, such as proton beam therapy and stereotactic radiosurgery, have demonstrated encouraging tumor control rates in small series, although with substantial risks of radiation-induced ocular complications [22, 23]. These experiences underscore the need for carefully selected and technically precise salvage treatments.

The present retrospective case series describes the use of fSRT as a definitive salvage modality in patients with locally recurrent uveal melanoma after failed TTT and prior ¹⁰⁶Ru brachytherapy. Given the small cohort size and heterogeneous dose-fractionation schedules, our findings should be interpreted as descriptive and hypothesis-generating rather than confirmatory.

Importantly, all patients in this series underwent TTT before fSRT, reflecting real-world clinical practice. The oncologic outcomes cannot be attributed solely to fSRT; it served as the final, non-invasive globe-sparing option in eyes with persistent or progressive disease following initial salvage therapy.

Globe preservation was achieved in the majority of patients, although two eyes ultimately required enucleation during long-term follow-up. These cases highlight both the potential and limitations of re-irradiation in advanced or biologically aggressive tumors. The observed toxicity profile was dominated by cataract formation; however, the lack of systematic toxicity grading represents an important limitation of this study.

Conclusions

Despite the constraints, our experience suggests that fSRT is a feasible globe-sparing salvage option for selected patients with locally recurrent uveal melanoma after plaque brachytherapy and unsuccessful TTT. Further studies with standardized endpoints, detailed dosimetric reporting, and longer follow-up, are required to better define its role and optimize treatment protocols.