LAUSR

Treatments for iridociliary melanomas include local resection, plaque radiotherapy, proton beam irradiation, or enucleation. For local tumor resection, iridocyclectomy, choridectomy, partial lamellar sclerouvectomy, and block excision with tectonic corneoscleral grafting have been reported. Because 90% of uveal melanomas invade the sclera histopathologically, we prefer Naumann’s radical removal of the tumor along with the adjacent full-thickness cornea and sclera in 1 block. Recently, chromosome 3 analysis or gene expression profiling have been used widely for stratification of the risk of metastasis and tumor-related death, with limited clinical impact up until now because no effective therapy exists against advanced metastatic uveal melanomas. This may change in the near future, with the advent of clinical trials testing the promising concept of tumor vaccination therapy in uveal melanoma. This method relies on the ex vivo sensitization of the patient’s own dendritic cells against tumor antigens of the patient’s uveal melanoma with subsequent reinjection of stimulated immune cells into the patient with the aim of initiating a vigorous immune response against melanoma cells within and outside the eye. A phase III trial is currently being carried out. Because this is an adjuvant therapy trial, only monosomy-3 patients likely developing metastatic disease are enrolled. In this context, our purpose was to analyze whether block excision of iridociliary tumors, while attempting to remove the tumor completely and retain good vision, allows for molecular tumor profiling and tissue sampling for immune-based vaccination therapy. Six consecutive patients with iridociliary tumors involving not more than 4 clock-hours (Table S1; available at www.aaojournal.org) underwent block excision of the tumor together with the adjacent cornea and sclera in full-thickness and subsequent tectonic corneoscleral grafting (Fig 1). Block excision specimens were bisected and one-half was formalin fixed and paraffin embedded for histopathologic and molecular-pathologic evaluations, including targeted next-generation sequencing for the detection of hotspot mutations in 17 melanoma-related genes (Table S2; available at www.aaojournal.org), fluorescence in situ hybridization for monosomy-3 detection, and sanger sequencing for EIF1AX and SF3B1. The second half of the tumor was used to prepare autologous tumor RNA for transfection of the patient’s dendritic cells produced by leukapheresis. Melanoma patients displaying monosomy-3 were invited to participate in the phase III adjuvant vaccination trial (NCT01983748), randomized in a control arm receiving standard care (observation only) as well as in the trial arm receiving eight intravenous vaccinations with 20 10 tumor RNA-loaded autologous dendritic cells at specific intervals over 2 years. All patients underwent standardized ophthalmic and systemic check-ups within 28 11 months of follow-up (range, 9e42). The surgical procedure was performed successfully in all cases. In 1 patient, a second resection was required after 2 weeks, owing to histopathologic tumor involvement of the posterior section margin; the second resection revealed tumor-free margins. In all cases, the eye could be conserved with maintained vision (median 20/45 preoperatively vs 20/50 at the end of follow-up) and without local recurrence. Perioperative complications including vitreous hemorrhage in 2 patients as well as corneal endothelial decompensation, steroid-induced intraocular pressure elevation and cystoid macular edema in 1 patient resolved within 6 weeks. Irreversible complications, such as proliferative vitreoretinopathy, atrophy of the globe, or epithelial ingrowth were not observed. One patient developed epiretinal membrane 6 months postoperatively, but refused membrane peeling despite slight metamorphopsia. Histopathologic examination revealed melanocytoma in 2 patients and melanoma in four patients. Two of the iridociliary melanomas showed invasion of the inner scleral layers. Next-generation sequencing was evaluable in 5 patients, and impossible in 1 case owing to fixation artifacts and low DNA concentrations. Two patients showed wild-type sequences for the gene regions analyzed, whereas 3 patients revealed a point mutation in codon 209 of the GNA11 gene (p.Q209L and p.Q209Y, respectively). In 2 of these patients, an additional mutation in codon 148 of the CDKN2A gene was detected. Fluorescence in situ hybridization was feasible in only 3 patients; the nonevaluable cases showed a strong background signal most likely owing to high tumor pigmentation. All 3 patients revealed a deletion of chromosome 3p; the average number of tumor cells containing 1 counting orange centromere 3 signals was 85%, 68%, and 41%, respectively. Sanger sequencing for EIF1AX and SF3B1 was feasible in patients 1 and 5, revealing wild-type sequences for the mutational hotspots of both genes. For patients 4 and 6, we saw wild-type sequences for EIF1AX. Sequencing of SF3B1 failed in these patients and in patients 2 and 3. All failures were mainly owing to high fragmentation of DNA from the available formalin-fixed tissue. One patient consented to participate in the vaccination trial, and here the excised melanoma tissue was sufficient to prepare the tumor vaccine. However, the patient could not finally be enrolled owing to logistical reasons. None of the 6 patients showed evidence of lymphatic spread, distant metastasis, or tumor-related death within follow-up. This study reveals 3 important findings. First, the surgical technique of block excision with tectonic corneoscleral grafting allows for complete resection of iridociliary tumors involving not more than 4 clock-hours, while maintaining good vision. In contrast with other resection strategies, such as iridocyclectomy or partial lamellar sclerouvectomy, our block excision technique decreases the risk of incomplete tumor resection within the scleral flap and subsequent intrascleral recurrence; histopathologic studies have shown that 90% of uveal melanomas invade the sclera, and 45% reveal invasion of >30% of scleral thickness. Second, our procedure provides tissue not only for histopathologic diagnosis (which is essential, even in clinically challenging cases), but also for molecular profiling, such as next-generation sequencing, fluorescence in