Misdiagnosis of uveal melanoma and other intraocular malignancies in enucleated eyes: a 28-year national review

Abstract Background Intraocular tumors are often diagnosed clinically rather than by biopsy, so a benign lesion can be mistaken for malignancy and lead to enucleation. We determined how often uveal melanoma (UM) and other intraocular tumors are clinically misdiagnosed among enucleated eyes at the Swedish national ophthalmic pathology laboratory over 28 years, and the lesions responsible. Methods In this retrospective, registry-based case series, all enucleation specimens accessioned at the St. Erik Ophthalmic Pathology Laboratory, Stockholm, between 1995 and 2022 under a clinical diagnosis of malignancy were reviewed against pathology reports, referral letters, and clinical records. Predefined criteria identified eyes in which a malignant tumor drove enucleation but was not confirmed histopathologically. In an exploratory analysis, the authors rated the projected visual acuity had enucleation been deferred. The misdiagnosis rate was modeled against calendar year by logistic regression and tested in sensitivity analyses. Results Of 24,242 specimens, 1,661 were enucleations, and after predefined exclusions, 1,419 eyes had been enucleated for a presumed tumor. Thirty-nine (2.7%) were misdiagnoses: 21 diagnosed as UM, 11 with another suspected neoplasm, and 7 with a suspected unspecified malignancy. Intraocular hemorrhage was the lesion most often mistaken for UM. Among the 21 μm eyes, the projected visual impact of enucleation was nil in 13 that had no useful vision at enucleation, a loss of useful vision in 1 (decimal acuity 0.25, a benign leiomyoma), and not estimable in 7. The misdiagnosis rate did not change appreciably (odds ratio 0.97 per year, 95% CI 0.94 to 1.01, P = 0.16), remained between 2.3% and 4.2% in sensitivity analyses, and did not differ between 1995 and 2010 (2.8%) and 2011–2022 (1.8%; P = 0.25). Misdiagnosed eyes were almost always blind, painful, or had opaque media, where clinical and imaging assessment is least reliable. Conclusions Clinical misdiagnosis preceded 2.7% of tumor-related enucleations over 28 years, at a stable rate, and intraocular hemorrhage was the most frequent mimic of UM. In an exploratory projection based on incomplete records, misdiagnosis seldom appeared to sacrifice useful vision. The findings reinforce the value of multimodal imaging, ocular-oncology consultation, and consideration of biopsy before enucleation when the diagnosis is uncertain.