Case Reports and Small Case Series
January 2001

Cardiac Metastasis From Uveal Melanoma

Author Affiliations

Copyright 2001 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2001

Arch Ophthalmol. 2001;119(1):139-140. doi:

Ruiz and coworkers1 recently described an elderly woman who developed cardiac metastasis from choroidal melanoma and died of a ruptured myocardium. Because of a scarcity of other reports, they concluded that cardiac metastasis from uveal melanoma was extremely rare, perhaps with the exception of patients with widespread micrometastases that might go undetected without autopsy.1 In our experience, macroscopically detectable metastasis from uveal melanoma to the heart is not uncommon, although it is true that it has seldom been reported in the literature. We are aware of 4 reports of cardiac metastasis from uveal melanoma in 14 patients.14 A fifth article mentions 2 patients with cardiac metastasis from an ocular primary melanoma, but it does not specify whether the primary site was the uvea or conjunctiva.5 The purpose of the present report is to describe 5 patients with cardiac metastasis and to estimate the frequency of metastasis to the heart from uveal melanoma.

Report of Cases

To get a reliable estimate of the spectrum and frequency of cardiac metastasis, we reviewed clinical and autopsy data of a consecutive series of 167 patients who had an eye enucleated between 1972 and 1981 because of a choroidal or ciliary body melanoma.6 These patients have complete, validated follow-up data. By December 1999, tumors had disseminated in 80 patients (48%). Autopsy had been performed on 27 (34%) of these 80 patients, who all died of disseminated uveal melanoma. Sites of metastases that were identified clinically or by macroscopic examination at autopsy were recorded. Metastases that were detected only incidentally by microscopy, including 1 in the heart, were disregarded. Statistical analysis was based on exact probability distributions (StatXact-3; Cytel Software, Cambridge, Mass).

Five patients (Table 1) had cardiac metastases, all of which were clinically asymptomatic. Macroscopic autopsy reports gave the following descriptions: Case 1. Three pigmented metastases in the left ventricular wall. Case 2. Multiple pigmented metastases 2 to 3 mm in diameter in the left ventricular epicardium and myocardium, and a subendocardial metastasis 2 mm in diameter in the anterior papillary muscle. Case 3. Multiple, small metastases in the pericardium, epicardium, and myocardium (Figure 1) in the apex of the heart. Case 4. One pigmented myocardial metastasis 1 cm in diameter in the posterior wall of the heart, together with several endo- and pericardial metastases 2 to 3 mm in diameter. Case 5. A solitary metastasis in the epicardium.

Image not available
Clinical Characteristics of 5 Patients With Malignant Uveal Melanoma*
Figure 1.
Image not available

Epithelioid melanoma cells infiltrate within the myocardium (A) diffusely (hematoxylin-eosin, original magnification ×225) and (B) as small nodules (hematoxylin-eosin, original magnification ×360) (case 3).

The frequency of cardiac metastasis at autopsy was 19% (5 of 27 cases) (95% confidence interval [CI], 6%-38%). This finding is essentially identical to that of a 1963 population-based Danish study, which reported pericardial and myocardial metastases in 9 (24%) of 38 patients who were autopsied (95% CI, 11%-40%).2 The frequency of cardiac metastasis in the entire series of 167 patients who had an eye enucleated was thus 3.0% (95% CI, 1%-7%), and the frequency of symptomatic cardiac metastasis was 0% (95% CI, 0%-2%). We recognize the limitations of our study in that the autopsy rate was not high and patients were not selected for autopsy at random. Autopsies were done on an individual basis according to the preference of the attending physician and the relatives of the deceased.

Retrospective evidence did not suggest any obvious bias regarding selection of patients for autopsy. Of the 27 patients who underwent autopsy, 5 were autopsied without suspicion of metastasis, which argues against selection for autopsy of patients with known widespread dissemination. Furthermore, no difference between autopsied and nonautopsied patients who died of disseminated uveal melanoma was observed in the age at enucleation (P = .25, Mann-Whitney test), the presence of metastatic symptoms (P = .16, Pearson χ2 test), or the time from enucleation to the onset of symptoms and signs of metastatic disease (P = .96, log-rank test). Comparison of the number of clinically known metastases at the time of death between autopsied and nonautopsied patients indicated no statistically significant difference between groups (mean, 1.3 vs 1.7, respectively; P = .18, Mann-Whitney test).

Presence of cardiac metastasis was statistically significantly associated with a large number of macroscopic metastatic sites detected at autopsy (Figure 2) (P<.001, Pearson χ2 test). None (95% CI, 0%-17%) of the 20 patients who had 3 or fewer metastatic sites had cardiac metastases, whereas 5 (71%; 95% CI, 29%-96%) of 7 patients with 4 or more metastatic sites had cardiac metastases. Presence of cardiac metastases among the 80 patients who died of disseminated uveal melanoma was not associated with the age at diagnosis of uveal melanoma (P = .47, Mann-Whitney test) or with the time from enucleation to death (P = .82, log-rank test).

Figure 2.
Image not available

Number of macroscopic metastatic sites detected in 27 autopsied patients with metastatic uveal melanoma. Except for one patient with cerebral metastasis, 10 of the 11 patients who had only one metastatic site had hepatic metastases. The 5 patients who had cardiac metastases had a higher number of metastatic sites than those who did not (P<.001, Pearson χ2 test). Bar indicates median.


Based on our data, we estimate that cardiac metastasis from uveal melanoma is found in approximately one fifth of patients who die of this tumor. It is possible that the ability to spawn widespread visceral metastasis, often with cardiac involvement, is a property developed by a subset of uveal melanomas, as opposed to being the result of longer survival time that would allow more extensive dissemination. Such metastases, however, seldom cause clinical symptoms. In addition to the case report of Ruiz and coworkers,1 we are aware of one lethal cardiac metastasis from uveal melanoma.4 This 69-year-old man developed an atrioventricular block and atrial fibrillation and died of cardiac arrest. At autopsy, he was found to have metastases in the heart, liver, pancreas, adrenal gland, and lymph nodes. The cardiac metastases involved the atrioventricular node and left and right bundle branches.4 On the other hand, patients who die of uveal melanoma are generally 56 years or older, and many of them have concurrent heart disease, which may obscure the origin of cardiac symptoms or of electrocardiographic findings related to metastatic uveal melanoma.

Since submission of the manuscript, Rosario et al7 recently described a woman who experienced syncope attacks because of endocardial and left ventricular metastasis from choroidal melanoma. She died of complications during cardiac surgery.

Corresponding author and reprints: Teemu Mäkitie, MD, Ophthalmic Pathology Laboratory, Department of Ophthalmology, Helsinki University Central Hospital, Haartmaninkatu 4 C, PL 220, FIN-00029 HUS, Helsinki, Finland (e-mail:

Ruiz  RSEl Harazi  SAlbert  DMBryar  PJ Cardiac metastasis of choroidal melanoma. Arch Ophthalmol. 1999;1171558- 1559
Jensen  OA Malignant melanomas of the uvea in Denmark 1943-1952: a clinical, histopathological, and prognostic study. Acta Ophthalmol. 1963;75suppl1- 220
Einhorn  LHBurgess  MAGottlieb  JA Metastatic patterns of choroidal melanoma. Cancer. 1974;341001- 1004Article
Matturri  LNappo  AVaresi  CRossi  L Metastasi cardiache di melanoma maligno. Pathologica. 1993;8597- 102
Glancy  DLRoberts  WC The heart in malignant melanoma: a study of 70 autopsy cases. Am J Cardiol. 1968;21555- 571Article
Mäkitie  TSummanen  PTarkkanen  AKivelä  T Microvascular loops and networks as prognostic indicators in choroidal and ciliary body melanomas. J Natl Cancer Inst. 1999;91359- 367Article
Rosario  RTDiMaio  DJLapham  RLSweeney  MSmalling  RBarasch  E Metastatic ocular melanoma to the left ventricle inducing near-syncope attacks in an 84-year-old woman. Chest. 2000;118551- 553Article