Ocular oncology and the study of rare cancers.

The seven papers in this issue of Cancer Control deal with six different types of neoplasms of the eye and orbit: choroidal melanoma, retinoblastoma, primary ocular lym-phoma, conjunctival melanoma, orbital rhabdomyosarco-ma, and orbital meningioma. These tumors all have an obvious anatomic relationship to the anterior visual pathway , yet each has a different histogenesis, clinical presentation , pathology, and prognosis for vision and life. Despite their differences, these tumors share one important feature: they are all uncommon, if not rare conditions. The American Cancer Society estimated that approximately 2,200 new cases of all types of primary ocular and orbital malignancy were diagnosed in the United States in 2002, based on incidence rates from the Surveillance, Epidemiology and End Results Program of the National Cancer Institute. 1 Uveal melanoma is the single most common malignancy of the eye and orbit and makes up nearly 70% of all malignancies of these tissues, or about 1,600 new cases each year. 2 By comparison, the age-adjusted incidence of cutaneous melanoma alone is 20 times greater than that of all types of ocular and orbital cancer combined. 2 In an era when physicians and patients rely on evidence based medicine to guide them in clinical decision making, uncommon cancers present challenging problems in how to design and conduct clinical trials. Until relatively recently, essentially all clinical decisions about the effectiveness of therapies for ocular cancer were based on retrospective series or single-armed studies with historical controls. The superiority of randomized clinical trials in determining effectiveness of new treatments is well established , but when dealing with uncommon conditions it may not be possible to recruit enough patients into a study to have sufficient statistical power to detect a treatment effect. In cancer trials, the dilemma is more perplexing because even a small risk reduction in the primary outcome of death is important to validate. The detection of a small treatment effect, however, requires enrollment of large numbers of patients, which is not realistic for most ocular cancers. Further complicating the assessment of treatment of ocular cancer is the real possibility that therapy itself will result in loss of vision or loss of an eye. Investigators have found that vision-related quality of life assessment is another measure of clinical outcome that may be just as important as other traditional measurements, including longevi-ty of life. 3 Clinical trialists are now developing and validating tools that can reliably …