Although a standardized and uniformly accepted cancer staging system is an essential and fundamental requirement to enable meaningful comparisons across patient populations, the sometimes capricious biologic behavior of melanoma makes developing such a staging system particularly difficult. Since the earliest well-documented attempts at classifying patients with cutaneous melanoma were described more than 50 years ago, the identification of increasingly powerful prognostic factors has led to sequential modifications of the cutaneous melanoma staging system. The current AJCC staging system is based on relatively well-established prognostic factors; however, several recent reports have identified additional prognostic factors not included in the current system, and other studies support the re-evaluation of some of the currently employed staging criteria. Some of the more controversial areas include the relevance of level of invasion versus tumor thickness, optimal cutoffs for tumor thickness, importance of ulceration, the grouping of satellites with in-transit metastases, the inclusion of microsatellites and local recurrences as a separate staging criterion, the replacement of size of nodal mass with number of positive nodes, the importance of nodal metastases in more than one nodal basin, and the prognostic significance of distant metastases. Therefore, future modifications of the staging system are anticipated to better incorporate these observations. Stage-specific staging recommendations for the patient with melanoma provide the clinician with a framework to most efficiently assess extent of disease in an era of cost-conscious clinical practice. In the asymptomatic patient with primary melanoma (stage I or II), we recommend a chest roentgenogram and evaluation of alkaline phosphatase and LDH levels; extensive radiologic evaluations are not indicated, because the rate of detection in this population is extremely low. Additional staging information should also be obtained by the technique of lymphatic mapping and sentinel lymphadenectomy. For patients with local-regional disease (stage III, satellites, and local recurrence), a selective approach to imaging studies is warranted. For this patient population, we recommend complete blood count, liver function tests including alkaline phosphatase and LDH, a chest roentgenogram, and a CT scan of the abdomen. Although the yield of these tests, particularly CT of the abdomen, in detecting distant metastases in asymptomatic patients is low, they may identify false-positive abnormalities and provide an important baseline for future studies in this high-risk population. For patients with disease below the waist or in the head and neck region, we recommend CT of the pelvis and CT of the neck, respectively. Additional studies should be done only if clinically indicated. Finally, patients with known systemic disease (stage IV) should be more comprehensively evaluated, because the likelihood of detecting asymptomatic metastases is higher. Accordingly, in addition to the work-up outlined previously for stage III patients, we also perform a CT scan of the chest and MR imaging of the brain; other studies (e.g., bone scan, gastrointestinal series) are performed on the basis of symptoms.