Previous studies of cutaneous eyelid lesions have been retrospective using multiple contributing surgeons. The purpose of this study was to determine prospectively the accuracy of the clinical diagnosis of benign, premalignant, and malignant cutaneous eyelid lesions and to determine if all clinical diagnoses require histopathologic confirmation, or if an experienced clinician can reliably distinguish benign from malignant lesions on the basis of history and clinical examination alone. This prospective study was conducted between January 1988 and January 1995. All patients presenting during this time with periocular cutaneous eyelid lesions were evaluated and a specific clinical diagnosis made in each case before biopsy and histopathologic evaluation. The lesions were categorized prospectively as benign, premalignant, or malignant, and histopathologic evaluation then correlated to determine the accuracy of the clinical diagnosis. Only patients who presented without previous biopsy were eligible for inclusion in the study. A biopsy was done on a total of 864 eyelid lesions during the 85 month study period. One hundred fifty-three lesions clinically were thought to represent malignancies. Of these, 140 (91.5%) were found to have malignant histopathologies. Nineteen lesions clinically were thought to represent premalignant processes. Histopathologic evaluation of these 19 lesions showed 16 to be actually premalignant, 1 to be a malignancy, and 2 to be benign. Six hundred ninety-two lesions clinically were thought to be benign. Of these, 13 (1.9%) proved on histopathologic evaluation to be malignant. These included 10 basal cell carcinomas, 1 squamous cell carcinoma, 1 non-Hodgkin lymphoma, and 1 adenoid cystic carcinoma. Three (0.4%) of the 692 clinically benign lesions were found to be premalignant. The 13 missed malignancies were distributed among a number of different clinical diagnoses, including papilloma, epidermal inclusion cyst, melanocytic nevus, hydrocystoma, and trichoepithelioma. Of the 153 clinically malignant lesions, 6 lesions that clinically were highly suspicious for malignancy had initial benign histopathologic diagnoses. Rebiopsy results in all of these six subsequently confirmed the suspected malignant diagnosis. Malignant eyelid lesions may masquerade as a number of different clinically benign conditions. The authors conclude that all excised eyelid lesions should be submitted for histopathologic confirmation because it is not possible to obtain 100% accuracy in diagnosing eyelid lesions on clinical grounds alone. However, strong clinical suspicion of a malignancy is highly significant, and if initial histopathologic evaluation does not agree with the malignant clinical diagnosis, repeat biopsy should be performed.