Delineation of the metastatic spread of lung cancer has been attempted by a variety of means. Controversy as to the indications for organ surveys, mediastinoscopy, scintillation scanning, and biopsy techniques still exists. Since definition of the micrometastatic state is yet unachieved, the staging of disease for therapy continues to be predicted on documented spread beyond the site of the origin. The records of 75 patients in whom the presence or absence of mediastinal metastases was known were retrospectively reviewed to establish the sensitivity, specificity, predictive values, and accuracy of 67Ga scintigrams. Comparisons were made with chest roentgenograms, mediastinal tomograms, and endoscopic findings. Five patients had 67Ga-negative studies. In the 70 patients having 67Ga-positive lung lesions, mediastinal 67Ga uptake had a sensitivity of 88%, a specificity of 86%, predictive values of 93% for a positive test and 76% for a negative test, and a test accuracy of 87%. These studies, plus those of others, permit selectivity of choosing candidates for mediastinoscopy. If the primary is 67Ga positive, a negative mediastinal scan obviates mediastinoscopy. If the mediastinum is 67Ga positive, mediastinal exploration is indicated. The level of involvement dictates whether resection is undertaken in suitable surgical candidates. This approach, employed since 1976, has lowered the costs of staging, and 67Ga has become our scintigram of choice.