T here is controversy regarding the reliability of 67Ga scanning for detecting hilar and mediastinal metastases in lung cancer. Approximately 90 percent of primary lung carcinomas ofall histologic types are avid for gallium,’ a finding confirmed by numerous investigators. However, there is considerably more variation in the figures reported in recent studies of detection of hilar and mediastinal metastases (Fig 1) with gallium. One study by Alazraki et al2 compared 67Ga scan results with mediastinoscopy in 25 patients with nonsmall-cell lung carcinoma. Eleven of the 25 patients had tumor deposits in the mediastinum, and each of these II had a positive gallium scan. Fourteen ofthe 25 patients had no tumor deposits in the mediastinum at mediastmnoscopy; of these patients, four had positive scans. The true positive ratio, or sensitivity, was 100 percent in this study. The true negative ratio, or specificity, was 71 percent. The authors concluded that a negative gallium scan obviates the need for mediastinoscopy, and that such patients may be referred directly for thoracotomy. This clear-cut result is not borne out by other studies. Reported sensitivity ranges from 30 percent to 100 percent (mean, 80 percent) while the range of specificity is from 50 percent to 94 percent (mean, 71 percent). 3 Gallium scanning cannot be used as the sole method for selection of patients for curative surgery. Nevertheless, the results appear good enough to justify the use ofgallium scanning in at least some cases. An attempt to use gallium scans to assess response to radiotherapy demonstrated no The published series to date are fairly small, and considerable variation should not be surprising. The type and quality of scanning equipment used and the experience of the nuclear physician are significant variables. The characteristics ofthe primary tumor are also important. Tumors near the mediastinum appear to be the most difficult to evaluate, with higher false positive rates than occur with peripheral primaries.4 Gallium scanning has also been evaluated for the detection of extrathoracic spread of lung carcinoma. The economy and logistic simplicity ofevaluating local and distant metastases with a single imaging procedure are attractive, especially since symptoms may not correctly identify the organ system affected by metastases.’5 In one study ofpatients with oat cell carcinoma, only 7 percent of extrathoracic metastases were found, although 84 percent of mediastinal disease was identifled.3 Another study, which included tumors ofall cell types, showed that 75 percent ofmetastatic sites found by all other clinical methods (isotope scans of individual organs, CT scans, and clinical examination) were detected by whole-body gallium scanning. The wholebody gallium scan accurately identified or excluded extrathoracic metastatic disease in 11 of 12 patients undergoing autopsy within three months ofthe scan.4 In a few patients (with kidney and adrenal metastases),