Analysis of published studies on the detection of extrathoracic metastases in patients presumed to have operable non-small cell lung cancer

Abstract

Background: A study was undertaken to determine the proportion of patients with potentially operable non-small cell lung cancer that could be spared thoracotomy by a systematic search for extrathoracic metastases.

Methods: An English language literature search was carried out using MEDLINE (1966-91) and bibliographic reviews of textbooks, review articles, and key articles. Studies were included in which at least 90% of the patients had histologically proven non-small cell cancer of the lung, were presumed otherwise operable, and for which the results of computed tomography of the head or abdomen, ultrasonography of the abdomen, or radionuclide imaging (scan) of bone or liver were available. Study quality and specific descriptive information concerning population, intervention, and outcome measurements were assessed.

Results: Of approximately 1500 citations which were screened, 100 were reviewed in detail and data abstracted from 16. The number of patients (total number, followed in square brackets by number of asymptomatic patients) shown to be inoperable directly as a result of the investigation and thus spared unnecessary thoracotomy was: computed tomography of the head: 26/785 (3.3%), 95% confidence interval (CI) 2.1% to 4.4% [14/353 (4.0%), 95% CI 2% to 6%], computed tomography of the adrenal glands: 30/632 (4.7%), 95% CI 3.0% to 6.4% [number asymptomatic indeterminate], bone scan: 45/480 (9.3%), 95% CI 6.7% to 12% [9/301 (3.0%), 95% CI 1.1% to 4.9%], liver imaging: 12/529 (2.3%), 95% CI 0.9% to 3.3% [4/268 (1.5%), 95% CI 0.1% to 3%].

Conclusions: A study with a large sample size and preferably incorporating thoracic computed tomography is required to narrow the confidence intervals around each test. All tests may play an important part in the preoperative evaluation of patients with non-small cell carcinoma of the lung who are presumed to be operable, including asymptomatic patients. Limitations of present data preclude definitive recommendations for asymptomatic patients.