Multimodality systematic approach to mediastinal lymph node staging in non-small cell lung cancer.

Establishing an accurate diagnosis and stage for non-small cell lung cancer has important implications for treatment and prognosis. Ideally, the process should be performed in a way that maximizes the information from each procedure while minimizing the risk to the patient. The concepts of decision analysis and Bayes' theorem form a basis to develop the strategy. In this framework, the pre-test probability of malignancy is estimated in the lung nodule or mass, the regional lymph nodes and in distant sites. Invasive diagnostic tests are performed in sites with a pre-test probability greater than the testing threshold, beginning with those sites that would yield the highest stage, if positive. Modalities are chosen that are able to biopsy the suspicious sites and present the least amount of risk to the patient. Following each test, the post-test probability of malignancy is calculated to determine if it crosses the testing or test-treatment thresholds. The process continues with further tests until a diagnosis and stage are established.