Overview
The American College of Chest Physicians (ACCP) chose the Duke University Center for Clinical Health Policy Research to perform formal systematic reviews of the current evidence in the five new non-small cell lung cancer (NSCLC) topic areas, as well as to provide a search for the existing guidelines, systematic reviews, and meta-analyses in all of the topics areas. In addition, the Agency for Healthcare Quality and Research) AHRQ agreed to fund the BlueCross BlueShield Association Technology Evaluation Center to perform the formal systematic review of literature on small cell lung cancer (SCLC). The Health Outcomes Research Group of the Department of Epidemiology and Biostatistics at Memorial Sloan-Kettering Cancer Center conducted a full-scale review of the literature since the first set of guidelines in the area of screening for lung cancer to assist that particular writing group.
The formal systematic reviews of the five new topic areas were guided by the appropriate chapter editors and their writing committees, in concert with the Executive Committee of the panel.
The two EPC research teams conducted a variety of systematic computerized bibliographic database searches including the following: (1) a search for systematic reviews, guidelines, and meta-analyses published since the last ACCP lung cancer guideline (MEDLINE, The Cochrane Library, National Guidelines Clearinghouse); (2) targeted searches for reviews in each of five selected treatment sections (solitary pulmonary nodules, stage I and II, stage IIIA, stage IIIB, stage IV); these searches, run in OVID version of MEDLINE, were performed in July and August 2005 and were limited to publication years since 1995, English language, and human subjects; and (3) searches related to SCLC are described in the evidence chapter on SCLC.
Search terms included the medical subject heading terms lung neoplasms (exploded) and bronchial neoplasms for the lung cancer concept. Each topic search utilized key words specific to the key questions of interest (complete search strategies are available on request from the authors).
Strategy Specific for Invasive Mediastinal Staging of Lung Cancer
The data presented here are based on a systematic search and evaluation of the published literature from January 1980 through June 2006. Articles published prior to July 2001 were identified according to the criteria laid out in the previous version of the American College of Chest Physicians lung cancer guidelines. Subsequent literature was identified by the authors using the same search strategy and selection criteria (briefly, studies published in the English language, peer-reviewed, nonoverlapping, having at least 20 patients, containing an adequate assessment of the true nodal status, and with the ability to calculate performance characteristics).
The data abstraction was performed for patients suspected of having lung cancer (e.g., non-small cell lung cancer [NSCLC] and small cell lung cancer [SCLC]). Patients suspected of a diagnosis other than lung cancer were excluded from the study, where possible. A definite diagnosis of any lung cancer in the mediastinal tissues was considered to be positive, while other diagnoses (e.g., benign disease or lymphoma) were coded as negative for lung cancer. Equivocal test results were considered to be negative. Biopsies that were aborted or yielded insufficient tissue are included as negative findings and are counted as such in the statistics. The reported feasibility of the test is also reported (i.e., the proportion of patients undergoing the test in whom an adequate biopsy was able to be obtained) in order to have an assessment of the technical success rate. The calculation of the subtotal or total summary performance characteristics was accomplished by the calculation of an average of the values (e.g., of sensitivity and specificity) from each study; in other words, no weighting according to study size was performed. This was chosen for simplicity, and because a comparison of the results using both methods revealed minimal differences (i.e., 1 to 2 percentage points).