2000 Guideline
The MEDLINE and CancerLit databases (Ovid) were searched from 1966 to July 1998 using the medical subject heading (MeSH) "breast neoplasms," "neoplasm staging," "neoplasm metastasis," "bone neoplasms/sc," "liver neoplasms/sc," and "lung neoplasms/sc," and the text words "preop:," "stag:," and "baseline." The search was updated in March and November, 1999 and again in April 2000. These terms were also used to search the Cochrane Library (1999, Issues 1 and 4 and 2000, Issue 1). Articles found by the searches, cited in the relevant papers, or known to the lead author of the practice guideline were retrieved and reviewed.
The literature search described above was not restricted by language; it uncovered three reports on bone scanning published in French and one in German. Because a large body of literature published in English was available and resources for translation were limited, these foreign-language publications were excluded from the practice guideline.
2003 Update
The literature search was updated using subject headings (breast neoplasms, neoplasm staging, neoplasm metastases, bone neoplasms/sc, liver neoplasms/sc, lung neoplasms/sc, clinical trial{s}, exp evaluation studies) and text words (breast, mammary, cancer, carcinoma, neoplasm, stag:, baseline). MEDLINE (2000-April 2003), the Cochrane Library (Issue 1, 2003), and the PDQ Clinical Trials Database (www.cancer.gov/search/clinical_trials, accessed April 30, 2003) were searched for clinical trials of bone scanning, liver ultrasonography or chest radiography as staging tests in breast cancer.
Inclusion Criteria
Studies were eligible for inclusion in this overview of the evidence if they reported the number of women with newly diagnosed breast cancer who had metastases detected by bone scan, liver ultrasound, or chest radiograph. These tests could be performed either before or after surgery. Both full reports and abstracts were eligible. Studies were included only if they reported the rates of positive tests by stage of disease and the staging system was similar to that currently in use (see Appendix 1 in the original guideline document).
The primary outcome of interest was the detection rate, that is, the number of patients with abnormal tests that were indicative of metastases divided by the total number of patients tested. Detection rates were calculated by the guideline authors from data appearing in the study reports. Also of interest were the false-positive and the false-negative rates; these were given in some of the study reports reviewed for this guideline.