• Karla Kerlikowske, Rebecca Smith-Bindman, Britt-Marie Ljung, and Deborah Grady.
• University of California, San Francisco, General Internal Medicine Section, Department of Veterans Affairs, San Francisco, California 94121, USA. kerliko@itsa.ucsf.edu
• Ann. Intern. Med. 2003 Aug 19;139(4):274-84.

BackgroundBecause approximately 1 in 10 women with a breast lump or abnormal mammography result will have breast cancer, a series of decisions must be taken by a primary care practitioner to exclude or establish a diagnosis of breast cancer among these women.PurposeTo determine the most accurate and least invasive means to evaluate an abnormal mammography result and a palpable breast abnormality.Data SourceMEDLINE search (January 1966 to March 2003) for articles and reviews describing the accuracy of clinical examination, biopsy procedures, and radiographic examination for patients with abnormal mammography results or palpable breast abnormalities.Study SelectionThe authors reviewed abstracts and selected articles that provided relevant primary data. Studies were included if 1) mammography, fine-needle aspiration biopsy, or core-needle biopsy was performed before a definitive diagnosis was obtained; 2) the study sample included 100 or more women; and 3) breast cancer status was determined from histopathology review of excisional biopsy specimens, from linkage with a state cancer registry or the Surveillance, Epidemiology, and End Results program, or from clinical follow-up of 95% or more of the study sample.Data ExtractionOne investigator abstracted results. Methods were evaluated for major potential biases, but methodologic scoring was not performed.Data SynthesisLikelihood ratios for first screening mammography were 0.1 for the Breast Imaging Reporting and Data System (BI-RADS) assessment category "negative or benign finding," 1.2 for "probably benign finding," 7 for "need additional imaging evaluation," 125 for "suspicious abnormality," and 2200 for "highly suggestive of malignancy." For fine-needle aspiration biopsy of a palpable lump performed by formally trained physicians, the likelihood ratio was infinity for an assessment of "malignant," 2.6 for "atypical/suspicious," and 0.02 for "benign." When diagnostic mammography was used to evaluate a palpable lump or nonpalpable breast abnormality, the positive likelihood ratios were 5.6 and 9.4, and the negative likelihood ratios were 0.15 and 0.19, respectively.ConclusionsWomen whose screening mammography results are interpreted as "suspicious abnormality" or "highly suggestive of malignancy" have a high risk for breast cancer and should undergo core-needle biopsy or needle localization with surgical biopsy. Women whose screening mammography results are interpreted as "need additional imaging evaluation" have a moderate risk for breast cancer and should undergo diagnostic mammography or ultrasonography to decide whether a nonpalpable breast lesion should be biopsied. Women whose screening mammography results are interpreted as "probably benign finding" have a low risk for breast cancer and can undergo follow-up mammography in 6 months. Either fine-needle aspiration biopsy or ultrasonography is recommended as the first diagnostic test of a palpable breast abnormality to distinguish simple cysts from solid masses. Fine-needle aspiration biopsy also allows characterization of a solid mass. Diagnostic mammography does not help determine whether a palpable breast mass should be biopsied and should not affect the decision to perform a biopsy.

21 keywords...

hide…