THE VALUE OF ASPIRATION CYTOLOGIC EXAMINATION OF THE BREAST - A STATISTICAL REVIEW OF THE MEDICAL LITERATURE

Citation
Rwm. Giard et J. Hermans, THE VALUE OF ASPIRATION CYTOLOGIC EXAMINATION OF THE BREAST - A STATISTICAL REVIEW OF THE MEDICAL LITERATURE, Cancer, 69(8), 1992, pp. 2104-2110
Citations number
41
Language
INGLESE
art.tipo
Review
Journal title
CancerACNP
ISSN journal
0008-543X
Volume
69
Issue
8
Year of publication
1992
Pages
2104 - 2110
Database
ISI
SICI code
0008-543X(1992)69:8<2104:TVOACE>2.0.ZU;2-4
Abstract
The decision to perform surgery in patients with a breast mass usually is made on the basis of combined diagnostic information, with fine-ne edle aspiration cytologic examination (FNAC) playing a central role. T o determine and compare the quality of FNAC of the breast, a search wa s performed of the English literature for articles with quantitative i nformation about their results. Twenty-nine such articles, containing 31,340 aspirations, were identified and summarized. Required data were extracted from these articles. These numbers were analyzed with the u se of a two-by-four contingency table to relate the FNAC result (defin itely malignant, suspect, benign, or unsatisfactory cytologic material ) with the final diagnosis (malignant or benign breast disease). Test characteristics such as sensitivity, specificity, and the likelihood r atios for the four different FNAC results were derived for each study and compared. There was a striking difference between studies with reg ard to the probability of a particular FNAC upshot (e.g., in patients with breast cancer, the chance of obtaining definitely malignant cytol ogic material ranged from 0.35 to 0.92), the sensitivity (range, 0.65 to 0.98), the specificity (range, 0.34 to 1.0), and likelihood ratios. In the opinion of the authors, it is virtually impossible to infer ge neral test characteristics of FNAC of the breast from the medical lite rature because of differences in methods and different biases. At best , the maximum attainable performance of this test can be described. Fo r the development of a policy for breast mass management, the local te st characteristics of this highly operator-dependent test should be es tablished.