Performance of users in tropical areas with the BI-RADS classification of breast lesions for predicting malignancy

To evaluate the diagnostic performance of radiologists in Cameroon using the BI-RADS classification to interpret mammograms and ultrasound scans together for the prediction of malignant breast lesions. This cross-sectional study took place at the Women's and Children's Hospital in Yaounde...

Full description

Saved in:
Bibliographic Details
Published in:Médecine et santé tropicales (Montrouge, France) Vol. 23; no. 4; p. 439
Main Authors: Gonsu Kamga, J E, Moifo, B, Sando, Z, Guegang Goudjou, E, Nko'o Amvene, S, Gonsu Fotsin, J
Format: Journal Article
Language:French
Published: France 01-10-2013
Subjects:
Online Access:Get more information
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:To evaluate the diagnostic performance of radiologists in Cameroon using the BI-RADS classification to interpret mammograms and ultrasound scans together for the prediction of malignant breast lesions. This cross-sectional study took place at the Women's and Children's Hospital in Yaounde from July 2009 to April 2010 and included 211 women with breast lesions identified on mammograms during a breast screening campaign and subsequently assessed with ultrasonography and histology. The BI-RADS classifications of these lesions were compared to the corresponding histology results to evaluate the accuracy of predictions of malignancy from the mammograms and ultrasound scans interpreted with the BI-RADS system. The rate of malignancy in each ACR-classified category was also compared to the standard ACR categories as stipulated in the ACR classification. In all, 339 women aged from 16 to 78 years were screened, and lesions requiring biopsies were identified for 211. The age group included most often was the 41-50 year-old group (n = 98, 46.4%). Overall, 135 (64%) women had benign lesions and 76 (36%) malignant. Invasive carcinoma was found in 49 (65%) of the malignant lesions, in situ intraductal carcinoma in 23 (30%), and sarcoma in 4 (5%). Based on the BI-RADS classification, 124 (58.7%) breast lesions were classified as ACR2, 15 (7.1%) as ACR3, 44 (20.8%) as ACR4, and 28 (13.3%) as ACR5. Comparison of the BI-RADS classification and the histological findings showed that 19% of ACR2-classified lesions were malignant, 13% of those classified ACR3, 66% ACR4, and 75% ACR5. The global accuracy in the prediction of malignancy the BI-RADS classification was 77.3%. The accuracy of the radiologists using the BI-RADS classification in our hospital was good at 77.3%, although shortcomings in the evaluation and interpretation of some lesions resulted in a relatively high prevalence of malignant lesions in categories ACR2 and ACR3.
ISSN:2261-2211
DOI:10.1684/mst.2013.0251