Date of Award

2-15-2010

Document Type

Dissertation

Degree Name

Master of Medicine (MMed)

First Supervisor/Advisor

Sudhir Vinayak

Department

Imaging and Diagnostic Radiology (East Africa)

Abstract

Purpose of the study: To which of the ACR sonographic BIRADS lexicon descriptors can be used reliably to differentiate benign from malignant solid breast masses.

Objectives: 1.Main Objective: To determine the association of sonographic ACR BI-RADS descriptors and solid breast masses 2. Specific Objective: To determine the predictive probabilities of sonographic BI-RADS lexicon descriptors

Methodology: This was a 15 months prospective cross sectional study in which 125 consecutive patients who met the inclusion criteria were enrolled. The study was carried out at theAga KhanUniversity hospitalradiology department between October 2008 and December 2009 inclusive. The scans were performed in ourradiology department by residents and consultant radiologists. The sonographic BI-RADS descriptors were then assigned to the lesions based on the consensus arrived at between the principal investigator and the consultant radiologist. The findings were then correlated with the histological/cytological diagnosis which was the gold standard. The study was approved by the University Research and Ethics Committee

Results: 66% (n=82) of the patients turned out to have benign lesions while 34% (n=43) had malignant breast lesions. All the sonographic BI-RADS descriptor variables demonstrated significant association with the histology results apart from vascularity. The predictive probability of malignancy was lowest (8.36 %) for lesions with parallel orientation and well circumscribed margins and highest (73.89%) for masses with non parallel and poorly circumscribed margins.

Conclusion: There is a significant association between all the sonographic ACR BI-RADS lexicon descriptors and solid breast masses apart from one descriptor: vascularity. Lesions with both non parallel orientation and non circumscribed margins have the highest predictive probability for malignancy.

Download

Share

COinS