Effect of clinical information on diagnostic performance in breast sonography

Abstract

OBJECTIVE: The purpose of this study was to assess whether the clinical information (CI) of patients affects the degree of suspicion for malignancy by radiologists performing breast sonography.

METHODS: We included 150 breast lesions in 144 patients who underwent breast sonography and sonographically guided core needle biopsy. A pathologic diagnosis was available for all 150 breast lesions: 78 (52%) were malignant, and 72 (48%) were benign. Three radiologists retrospectively reviewed the sonograms of all lesions twice at 8-week intervals first without any CI for the patients (first review) and then with CI such as patient age, palpability, and personal history of risk factors for breast cancer (second review). The reviewers categorized the final assessment according to the American College of Radiology Breast Imaging Reporting and Data System. We compared diagnostic performance such as sensitivity and specificity and the degree of suspicion for malignancy between the image reviews with and without CI.

RESULTS: In the second review, sensitivity was improved in all 3 reviewers (94.0 to 99.2%; P < .05), and specificity was decreased (39.8 to 30.8%; P = .04). There was a significant increase of suspicion for malignancy with the patients' CI (P < .05).

CONCLUSIONS: Clinical information about a patient's breast cancer history and clinical presentation with a palpable mass can increase the suspicion for malignancy on sonography and the sensitivity of sonographic interpretation.