Comparison of the diagnostic performance of abbreviated MRI and full diagnostic MRI using a computer-aided diagnosis (CAD) system in patients with a personal history of breast cancer: the effect of CAD-generated kinetic features on reader performance

Abstract

AIM: To compare the diagnostic performance of abbreviated magnetic resonance imaging (MRI) and full diagnostic MRI with computer-aided diagnosis (CAD) system in patients with a personal history of breast cancer and to evaluate how the kinetic features affect the performance of two radiologists. MATERIALS AND METHODS: Between 1 January 2014 and 31 December 2017, 3,834 breast MRI examinations in 2,310 patients with a history of breast cancer comprised the study population. MRI images were reviewed retrospectively by two radiologists. First, two radiologists independently reviewed T1-weighted images scanned at 90 seconds after the contrast medium injection and T2-weighted images. After 6 months, the two readers reviewed contrast-enhanced T1-weighted images with five consecutive delayed images using CAD. The diagnostic performance of the abbreviated-sequence and full-sequence MRI were compared. RESULTS: Fifty-one intramammary recurrences were detected with breast MRI in 47 patients. Of the 51 tumour recurrences, 36 (70.6%) lesions occurred >3 years after initial cancer surgery and seven (13.7%) lesions at <2 years after initial surgery. The sensitivity and specificity were 92.2-94.1% and 97.6-98.6% on the abbreviated sequence and 94.1-96.1% and 97.9-98.3% on the full diagnostic MRI. Of 51 malignant lesions, six showed delayed persistent pattern, of which three lesions were non-mass enhancement and three lesions were small enhancing masses <1 cm. CONCLUSION: Overall diagnostic performances of abbreviated MRI and full diagnostic MRI were similar in both readers. The CAD-generated kinetic features could affect reader performance and the sensitivity could be improved or the specificity improved according to the reader.