What is the sensitivity of mammography and dynamic MR imaging for DCIS if the whole-breast histopathology is used as a reference standard?

Abstract

Our purpose was to compare mammography and dynamic contrast-enhanced magnetic resonance imaging (MRI) in the detection of ductal carcinoma in situ (DCIS). Ninety patients (aged 58.6+/-16.1 years) who were candidates for unilateral (n=81) or bilateral (n=9) mastectomy underwent mammography and dynamic contrast-enhanced breast MRI using a coronal three-dimensional gradient-echo sequence with slice thickness < or =3 mm before and after intravenous injection of gadoteridol (0.1 mmol/kg). Mammographic and MR images were evaluated by two offsite readers working in consensus. Pathological examination performed on 5-mm sections covering the whole breast was used as a reference standard. Out of 99 breasts, pathology revealed 26 DCIS in 14 breasts of 14 patients, aged 52.0 +/- 9.6 years. Lesion diameter at pathology was <5 mm (n=4); > or =5 and <10 mm (n=7); > or =10 and <20 mm (n=3); > or =20 mm (n=2); not assessed (n=10). Sensitivity was 35% (9/26) for mammography and 38% (10/26) for MRI (not significant difference, McNemar test). Both mammography and MRI provided a true positive result in seven cases (four of them measured at pathology, with a diameter of 20.0+/-12.9 mm; median 20 mm) and a false negative result in 14 cases (10 of them measured at pathology, with a diameter of 4.2+/-1.9 mm; median 4.6 mm) (p=0.024, Mann-Whitney U test). Only 46% (12/26) of DCIS were detected at mammography and/or MRI; the remaining 54% (14/26) were diagnosed only at pathological examination. When the whole breast is used as the histopathological reference standard, both mammography and MRI show low sensitivity for DCIS.