To assess the role of fusion of two methods of imaging in the evaluation of breast cancer: prone FDG PET/CT and Magnetic Resonance of the breast .
We prospectively recruited 85 women with newly diagnosed breast cancer (cT2- cT3) from the second semester of 2012 to October 2013. Within 10 days dynamic contrast-enhanced MR of the breast and PET/CT were performed. Patients underwent FDG PET/CT for initial staging, acquired as standard whole body PET/CT scan in supine position followed by a delayed prone acquisition 2 hours post-injection with a dedicated breast device the same as the one used in MR. Three groups were created: patients with single lesions (n. 38), with multifocal lesions (n. 16 women with 48 lesions) and with multicentric lesions (n. 31 patients with 71 lesions). There were 13 suspected bilateral malignant lesions and 40 benign neoformations. The image interpretation in PET/CT was visual and semiquantitative with calculation of SUVmax of the lesions. On MR all lesions were classified as mass or nonmass-like according to the Breast Imaging reporting Data System (BI-RADS). In the fusion the only criterion for suspicious lesion was any increase in tracer uptake corresponding to the site of the enhancing lesion detected on MR.
By pathology, 161 lesions were malignant and 53 were benign. The sensitivity, specificity, PPV, NPV and accuracy were: for MR 100%, 75%, 93%, 100% and 94%; for PET/CT 90%, 98%, 99%, 76% and 92%; in fused PET/MR 94%, 100%, 100%, 85% and 96%. PET/CT had a great number of FN in multifocal/multicentric groups, whereas MR had a consistent number of FP in the contralateral breasts. Both results greatly improved using fusion PET/RM.
Integrated information between PET and MR images increased the sensitivity and accuracy of PET, the specificity, PPV and accuracy of MR in diagnostic work-up of breast cancer.