Positron emission mammography in breast cancer presurgical planning: Comparisons with magnetic resonance imaging

Abstract

Purpose: The objective of this study was to compare the performance characteristics of 18F-fluorodeoxyglucose (FDG) positron emission mammography (PEM) with breast magnetic resonance imaging (MRI) as a presurgical imaging and planning option for index and ipsilateral lesions in patients with newly diagnosed, biopsy-proven breast cancer. Methods: Two hundred and eight women >25 years of age (median age = 59.7 ± 14.1 years) with biopsy-proven primary breast cancer enrolled in this prospective, single-site study. MRI, PEM, and whole-body positron emission tomography (WBPET) were conducted on each patient within 7 business days. PEM and WBPET images were acquired on the same day after intravenous administration of 370 MBq of FDG (median = 432.9 MBq). PEM and MRI images were blindly evaluated, compared with final surgical histopathology, and the sensitivity determined. Substudy analysis compared the sensitivity of PEM versus MRI in patients with different menopausal status, breast density, and use of hormone replacement therapy (HRT) as well as determination of performance characteristics for additional ipsilateral lesion detection. Results: Two hundred and eight patients enrolled in the study of which 87% (182/208) were analyzable. Of these analyzable patients, 26.4% (48/182), 7.1% (13/182), and 64.2% (120/182) were pre-, peri-, and postmenopausal, respectively, and 48.4% (88/182) had extremely or heterogeneously dense breast tissue, while 33.5% (61/182) had a history of HRT use. Ninety-two percent (167/182) underwent core biopsy for index lesion diagnosis. Invasive cancer was found in 77.5% (141/182), while ductal carcinoma in situ (DCIS) and/or Paget's disease were found in 22.5% (41/182) of patients. Both PEM and MRI had index lesion depiction sensitivity of 92.8% and both were significantly better than WBPET (67.9%, p < 0.001, McNemar's test). For index lesions, PEM and MRI had equivalent sensitivity of various tumors, categorized by tumor stage as well as similar invasive tumor size predictions with Spearman's correlation coefficient of 0.61 for both PEM and MRI compared to surgical pathology. Menopausal status, breast density, and HRT did not influence the sensitivity of PEM or MRI. For 67 additional unsuspected ipsilateral lesions or multifocal lesions, PEM had sensitivity of 85% (34/40) and specificity of 74%, (20/27) compared to MRI's sensitivity of 98% (39/40) and specificity of 48% (13/27) [p = 0.074, for sensitivity; p = 0.096 for specificity] Conclusion: PEM is a good alternative to MRI as a presurgical breast imaging option and its performance characteristics are not affected by patient menopausal/hormonal status or breast density. © 2010 The Author(s).