Purpose: Nipple discharge is the third most common breast-related complaint. It is recommended to exclude malignancy in pathologic nipple discharge (PND). Mammography and ultrasound are the first-line conventional diagnostic (CD) imaging. Although magnetic resonance is often used as a complementary modality, molecular breast imaging (MBI) with Tc-99m sestamibi may be a suitable alternative. Considering the lack of information on this subject and its clinical importance, this study aimed to evaluate the role of MBI in ruling out malignancy in patients with PND and negative/indeterminate CD. Patients and Methods: Retrospective cohort single-center study including all patients with PND evaluated by CD and MBI between 2012 and 2020. Pathology was considered the gold standard. Follow-up was used when pathology was not available. Results: Of the 96 cases of PND included, 78 were benign, and 18 (20%) corresponded to breast cancer (BC). Although CD and MBI were concordant in the BIRADS classification in 81% (78/96), half of BC were detected by MBI only. BC was located directly behind the nipple in a minority of patients (11%), meaning that MBI could significantly prevent futile central ductal excision. MBI presented higher sensitivity (83% vs. 33%) and negative predictive value (96% vs. 86%) than CD alone, with similar specificity (89% vs. 92%) and positive predictive value (63% vs. 50%). The area under the curve of MBI and CD was 0,86 (P-value<0.001 [95% CI: 0.75-0.97]) and 0.63 (P-value=0.091 [95% CI: 0.47-0.79]), respectively. Conclusions: MBI showed good diagnostic accuracy for detecting BC in patients with PND with negative/indeterminate findings on CD imaging.

Purpose
This study evaluates the semi-quantitative single-photon emission computed tomography (SPECT) parameters of prone SPECT using [99mTc]Tc-sestamibi and compares them with Molecular Breast Imaging (MBI)-derived semi-quantitative parameters for the potential use of response prediction in women with locally advanced breast cancer (LABC).

Procedures
Patients with proven LABC with a tumor ≥ 2 cm on mammography and an indication for MBI using [99mTc]Tc-sestamibi were prospectively enrolled. All patients underwent a prone SPECT/CT at 5 min (early exam) and an additional scan at 90 min (delayed exam) after injection of 600 MBq [99mTc]Tc-sestamibi to compose wash-out rates (WOR). All patients underwent MBI after early SPECT/CT. Volumes of interest of the primary tumor were drawn semi-automatically on early and delayed SPECT images. Semi-quantitative analysis included maximum and mean standardized uptake values (SUVmax, SUVmean,), functional tumor volume (FTVSPECT), total lesion mitochondrial uptake (TLMU), tumor-to-background ratios (TBRmax and TBRmean), WOR and coefficient of variation (COVSPECT). Subsequently, the FTVSPECT, TBRSPECT and COVSPECT were compared to FTVMBI, TBRMBI and COVMBI.

Results
Eighteen patients were included. Early SUVmax, and TBRmax showed significantly higher interquartile range (IQR) compared to SUVmean and TBRmean, respectively 2.22 (2.33) g/mL, 6.86 (8.69), 1.29 (1.39) g/mL and 3.99 (5.07) (median (IQR), p < 0.05). WOR showed a large IQR (62.28), indicating that there is WOR variation among the LABC patients. FTV showed no difference between MBI and early SPECT semi-quantitative parameter (p = 0.46).

Conclusions
In LABC patients it is feasible to obtain semi-quantitative parameters from prone SPECT/CT. The FTV derived from early prone SPECT/CT is comparable with MBI-based FTV. Studies with comprehensive clinical parameters are needed to establish the clinical relevance of these semi-quantitative parameters, including WOR, for response prediction before its use in clinical routine.