RSNA 2016

Abstract Archives of the RSNA, 2016


BR248-SD-TUB4

Bilateral Contrast-enhanced Spectral Mammography Compared to Breast Magnetic Resonance Imaging for Presurgical Imaging Evaluation In Newly Diagnosed Breast Cancer

Tuesday, Nov. 29 12:45PM - 1:15PM Room: BR Community, Learning Center Station #4



Stephanie A. Lee-Felker, MD, Los Angeles, CA (Presenter) Nothing to Disclose
Leena Tekchandani, MD, Mineola, NY (Abstract Co-Author) Nothing to Disclose
Mariam Thomas, MD, Sylmar, CA (Abstract Co-Author) Nothing to Disclose
Denise M. Andrews-Tang, MD, Sylmar, CA (Abstract Co-Author) Nothing to Disclose
Antoinette R. Roth, MD, Sylmar, CA (Abstract Co-Author) Nothing to Disclose
Esha A. Gupta, MD, Los Angeles, CA (Abstract Co-Author) Nothing to Disclose
James Sayre, PhD, Los Angeles, CA (Abstract Co-Author) Nothing to Disclose
Guita Rahbar, MD, Beverly Hills, CA (Abstract Co-Author) Nothing to Disclose
PURPOSE

To compare the diagnostic performance of contrast-enhanced spectral mammography (CESM) and breast magnetic resonance imaging (MRI) for presurgical imaging evaluation of extent of disease in newly diagnosed breast cancer, using histology as the gold standard.

METHOD AND MATERIALS

This IRB-approved, HIPAA-compliant, retrospective study included consecutive women with newly diagnosed unilateral breast cancer on core needle biopsy (CNB) who underwent presurgical CESM and MRI, followed by CNB of additional imaging-detected suspicious secondary cancer sites, at a single institution between April 2014 and October 2015.Exclusion criteria included: inflammatory breast cancer (n = 1) and contraindications for intravenous contrast administration (n = 3) or MRI (n = 2).Images were analyzed by 1 of 5 breast radiologists with 2-17 years of experience.Sensitivity (Sn), positive predictive value (PPV), and accuracy were calculated with 95% confidence intervals for both modalities.  Specificity (Sp) and negative predictive value (NPV) were also calculated for CESM but not for MRI because the latter did not have true negative lesions.  The number of false positive (FP) lesions for each modality was compared using the McNemar exact test.  A p value <0.05 was significant.  

RESULTS

52 women with 86 biopsy-proven breast lesions were included for analysis (mean age 50 years; range, 29-73 years).  The performance of CESM was: Sn 98.4% [90.3%-99.9%], Sp 82.6% [59.9%-94.0%], PPV 93.9% [84.4%-98.0%], NPV 95.0% [71.9%-99.7%], and accuracy 94.2% [86.8%-98.0%].  In comparison, the performance of MRI was: Sn 98.4% [90.3%-99.9%], PPV 72.9% [62.9%-82.5%], and accuracy 72.1% [62.2%-82.0%].  Of 8 additional biopsy-proven secondary cancer sites, CESM detected all 8 of 8 (100.0%) and MRI detected 7 of 8 (87.5%).CESM and MRI each had 1 false negative for DCIS in 2 different women.CESM had significantly fewer FPs than MRI (p <0.001): 4 FPs (2 high risk lesions, 1 fibroadenoma, and 1 non-high risk benign lesion) on CESM and 23 FPs (8 high risk lesions, 4 fibroadenomas, and 11 non-high risk benign lesions) on MRI.  

CONCLUSION

In this limited study, CESM was as sensitive and more specific than MRI for detecting breast cancer.

CLINICAL RELEVANCE/APPLICATION

In the appropriate clinical setting, CESM may substitute for breast MRI in presurgical imaging evaluation for extent of disease in newly diagnosed breast cancer.