Role of Magnetic Resonance Spectroscopic Imaging in Recategorization of BIRADS 4 Breast Lesions

Abstract
Background: Breast cancer is one of the major causes of death among females, its early diagnosis allow better prognosis and prolonged survival rates of the patients. Post contrast dynamic MRI is the most sensitive method for breast cancer detection which has also good diagnostic performance for evaluation of different breast pathological lesions. Magnetic Resonance (MR) spectroscopy is a functional MRI technique which is used to study the metabolic profile of different breast lesions to improve the diagnostic accuracy of MRI especially in BIRADS 4 breast lesions in order to reduce the unnecessary biopsies.
Aim of Study: Evaluation of the role of multi-voxel MR spectroscopy in the recategorization of BIRADS 4 breast lesions when combined with Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI).
Material and Methods: This study included 63 patients, their ages ranged from 28 to 73 years old with mean age 46.7 years old. All the patients included in this study were subjected to history, clinical examination, MRI examination and his-topathology. Initially the included breast lesions were catego-rized as BIRADS 4 by conventional post contrast dynamic MRI techniques, and then we added the results of the magnetic resonance spectroscopic imaging to detect whether the spec-troscopic data can improve the diagnostic accuracy of the MRI study. Finally the reported diagnosis was confirmed by histopathological specimens obtained after excision biopsy and core needle biopsy. We used 1.5 T MRI system and the sequences performed included T1 & T2 pulse sequences, STIR and dynamic contrast enhanced MRI followed by MRS for the suspected lesions. For statistical analysis we used T-test and ROC curve.
Results: This study included 63 female patients, their ages ranged from 28 to 73 years old with mean age 46.7 years old. All the breast lesion were categorized as BI-RADS 4 breast lesions based on the Dynamic Contrast-Enhanced Magnetic Resonance Imaging (DCE-MRI) then after magnetic resonance spectroscopy the breast lesions were re-categorized into BIRADS 3, 4 and 5 categories. Finally the results were compared to the histopathological results. The most common clinical presentation is palpable mass. We had 33 pathologically proven benign lesions and 30 pathologically proven malignant lesions. The diagnosis was downgraded into BIRADS 3 in 29 cases and upgraded into BIRADS 5 in 28 cases and in 6 cases there was no change in the BIRADS 4 categorization after MRS. During MRS examination the choline resonance peak was detected at a frequency of 3.23ppm, and the choline SNR was measured at cut off value of 2. There were statistically significant differences between malignant and benign breast lesions, the choline SNR measured 6.31±0.99 (mean ± SD) in the malignant lesions and it measured 1.902±0.47 (mean ± SD) in the benign lesions. The sensitivity, specificity, PPV, NPV and accuracy measured 81.3%, 62.6%, 76.5%, 67.1% and 73.4% for the DCE-MRI and measured 87.9%, 89.1%, 85.2%, 88.9% and 94.8% for combined assessment by MRS and DCE-MRI.
Conclusion: Magnetic resonance spectroscopic imaging can be problem solving method for BIRADS 4 breast lesions when combined with dynamic contrast enhanced MRI to improve its diagnostic accuracy.