Archana Singh1, Shweta Kothari2, Utpalendu Das3, Mithun Chaudhuri4
BACKGROUND
To evaluate whether the detection of choline-containing compounds in single voxel MR spectroscopy of breast lesions can differentiate between benign and malignant breast lesions.
MATERIALS AND METHODS
This prospective observational study included 99 breast lesions in 77 patients (between 18 and 86 years age) who underwent 3T breast MRI including proton MR spectroscopy before biopsy. Following dynamic contrast-enhanced study, single voxel, water and fat suppressed proton MR spectroscopy was performed. The position and size of the Volume of Interest (VOI) was placed so as to include the enhancing part of the lesion and excluding as much as possible the nonenhancing gland parenchyma, adjoining fat or necrotic part of the lesion. Choline peak at 3.2 ppm was qualitatively evaluated. MRI BIRADS scoring was done for each lesion. Sensitivity and specificity of the (1) H-MRS were calculated. Final histopathological diagnosis was taken as the gold standard.
RESULTS
According to histopathology, 53 lesions were malignant and 46 were benign. The qualitative approach based on presence or absence of choline peak yielded 88.68% [95% CI 76.97% to 95.73%] sensitivity and 76.09% [95% CI 61.23 to 87.41%] specificity for differentiating malignant and benign lesions (p<0.0001).
CONCLUSION
In vivo proton, MR spectroscopy can be used as an adjunctive tool for characterising breast lesions. However, the detection of choline-containing compounds is not specific for malignancy. Benign breast lesions may also demonstrate choline peak.
