Diffusion-weighted magnetic resonance imaging of breast lesions: the influence of different fat-suppression techniques on quantitative measurements and their reproducibility.

OBJECTIVE

The aim of this study was to evaluate the influence of different fat-suppression techniques on quantitative measurements and their reproducibility when applied to diffusion-weighted imaging (DWI) of breast lesions.

METHODS

Twenty-five patients with different types of breast lesions were examined on a clinical 1.5-T magnetic resonance imaging (MRI) system. Two diffusion-weighted sequences with different fat-suppression methods were applied: one with spectral presaturation by inversion recovery (SPIR), and one with short-TI inversion recovery (STIR). The acquisition of both sequence variants was repeated with modified shim volume. Lesion-to-background contrast (LBC), apparent diffusion coefficients (ADC) ADC(0,1000) and ADC(50,1000), and their coefficients of variation (CV) were determined.

RESULTS

In four patients, the image quality of DWI with SPIR was insufficient. In the other 21 patients, 46 regions of interest (ROI), including 11 malignant and 35 benign lesions, were analysed. The LBC, ADC(0,1000) and ADC(50,1000) values, which did not differ between initial and repeated measurements, were significantly higher for STIR than for SPIR. The mean CV improved from 10.8 % to 4.0 % (P = 0.0047) for LBC, from 6.3 % to 2.9 % (P = 0.0041) for ADC(0,1000), and from 6.3 % to 2.6 % (P = 0.0049) for ADC(50,1000).

CONCLUSION

For STIR compared to SPIR fat suppression, improved lesion conspicuity, higher ADC values, and better measurement reproducibility were found in breast DWI.

KEY POINTS

• Quality of fat suppression influences quantitative DWI breast lesion measurements. • In breast DWI, STIR fat suppression worked more reliably than SPIR. • Lesion-to-background contrast and its reproducibility were significantly higher with STIR fat suppression. • Lesional ADCs and their reproducibility were significantly higher with STIR fat suppression.