Diffusion-weighted imaging with fat suppression using short-tau inversion recovery: Clinical utility for diagnosis of...

To compare the utility of conventional diffusion-weighted imaging (DWI) with fat suppression using short-tau inversion recovery (STIR-DWI) for the detection of breast lesions.

Magnetic resonance imaging (MRI) images of 56 patients (both DWI and STIR-DWI performed) were retrospectively analysed. Parameters compared between DWI and STIR-DWI were image artefacts, image signal-to-noise ratio (SNR), apparent diffusion coefficient (ADC), and contrast-to-noise ratio (CNR). Diagnostic utility was assessed using receiver operating characteristic (ROC) analysis.

No abnormality was detected in 17 patients, with lesions observed in 39 patients (16 benign, 23 malignant; confirmed by biopsy or surgical histopathology). The rate of image artefacts was significantly lower for STIR-DWI (p < 0.01): quality levels 1 (best), 2, and 3 accounted for 50%, 35.7%, and 14.3% of DWI images, and 96.4%, 3.6% and 0% of STIR-DWI images, respectively. The SNR was not significantly different. ADC values of breast lesions and normal glands were significantly lower for DWI than for STIR-DWI (p = 0.03 and 0.034). ADC values of malignant lesions, but not benign lesions, were significantly lower for DWI than for STIR-DWI (p = 0.02). CNR values of both benign and malignant lesions were not significantly different between DWI and STIR-DWI. The area under the ROC curve, for the use of ADC values to differentiate benign from malignant lesions, was not significantly different between DWI (0.931) and STIR-DWI (0.914). Taking a threshold ADC value of 1.23 × 10(-3) mm(2)/s, the sensitivity and specificity were 87.5% and 87% for DWI, and 87.5% and 82.6% for STIR-DWI, respectively.

STIR-DWI is adequate for clinical use in breast MRI investigations.