Application of the diffusion kurtosis model for the study of breast lesions

• Autor(a) principal: Nogueira, Luísa
• Data de Publicação: 2014
• Outros Autores: Brandão, Sofia, Matos, Eduarda, Nunes, Rita Gouveia, Loureiro, Joana, Ramos, Isabel, Ferreira, Hugo Alexandre
• Tipo de documento: Artigo
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/14206
• topic: Breast Neoplasms; Diagnosis, Differential; Diffusion Magnetic Resonance Imaging; Female; Follow-Up Studies; Image Interpretation, Computer-Assisted; Middle Aged; Prospective Studies; Reproducibility of Results; Algorithms

Objectives To evaluate diffusion-weighted imaging (DWI) and diffusion kurtosis imaging (DKI) in the differentiation and characterisation of breast lesions. Methods Thirty-six women underwent breast magnetic resonance imaging (MRI) including a DWI sequence with multiple b-values (50–3,000 s/mm2). Mean values for apparent diffusion coefficient (ADC), mean diffusivity (MD) and mean kurtosis (MK) were calculated by lesion type and histological subtype. Differences and correlation between parameters were determined. Results Forty-four lesions were found. There were significant differences between benign and malignant lesions for all parameters (ADC, p = 0.017; MD, p = 0.028; MK, p = 0.017). ADC and MD were higher for benign (1.96 ± 0.41 × 10−3 and 2.17 ± 0.42 × 10−3 mm2/s, respectively) than for malignant lesions (1.33 ± 0.18 × 10−3 and 1.52 ± 0.50 × 10−3 mm2/s). MK was higher for malignant (0.61 ± 0.27) than benign lesions (0.37 ± 0.18). We found differences between invasive ductal carcinoma (IDC) and fibroadenoma (FA) for all parameters (ADC, MD and MK): p = 0.016, 0.022 and 0.016, respectively. FA and fibrocystic change (FC) showed differences only in MK (p = 0.016). Conclusions Diffusion in breast lesions follows a non-Gaussian distribution. MK enables differentiation and characterisation of breast lesions, providing new insights into microstructural complexity. To confirm these results, further investigation in a broader sample should be performed.