Quantitative liver ADC measurements using diffusion-weighted MRI at 3 Tesla: evaluation of reproducibility and perfusion dependence using different techniques for respiratory compensation

Diffusion weighted imaging (DWI) of the liver suffers from low signal to noise making 3 Tesla (3 T) an attractive option, but 3 T data is scarce. It was the aim to study the influence of different b values and respiratory compensation methods (RCM) on the apparent diffusion coefficient (ADC) level and on ADC reproducibility at 3 T. Ten healthy volunteers and 12 patients with malignant liver lesions underwent repeated (2-22 days) breathhold, free-breathing and respiratory triggered DWI at 3 T using b values between 0 and 1,000 s/mm(2). The ADCs changed up to 150 % in healthy livers and up to 48 % in malignant lesions depending on b value combinations. Best ADC reproducibility in healthy livers were obtained with respiratory triggering (95 % limits of agreement: +/- 0.12) and free-breathing (+/- 0.14). In malignant lesions equivalent reproducibility was obtained with less RCM dependence. The use of a lower maximum b value (b = 500) decreased reproducibility (+/- 0.14 to +/- 0.32) in both normal liver and malignant lesions. Large differences in absolute ADC values and reproducibility caused by varying combinations of clinically realistic b values were demonstrated. Different RCMs caused smaller differences. Lowering maximum b value to 500 increased limits of agreement up to a factor of two. Serial ADC changes larger than approximately 15 % can be detected confidently on an individual basis in both malignant lesions and normal liver parenchyma at 3 T using appropriate b values and respiratory compensation.