OBJECTIVES:To assess the reliability of diffusion tensor imaging (DTI)-based fibre tractography (FT), which is a prerequisite for clinical applications of this technique. Here we assess the test-retest reproducibility of the architectural and microstructural features of two clinically relevant tracts reconstructed with DTI-FT. METHODS:The corticospinal tract (CST), arcuate fasciculus (AF) and its long segment (AFl) were reconstructed in 17 healthy subjects imaged twice using a deterministic approach. Coefficients of variation (CVs) of diffusion-derived tract values were used to assess the microstructural reproducibility. Spatial correlation and fibre overlap were used to assess the architectural reproducibility. RESULTS:Spatial correlation was 68 % for the CST and AF, and 69 % for the AFl. Overlap was 69 % for the CST, 61 % for the AF, and 59 % for the AFl. This was comparable to 2-mm tract shift variability. CVs of diffusion-derived tract values were at most 3.4 %. CONCLUSIONS:The results showed low architectural and microstructural variability for the reconstruction of the tracts. The architectural reproducibility results encourage the further investigation of the use of DTI-FT for neurosurgical planning. The high microstructural reproducibility results are promising for using DTI-FT in neurology to assess or predict functional recovery.