Quantitative and reproducibility study of four tractography algorithms used in clinical routine

Abstract

Purpose: To evaluate fiber tracking strategy in terms of acquisition schemes in conjunction with four algorithms used in clinical routine, we studied one of the major tracts, anatomically well known, and which should be preserved as much as possible during neurosurgery: the corticospinal tract. Materials and Methods:Two identical exams, composed of three DTI acquisition schemes (6, 15, and 32 gradient directions), were performed on 12 healthy subjects during two different sessions. For each subject, intra-operator, and inter-exam reproducibility was quantitatively calculated from different fiber tracking algorithms: three deterministic and a probabilistic one. Inter-exam reproducibility was evaluated comparing fiber tracking results from the repetition of the same acquisition one month apart and variation of the fiber density distribution percentile.Results: For each fiber tracking algorithm, the best reproducibility result is obtained in case of 50% of fiber density and for the number of directions equal to 32. The reproducibility is improved using the probabilistic algorithm. Conclusion:This study highlights increased reliability of reproducibility results based on the number of directions used during the acquisition. The method of tractography used and the choice of adequate density fiber tract greatly improve the results. DIFFUSION TENSOR IMAGING (DTI) and tractography have the potential to depict fiber architecture within the white matter. DTI has emerged as a noninvasive imaging modality able to provide in vivo information about the white matter structure (1). The method is based on the sensitivity of measured diffusivity of the water protons to the microstructural environment. The diffusion paths of water molecules are longer along the axis of the myelinated fibers than perpendicular according to the degree of myelination, axonal membrane and subvoxel coherence (2,3). White matter tractography is a promising application of diffusion weighted imaging (DWI) (4-8). It uses the directional information of diffusion tensor maps to estimate connection pathways in white matter. Tractography studies of major white matter tracts in the human brain appear to be in good agreement with the anatomical results obtained with dissection or histological methods (5,6,(8)(9)(10). The knowledge of the architectonical organization and connectivity may contribute to a better understanding of brain anatomy, both in physiological and pathological conditions. It may be useful to assess how brain organization is affected by disorders such as trauma or tumor growth. Therefore, it is very important to determine the relationship between the lesions and the eloquent white matter tracts so that appropriate neurosurgical approaches can be designed to avoid destroying important fiber tracts. However, the question remains whether the DTI fiber tractography results are reproducible and reliable enough to be used as a tool for clinical routine. The reliability and reproducibility of these techniques are known to be limited by the quality of ...

Deterministic Diffusion Fiber Tracking Improved by Quantitative Anisotropy

“…Nonetheless, this improvement has not yet been rigorously examined. Although several tractography evaluation studies have been conducted [29–31], they examined the overall tracking results, which can also be affected by the applied reconstruction methods and tractography algorithms. A study specifically comparing voxel-based and ODF-based index is yet to be conducted.…”

Hemispheric Asymmetry in White Matter Connectivity of the Temporoparietal Junction with the Insula and Prefrontal Cortex

“…A previous probabilistic tractography study reported improved sensitivity comparing 12 to 60 directions data within the same subjects to examine major fibre pathways [76]. Tensaouti et al [77] also reported a significant increase in corticospinal tract volume between 6 and 32 diffusion-encoding direction data, but only a small volume increase between 15 and 32 direction data. In our study, 25 directions seemed to be sufficient for defining tracts between most of the seeds and targets.…”

A hitchhiker's guide to diffusion tensor imaging

“…Fiber tracking can be performed with different algorithms divided in two main categories: deterministic and probabilistic (Jones, 2008, 2010a; Descoteaux et al, 2009; Chung et al, 2011; Fillard et al, 2011; Tensaouti et al, 2011; Tournier et al, 2011). Deterministic tractography aims to model the data and, in practical terms, can be thought of as generating/reconstructing one fiber from each seed.…”