Improving tractography in peri-lesional oedema using constrained spherical deconvolution in paediatric neurosurgery patients.
AI interpretation is pending for this paper.
Open original publication →What the AI sees
Not AI summarized yet.
Research significance
Pending deeper interpretation.
Source abstract
UNLABELLED: Tractography guides neurosurgical planning by balancing lesion resection with functional preservation. A common challenge to this approach is that some lesions have accompanying oedema, which can obscure white matter (WM) fibre orientation through increased isotropic diffusion resulting in diffusion-weighted signal properties resembling grey matter (GM) and cerebrospinal fluid (CSF). Constrained spherical deconvolution (CSD) can improve tract reconstruction in complex environments compared to diffusion tensor approaches. We assessed four CSD methods for improving tractography accuracy in peri-lesional oedema. We retrospectively analysed 14 paediatric neurosurgical patients (2.1–19 years of age) with brain tumours or epilepsy-associated lesions exhibiting peri-lesional oedema. Multi-shell diffusion-weighted imaging data acquired at 3T were analysed using two single-tissue models (1-tissue CSD and 1-tissue CSD with free-water correction) and two multi-tissue models (3-tissue [WM, GM, CSF] and 2-tissue [WM, CSF] MSMT-CSD). Probabilistic tractography was performed within a combined white matter and peri-lesional oedema mask. Streamline count, length, and tract volume were compared using linear mixed effects models. The volume of peri-lesional oedema ranged from 3,046 to 132,644 mm³ (average: 49,499 mm³). Applying free-water correction to DWI data did not increase the average streamline count (p = .04) but was associated with shorter streamlines and reduced tract volume (p ≤ .006) compared with standard 1-tissue CSD. For multi-tissue models, 2-tissue MSMT-CSD produced higher average streamline count and length, and proportion of tracts within oedematous regions compared to 3-tissue MSMT-CSD (p ≤ .008). Our findings demonstrate: (1) two-tissue MSMT-CSD reconstructs more and longer streamlines and larger tract volumes within peri-lesional oedema compared to three-tissue MSMT-CSD, and (2) free-water correction with single-tissue CSD provides no improvement in tract reconstruction. We determine that the two-tissue MSMT-CSD approach shows promise for improving surgical planning in cases with peri-lesional oedema. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1007/s00429-026-03108-6.