Frontiers in Oncology (Mar 2023)

Automatic bundle-specific white matter fiber tracking tool using diffusion tensor imaging data: A pilot trial in the application of language-related glioma resection

  • Yifan Yuan,
  • Yifan Yuan,
  • Yifan Yuan,
  • Yifan Yuan,
  • Yifan Yuan,
  • Yifan Yuan,
  • Tianming Qiu,
  • Tianming Qiu,
  • Tianming Qiu,
  • Tianming Qiu,
  • Tianming Qiu,
  • Tianming Qiu,
  • Shin Tai Chong,
  • Sanford Pin-Chuan Hsu,
  • Ying-Hua Chu,
  • Yi-Cheng Hsu,
  • Geng Xu,
  • Geng Xu,
  • Geng Xu,
  • Geng Xu,
  • Geng Xu,
  • Geng Xu,
  • Yu-Ting Ko,
  • Kuan-Tsen Kuo,
  • Zixiao Yang,
  • Zixiao Yang,
  • Zixiao Yang,
  • Zixiao Yang,
  • Zixiao Yang,
  • Zixiao Yang,
  • Wei Zhu,
  • Wei Zhu,
  • Wei Zhu,
  • Wei Zhu,
  • Wei Zhu,
  • Wei Zhu,
  • Ching-Po Lin,
  • Jianping Song,
  • Jianping Song,
  • Jianping Song,
  • Jianping Song,
  • Jianping Song,
  • Jianping Song,
  • Jianping Song

DOI
https://doi.org/10.3389/fonc.2023.1089923
Journal volume & issue
Vol. 13

Abstract

Read online

Cerebral neoplasms like gliomas may cause intracranial pressure increasing, neural tract deviation, infiltration, or destruction in peritumoral areas, leading to neuro-functional deficits. Novel tracking technology, such as DTI, can objectively reveal and visualize three-dimensional white matter trajectories; in combination with intraoperative navigation, it can help achieve maximum resection whilst minimizing neurological deficit. Since the reconstruction of DTI raw data largely relies on the technical engineering and anatomical experience of the operator; it is time-consuming and prone to operator-induced bias. Here, we develop new user-friendly software to automatically segment and reconstruct functionally active areas to facilitate precise surgery. In this pilot trial, we used an in-house developed software (DiffusionGo) specially designed for neurosurgeons, which integrated a reliable diffusion-weighted image (DWI) preprocessing pipeline that embedded several functionalities from software packages of FSL, MRtrix3, and ANTs. The preprocessing pipeline is as follows: 1. DWI denoising, 2. Gibbs-ringing removing, 3. Susceptibility distortion correction (process if opposite polarity data were acquired), 4. Eddy current and motion correction, and 5. Bias correction. Then, this fully automatic multiple assigned criteria algorithms for fiber tracking were used to achieve easy modeling and assist precision surgery. We demonstrated the application with three language-related cases in three different centers, including a left frontal, a left temporal, and a left frontal-temporal glioma, to achieve a favorable surgical outcome with language function preservation or recovery. The DTI tracking result using DiffusionGo showed robust consistency with direct cortical stimulation (DCS) finding. We believe that this fully automatic processing pipeline provides the neurosurgeon with a solution that may reduce time costs and operating errors and improve care quality and surgical procedure quality across different neurosurgical centers.

Keywords