Three-dimensional visualization of brain tumor progression based accurate segmentation via comparative holographic projection

PLoS One. 2020 Jul 30;15(7):e0236835. doi: 10.1371/journal.pone.0236835. eCollection 2020.

Abstract

We propose a new optical method based on comparative holographic projection for visual comparison between two abnormal follow-up magnetic resonance (MR) exams of glioblastoma patients to effectively visualize and assess tumor progression. First, the brain tissue and tumor areas are segmented from the MR exams using the fast marching method (FMM). The FMM approach is implemented on a computed pixel weight matrix based on an automated selection of a set of initialized target points. Thereafter, the associated phase holograms are calculated for the segmented structures based on an adaptive iterative Fourier transform algorithm (AIFTA). Within this approach, a spatial multiplexing is applied to reduce the speckle noise. Furthermore, hologram modulation is performed to represent two different reconstruction schemes. In both schemes, all calculated holograms are superimposed into a single two-dimensional (2D) hologram which is then displayed on a reflective phase-only spatial light modulator (SLM) for optical reconstruction. The optical reconstruction of the first scheme displays a 3D map of the tumor allowing to visualize the volume of the tumor after treatment and at the progression. Whereas, the second scheme displays the follow-up exams in a side-by-side mode highlighting tumor areas, so the assessment of each case can be fast achieved. The proposed system can be used as a valuable tool for interpretation and assessment of the tumor progression with respect to the treatment method providing an improvement in diagnosis and treatment planning.

MeSH terms

  • Algorithms*
  • Brain / anatomy & histology*
  • Brain / diagnostic imaging
  • Brain Neoplasms / diagnostic imaging
  • Brain Neoplasms / pathology*
  • Disease Progression
  • Holography / methods*
  • Humans
  • Imaging, Three-Dimensional / methods*
  • Magnetic Resonance Imaging / methods*
  • Signal Processing, Computer-Assisted

Grants and funding

The author(s) received no specific funding for this work.