Skip to content

v0.2.0: Automatic level detection + Multi-level analysis + automatic dural sac

Latest
Latest

Choose a tag to compare

View all tags
@neuromath neuromath released this 25 Jun 17:12
v0.2.0
794c82d

[0.2.0] - 2026-06-25

Added

  • Automatic lumbar level detection via TotalSpineSeg: labels intervertebral discs (L1-L2 … L5-S) and the L3 vertebral body, then maps them onto the axial series with gap detection for levels not covered by the axial volume.
  • Automatic dural sac segmentation: cross-sectional area is computed by resampling the TotalSpineSeg canal mask onto each axial slice in world space, replacing the previous manual polygon ROI.
  • Multi-level analysis ("Analyze All Levels"): paraspinal muscle metrics + dural sac CSA at every detected level in a single run. Levels falling in axial gaps are analyzed from the nearest slice and flagged with the offset distance.
  • Canal narrowing flags: neutral comparison of dural sac CSA against published thresholds (absolute <75, relative <100, early <130 mm²), with a caveat at L5-S where the thresholds are less reliable. Reported as a research aid for the radiologist, not a diagnosis.
  • L3-based sarcopenia in the multi-level workflow, with a nearest-slice fallback when the L3 vertebral body falls in an axial gap. PMI is reported only at L3; other levels report muscle CSA / fat fraction.
  • Multi-level PDF and Excel reports: per-level table, muscle detail, sarcopenia summary, stenosis flags, and disclaimers. The existing report buttons produce the multi-level report when a multi-level analysis has been run.
  • GUI: "Detect Levels" panel with a clickable level list (jumps to the corresponding axial slice), sagittal level overlay with labels, axial canal overlay, and "Analyze All Levels" action.
  • TotalSpineSeg MPS (Apple Silicon GPU) support patch under patches/.

Changed

  • Dural sac CSA is now automatic; the manual ROI panel ("Draw ROI" / "Clear") has been removed.
  • Analyze now also overlays the segmented dural sac and reports its CSA automatically.
  • Sarcopenia notes and batch messages translated to English for consistency.

Fixed

  • Axial canal overlay was rendered at the wrong location due to (1) stale TotalSpineSeg output files from prior runs and (2) a basename match that incorrectly paired sagittal_s2 with sagittal_s201. The output directory is now cleared before each run and the canal NIfTI is matched by exact stem.
  • Sagittal level overlay vertical positioning corrected (consistent cranial→caudal mapping in both directions).

Notes

  • Research use only. Stenosis thresholds are from the literature and have not been validated against a local cohort. Supine MRI may underestimate canal narrowing relative to axial-loaded/standing imaging.
Assets 2
Loading