Abstract
Scoliosis monitoring in resource-limited settings remains constrained by limited access to radiography and concerns related to repeated radiation exposure. We present an open-source workflow for automated 3D spine reconstruction from tracked B-mode ultrasound, evaluated on a Senegalese cohort representing an underrepresented population in scoliosis imaging research. The workflow combines optical tracking, PLUS, OpenIGTLink, 3D Slicer, and automated vertebral segmentation using UltraSAM. We compare zero-shot and fine-tuned UltraSAM with nnU-Net and US-Medical Matting on held-out healthy and scoliotic ultrasound sequences. Fine-tuned UltraSAM achieved the best overall performance, with a DSC of $0.655$ and HD95 of $4.254$, and maintained improved performance in scoliotic subjects despite more complex anatomy. Qualitative reconstruction results showed better preservation of the central vertebral trajectory, supporting its use for downstream curvature assessment. This work provides a reproducible, radiation-free framework for spine assessment in low-resource clinical environments.
Links to Paper and Supplementary Materials
Main Paper (Open Access Version): https://papers.miccai.org/miccai-2026-sat/paper/MIRASOL_109.pdf
SharedIt Link: Not yet available
SpringerLink (DOI): Not yet available
Supplementary Material: Not Submitted
Link to Open Review
Open Review Page: https://openreview.net/forum?id=pH6HZyDGN7
BibTex
@InProceedings{ThiPap_An_MICCAISAT2026,
author = { Thiao, Pape Mady AND d’Albenzio, Gabriella AND Kim, Andrew AND Nguyen, Vivian AND Guo, Zhuoxin AND Seck, Idrissa AND Sunderland, Kyle AND Ba, Ousmane Seydina AND Camara, Mamadou Samba AND Diagne, Ngor Sine AND Ungi, Tamas AND Fichtinger, Gabor},
title = { { An Open-Source Workflow for Automated 3D Spine Reconstruction from Tracked Ultrasound: A Pilot Study in Senegal } },
booktitle = {Medical Image Computing and Computer Assisted Intervention -- MICCAI 2026 Workshops and Challenges},
year = {2026},
publisher = {Springer Nature Switzerland},
volume = {LNCS 17265},
month = {pending},
page = {pending}
}