Reproduce “THA-Flow Generative Model: Prosthesis Geometry Prediction from Preoperative CT” on OpenResearch

Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.

Universal app for Apple silicon and Intel. Requires macOS 11 or later.

About this paper

Yiping Wang, Jie Li, Jingyu Shen, Liao Wang

Researchers from Changzhou Jinse Medical Information Technology Co., Ltd. and Shanghai Jiao Tong University School of Medicine developed THA-Flow, a conditional flow-matching model that generates three-dimensional prosthesis geometry directly from preoperative CT scans. This system produces a distribution of patient-matched, anatomically appropriate solutions for Total Hip Arthroplasty (THA), demonstrating high reconstruction quality for prosthesis representations and interactive generation times.

Key insights

  • Reframing THA planning as the conditional generation of a 3D prosthesis distribution better reflects the clinical reality of multiple acceptable solutions for a given patient anatomy.
  • Direct generation of explicit 3D prosthesis geometry, rather than 2D synthetic images, enables direct superimposition onto unaltered patient anatomy for accurate measurement, retrieval, and geometric analysis.
  • Independent registration of pelvic and femoral components in preoperative space, combined with a dual-channel prosthesis representation, provides robust supervision and a foundation for explicit leg-length correction calculations.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.