Musculoskeletal (MSK)AI / InformaticsResearch

Combined MRI radiomics and clinical model predicts spine/pelvis chordoma recurrence with test AUC 0.897

European spine journal : official publication of the European Spine Society, the European Spinal Deformity Society, and the European Section of the Cervical Spine Research Societyyesterday

In 135 patients with spine/pelvis chordoma, an MRI radiomics model predicted recurrence with test AUC 0.847; adding age, tumor size, and resection mode lifted test AUC to 0.897 vs 0.700 for clinical factors alone.

  • Random Forest radiomics model using T2-weighted and contrast-enhanced T1 MRI achieved training/test AUCs of 0.929/0.847, outperforming clinical factors (0.680/0.700).
  • Combined radiomics-clinical nomogram reached test AUC 0.897 and highest net clinical benefit on decision curve analysis.
  • Design: single-center retrospective study of 135 patients; age, tumor size, and resection mode were independent predictors.
Read the source

Automated summary

RadPigeon summaries are original and for information only. They are not clinical advice.