SPECT/CT parameters improve ML differentiation of spinal TB from pyogenic infection

Integrating quantitative SPECT/CT metrics into a machine-learning model improved differentiation of spinal tuberculosis from pyogenic spondylitis, adding 0.110 to AUC (95% CI 0.054–0.172) over conventional data.

• In the RBF-SVM analysis, the fusion model achieved an AUC of 0.957, but the incremental gain over baseline was not significant (ΔAUC 0.021; 95% CI -0.003 to 0.050).
• Retrospective, single-center study with internal validation only; clinical utility requires multicenter prospective confirmation.

Automated summary

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