Random forest model achieves AUC 0.979 for classifying pulmonary nodules on dual-phase SPECT/CT

Random forest model using dual-phase 99mTc-MIBI SPECT/CT semiquantitative parameters achieved AUC 0.979 (95% CI 0.951-0.995) for differentiating malignant from benign pulmonary nodules in a retrospective study of 132 patients.

• Retrospective study of 132 patients (30 benign, 102 malignant) who underwent dual-phase 99mTc-MIBI SPECT/CT; early-phase uptake and retention index (RI) differed significantly between groups.
• Multivariable analysis identified elevated CEA and RImax as independent predictors of malignancy; other models (SVM, LR, ANN) also showed high AUCs.
• All model AUCs were reported on the training set; no independent test set, external validation, or prospective evaluation was performed.

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