Table 3 The performance of the clinical model for classifying three tumors
Classification models | Tumor Type | Training set | Validation set | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
Accuracy | Sensitivity | Specificity | AUC | Accuracy | Sensitivity | Specificity | AUC | ||||
Binary classification model | Ganglioneuroma vs. non-Ganglioneuroma | 75.3% | 90.0% | 73.1% | 0.901 | 73.1% | 75% | 72.7% | 0.682 | ||
Ganglioneuroblastoma vs. Neuroblastoma | 73.1% | 55.6% | 79.6% | 0.760 | 59.1% | 66.7% | 56.3% | 0.708 | |||
Ganglioneuroma | Ganglioneuroblastoma | Neuroblastoma | Overall accuracy | Ganglioneuroma | Ganglioneuroblastoma | Neuroblastoma | Overall accuracy | ||||
Three classification model | Ganglioneuroma vs. Ganglioneuroblastoma vs. Neuroblastoma | 90.0% | 11.1% | 79.6% | 64.9% | 75.0% | 0% | 56.3% | 46.2% | ||