Table 4 Input data modality ablation study for skin lesion management prediction results \(\rm{MGMT}_{\rm{pred, multi}}\) obtained using a multi-task model.
From: Predicting the clinical management of skin lesions using deep learning
Experiment name | Input data | Metrics | ||||||
|---|---|---|---|---|---|---|---|---|
Clinical image | Dermoscopic image | Patient metadata | Sensitivity | Specificity | Precision | AUROC | Overall accuracy | |
\({\rm{C}}\) | ✓ | ✗ | ✗ | 0.5997 | 0.7911 | 0.5466 | 0.7781 | 0.6051 |
\({\rm{CM}}\) | ✓ | ✗ | ✓ | 0.6050 | 0.7983 | 0.5684 | 0.7852 | 0.6405 |
\({\rm{D}}\) | ✗ | ✓ | ✗ | 0.6935 | 0.8384 | 0.6265 | 0.8630 | 0.6962 |
\({\rm{DM}}\) | ✗ | ✓ | ✓ | 0.7126 | 0.8424 | 0.6622 | 0.8644 | 0.7215 |
\({\rm{CD}}\) | ✓ | ✓ | ✗ | 0.5830 | 0.8060 | 0.7393 | 0.8335 | 0.7342 |
\({\rm{CDM}}\) | ✓ | ✓ | ✓ | 0.6893 | 0.8299 | 0.7134 | 0.8505 | 0.7367 |
