Table 2.
Evaluation of the diagnosis segmentation model using automatic extraction
| Textual elements | First round with the validation set (n = 496) | Second round with the validation set (n = 488) | Test set (n = 532) | |||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| TP | FP | FN | P (%) | R (%) | F1 (%) | TP | FP | FN | P (%) | R (%) | F1 (%) | TP | FP | FN | P (%) | R (%) | F1 (%) | |
| Tumor recurrence | 134 | 0 | 0 | 100.0 | 100.0 | 100.0 | 134 | 0 | 0 | 100.0 | 100.0 | 100.0 | 98 | 0 | 0 | 100.0 | 100.0 | 100.0 |
| Tumor location | 477 | 0 | 4 | 100.0 | 99.2 | 99.6 | 476 | 0 | 0 | 100.0 | 100.0 | 100.0 | 531 | 0 | 1 | 100.0 | 100.0 | 100.0 |
| Invasiveness | 166 | 0 | 1 | 100.0 | 99.4 | 99.7 | 165 | 0 | 0 | 100.0 | 100.0 | 100.0 | 128 | 0 | 0 | 100.0 | 100.0 | 100.0 |
| Endocrine status | 410 | 0 | 32 | 100.0 | 92.8 | 96.3 | 437 | 0 | 0 | 100.0 | 100.0 | 100.0 | 464 | 1 | 5 | 99.8 | 98.9 | 99.3 |
| Tumor size | 341 | 0 | 0 | 100.0 | 100.0 | 100.0 | 338 | 0 | 0 | 100.0 | 100.0 | 100.0 | 394 | 0 | 0 | 100.0 | 100.0 | 100.0 |
| Histopathology | 477 | 0 | 19 | 100.0 | 96.2 | 98.1 | 488 | 0 | 0 | 100.0 | 100.0 | 100.0 | 529 | 0 | 3 | 100.0 | 99.4 | 99.7 |
| Knosp grading | 44 | 0 | 3 | 100.0 | 93.6 | 96.7 | 46 | 0 | 0 | 100.0 | 100.0 | 100.0 | 2 | 0 | 0 | 100.0 | 100.0 | 100.0 |
| Residual tumor | 4 | 0 | 0 | 100.0 | 100.0 | 100.0 | 4 | 0 | 0 | 100.0 | 100.0 | 100.0 | 8 | 0 | 0 | 100.0 | 100.0 | 100.0 |
| Diagnostic confirmation | 76 | 0 | 3 | 100.0 | 96.2 | 98.1 | 78 | 0 | 0 | 100.0 | 100.0 | 100.0 | 18 | 0 | 1 | 100.0 | 94.7 | 97.3 |
| Refractoriness* | 2 | 0 | 0 | 100.0 | 100.0 | 100.0 | 2 | 0 | 0 | 100.0 | 100.0 | 100.0 | - | - | - | - | - | - |
*No PA diagnostic statement contained textual element of refractoriness in the test set