Table 2. Validation of 272 ex vivo needle biopsies (100 patients).
| Non-Malignant | Malignant | |||||
|---|---|---|---|---|---|---|
| Benign$ | Oncocytoma | Normal | Clear Cell | Papillary | Chromophobe | |
| Based on Biopsy (N = 272) | ||||||
| Ex Vivo Biopsy (N) | 26 | 26 | 102 | 98 | 14 | 6 |
| Correctly Predicted Subtype (N, %) | 11 (73%)* | 15 (58%) | 100 (98%) | 89 (91%) | 9 (64%) | 6 (100%) |
| Correctly Predicted Non-Malignant or Malignant (N, %) |
12 (80%)* | 26 (100%) | 100 (98%) | 89 (91%) | 12 (86%) | 6 (100%) |
| Based on Tumors (N = 100) | ||||||
| Tumors (N) | 13 | 16 | - | 59 | 8 | 3 |
| Correctly Predicted Subtype (N, %)1 | 6 (75%)* | 7 (44%) | - | 53 (90%) | 5 (63%) | 3 (100%) |
| Correctly Predicted Non-Malignant or Malignant (N, %) |
6 (75%)* | 16 (100%) | - | 54 (92%) | 7 (88%) | 3 (100%) |
$ - consists of angiomyolipoma and other uncommon non-malignant lesions (i.e. capillary hemangioma, renal tubular hyperplasia, etc.)
1 - patient assigned subtype of biopsy with maximum posterior probability
* - prediction only of angiomyolipoma (N = 15 ex vivo samples, N = 8 tumors)