Table 3 Interpretation results of WARS1, TYMP and GBP1 antibodies by immunohistochemistry for predicting active ABMR by pathologists and deep learning.
Diagnosis | WARS1 positivity for active ABMR | TYMP positivity for active ABMR | GBP1 positivity for active ABMR | ||||||
|---|---|---|---|---|---|---|---|---|---|
Number of total cases (n = 53) | Number of cases considered positive, n (%) | Number of total cases (n = 52) | Number of cases considered positive, n (%) | Number of total cases (n = 52) | Number of cases considered positive, n (%) | ||||
Pathologists | Deep learning | Pathologists | Deep learning | Pathologists | Deep learning | ||||
Active ABMR | 16 | 12 (75) | 13 (81) | 16 | 14 (88) | 13 (81) | 15 | 7 (47) | 14 (93) |
Other diagnosis | 37 | 5 (14) | 2 (5) | 36 | 3 (8) | 4 (11) | 37 | 3 (8) | 5 (14) |
Including | |||||||||
Active ABMR, C4d positive | 5 | 5 (100) | 5 (100) | 4 | 3 (75) | 3 (75) | 4 | 1 (25) | 4 (100) |
Active ABMR, C4d negative | 11 | 7 (64) | 8 (73) | 12 | 11 (92) | 10 (83) | 11 | 6 (55) | 10 (91) |
Non-active cABMR | 5 | 2 (40) | 1 (20) | 5 | 2 (40) | 1 (20) | 5 | 1 (20) | 1 (20) |
Isolated C4d | 3 | 1 (33) | 0 (0) | 3 | 1 (33) | 1 (33) | 3 | 1 (33) | 0 (0) |
SG ABO incompatible | 3 | 0 (0) | 0 (0) | 3 | 0 (0) | 0 (0) | 3 | 0 (0) | 0 (0) |
Non-humoral TMA | 5 | 0 (0) | 0 (0) | 5 | 0 (0) | 0 (0) | 5 | 0 (0) | 0 (0) |
Acute TCMR | 6 | 1 (17) | 1 (17) | 6 | 0 (0) | 1 (17) | 6 | 0 (0) | 3 (50) |
Infection (PVN, APN) | 5 | 1 (20) | 0 (0) | 4 | 0 (0) | 1 (25) | 5 | 1 (20) | 1 (20) |
ATI | 5 | 0 (0) | 0 (0) | 5 | 0 (0) | 0 (0) | 5 | 0 (0) | 0 (0) |
Recurrent GN | 5 | 0 (0) | 0 (0) | 5 | 0 (0) | 0 (0) | 5 | 0 (0) | 0 (0) |