Table 2.
Variables associated with EBUS-TBNA diagnostic accuracy (bivariate analysis)
| Variable | Diagnostic success, N=533 | Diagnostic failure, N=127 | OR | 95% CI | P value | ||
| Nodal station 10L, n (%) | 32 | (6%) | 15 | (12%) | 0.477 | (0.250 to 0.910) | 0.025 |
| Nodal station 10R, n (%) | 73 | (14%) | 29 | (23%) | 0.536 | (0.331 to 0.869) | 0.011 |
| Nodal station 11R, n (%) | 10 | (2%) | 8 | (6%) | 0.284 | (0.110 to 0.736) | 0.010 |
| Nodal station 4R, n (%) | 146 | (27%) | 18 | (14%) | 2.285 | (1.340 to 3.896) | 0.002 |
| Size<1 cm, n (%) | 87 | (16%) | 33 | (26%) | 0.540 | (0.340 to 0.858) | 0.009 |
| Pathologist #6*, n (%) | 38 | (7%) | 14 | (11%) | 0.399 | (0.201 to 0.791) | 0.009 |
| Pathologist #7*, n (%) | 73 | (14%) | 21 | (17%) | 0.511 | (0.288 to 0.906) | 0.022 |
| Pathologist #9*, n (%) | 55 | (10%) | 19 | (15%) | 0.425 | (0.232 to 0.778) | 0.006 |
| Pathologist #15,* n (%) | 43 | (8%) | 24 | (19%) | 0.263 | (0.146 to 0.473) | <0.001 |
| Trainee bronchoscopist #12, n (%) | 154 | (29%) | 52 | (41%) | 0.589 | (0.395 to 0.879) | 0.010 |
| Senior bronchoscopist #2, n (%) | 269 | (50%) | 45 | (35%) | 1.857 | (1.243 to 2.774) | 0.003 |
| Senior bronchoscopist #4, n (%) | 25 | (5%) | 14 | (11%) | 0.397 | (0.200 to 0.788) | 0.008 |
| Lung cancer (NSCLC), n (%) | 194 | (36%) | 21 | (17%) | 2.889 | (1.752 to 4.764) | <0.001 |
| Other cancer (non-lung cancer), n (%) | 80 | (15%) | 3 | (2%) | 7.299 | (2.266 to 23.509) | 0.001 |
| Normal lymph node, n (%) | 141 | (26%) | 13 | (10%) | 3.154 | (1.722 to 5.777) | <0.001 |
| Granulomatous disease (non-sarcoidosis, non-TB), n (%) | 7 | (1%) | 7 | (6%) | 0.228 | (0.079 to 0.663) | 0.007 |
| Size mm, median (IQR) | 14 | (10–20) | 12 | (9–18) | 1.554 | (1.033 to 2.337) | 0.034 |
| Senior bronchoscopist experience (in number of EBUS), median (IQR) | 330 | (218–458) | 317 | (180–429) | 1.001 | (1.000 to 1.003) | 0.050 |
A normal lymph node was defined as a TBNA cytology negative for cancer or other conditions, but showing a lymphocyte fund, undisputed by the findings of surgical biopsy/resection, postmortem analysis or follow-up.
*The reference group was pathologists with a diagnostic accuracy >80%; bivariate analysis performed using binary logistic regression at a significance level of p<0.05 (two-tailed).
EBUS-TBNA, endobronchial ultrasound transbronchial needle aspiration; NLCLC, non-small cell lung cancer; TB, tuberculosis.