TABLE 2.
Lymph Node Size and Morphologic and Pathologic Findings
| Ultrasound Feature | Finding on Final Pathologic Analysis | P | Odds Ratio (95% CI)a | |
|---|---|---|---|---|
|
| ||||
| Node-Negative Disease |
Node-Positive Disease |
|||
|
| ||||
| Short-axis diameter of lymph node (mm) (n = 501) | < 0.0001 | |||
| Median (range) | 6.0 (0–14.0) | 6.0 (0–24.0) | 0.83 (0.77–0.90) | |
| Mean (SD) | 5.6 (2.4) | 7.0 (3.1) | ||
| Long-axis diameter of lymph node (mm) (n = 501) | 0.016 | |||
| Median (range) | 12.0 (0–37.0) | 13.0 (0–46.0) | 0.96 (0.93–0.99) | |
| Mean (SD) | 12.7 (5.7) | 14.2(6.4) | ||
| Ratio of long-axis diameter to short-axis diameter (mm) (n = 501) | 0.045 | |||
| Median (range) | 2.25 (1.0–6.0) | 2.0 (1.0–16.0) | 1.22 (1.00–1.50) | |
| Mean (SD) | 2.38 (0.92) | 2.18 (1.09) | ||
| Node morphologic finding (n = 611)b | 0.004 | |||
| Node not seen | 39 (42.4) | 53 (57.6) | Reference | |
| Type I | 40 (44.0) | 51 (56.0) | 1.07 (0.59–1.91) | |
| Type II | 108 (43.9) | 138 (56.1) | 1.06 (0.66–1.73) | |
| Type III | 20 (40.8) | 29 (59.2) | 0.94 (0.46–1.90) | |
| Type IV | 9 (20.5) | 35 (79.5) | 0.35 (0.15–0.81) | |
| Type V | 11 (32.4) | 23 (67.6) | 0.65 (0.28–1.49) | |
| Type VI | 11 (20.0) | 44 (80.0) | 0.34 (0.16–0.74) | |
| Fatty hilum (n = 519)c | 0.0013d | |||
| Not visible | 11(19.0) | 47(81.0) | Reference | |
| Visible | 188 (40.8) | 273 (59.2) | 2.94 (1.49–5.82) | |
| Cortical thickness (n = 493)e | < 0.0001d | |||
| ≤ 3 mm | 152 (44.2) | 192 (55.8) | Reference | |
| > 3 mm | 38 (25.5) | 111 (74.5) | 0.43 (0.28–0.66) | |
a
For nodes with clinical nodal category N0.
b
Type I lymph nodes were hyperechoic with no visible cortex, type II nodes were thin (< 3 mm) with a hypoechoic cortex; type III nodes were hypoechoic with a cortex thicker than 3 mm, type IV nodes had a generalized lobulated hypoechoic cortex, type V nodes had focal hypoechoic cortical lobulation, and type VI nodes were totally hypoechoic with no hilum.
c
Limited to cases with lymph nodes visualized on axillary ultrasound.
d
Statistically significant.
e
Limited to cases with DICOM images and visible residual nodes for central review measurement.