Skip to main content

Table 2 Pre-scores of prediction models and their compositions of TET patients in the training cohort

From: Development and validation of a CT-texture analysis nomogram for preoperatively differentiating thymic epithelial tumor histologic subtypes

Short diameter (mean ± SD, mm) LTET (n = 61) HTET (n = 59) p-value
35.48 ± 16.44 29.85 ± 12.52 0.04
Boundary (%)
 Smooth 53 (86.89%) 30 (50.85%) 0.00
 Rough 8 (13.11%) 29 (49.15%)
Mediastinum fat line (%)
 clear 45 (73.77%) 20 (33.90%) 0.00
 unclear 16 (26.23%) 39 (66.10%)
MeanValue 68.63 (56.52, 78.55) a 48.84 (40.82, 58.18) a 0.00
SHAPE_Sphericity 0.95 (0.93, 0.97) a 0.93 (0.90, 0.96) a 0.01
NGLDM_Busyness 1.13 (0.36, 3.64) a 0.91 (0.43, 1.63) a 0.01
Pre-scores (Clinical model) −0.11 (−0.23, 0.06) a −0.01 (−0.11, 0.13) a 0.00
Pre-scores (CT model) − 0.37 (− 0.50, − 0.22) a 0.29 (− 0.33, 0.63) a 0.00
Pre-scores (TA model) −0.53 (− 0.85, 0.02) a 0.32 (− 0.01, 0.76) a 0.00
Pre-scores (Combined model) − 0.53 (− 0.99, − 0.05) a 0.52 (0.03, 0.84) a 0.00
  1. Note: NGLDM Neighboring Gray-Level Dependence Matrix, CT computed tomography, TA texture analysis; aValues refer to median (interquartile range (IQR))