Skip to main content
. Author manuscript; available in PMC: 2023 Apr 16.
Published before final editing as: Eur J Cancer. 2021 Oct 16;158:111–122. doi: 10.1016/j.ejca.2021.09.008

Table 2.

Multivariable model for the risk of recurrence prediction

Predictor β SE(P) HR (95% CI) P-value Risk points (max. 100)

Histotype Squamous cell Reference 0
Adenocarcinoma 0.342 0.116 1.408 (1.120; 1.771) 0.003 7
Adenosquamous 0.598 0.164 1.819 (1.317; 2.513) < 0.001 11
Neuroendocrine 1.741 0.246 5.704 (3.514; 9.260) < 0.001 33
Other 1.145 0.270 3.144 (1.848; 5.349) < 0.001 22

Tumour diameter < 0.5 cm Reference 0
0.5–1.99 cm 0.501 0.237 1.651 (1.035; 2.634) 0.035 10
2–3.99 cm 1.115 0.236 3.051 (1.915; 4.858) < 0.001 21
≥ 4 cm 1.556 0.245 4.738 (2.925; 7.674) < 0.001 30

Grade 1 Reference 0
2 0.260 0.214 1.297 (0.852; 1.976) 0.235 5
3 0.457 0.247 1.579 (0.970; 2.570) 0.085 9

Positive pelvic LN 0 / not assessed Reference 0
1 0.255 0.154 1.291 (0.953; 1.748) 0.098 5
2 0.482 0.170 1.619 (1.158; 2.264) 0.005 9
≥ 3 0.939 0.144 2.557 (1.927; 3.394) < 0.001 18

LVSI No / not assessed Reference 0
Yes 0.538 0.106 1.713 (1.390; 2.111) < 0.001 10

β: beta coefficient; CI: confidence interval; LN: lymph node; LVSI: lymphovascular space invasion; SE: standard error.