Skip to main content
. 2020 Feb 12;27(7):2516–2524. doi: 10.1245/s10434-020-08235-w

Table 2.

Univariable and multivariable Cox regression analyses according to the Amsterdam model in patients with resected pancreatic cancer diagnosed between 2005 and 2016

Clinicopathological parameter Median OS, months 5-year survival (%) Univariable analysis
HR (95% CI)
Multivariable analysis
HR (95% CI)b
p valuec
Age (each incremental year) – – 1.01 (1.00–1.01) 1.00 (1.00–1.01) 0.04
Tumor differentiation grade
 Well-differentiated 27 27 1.00 (reference) 1.00 (reference)
 Moderately differentiated 19 16 1.41 (1.21–1.65) 1.27 (1.11–1.46) 0.001
 Poorly or undifferentiated 14 12 1.94 (1.66–2.28) 1.74 (1.51–2.00) < 0.001
Lymph node ratioa
 0 (lymph node-negative) 25 28 1.00 reference 1.00 (reference)
 > 0 and ≤ 0.18 18 13 1.44 (1.27–1.63) 1.47 (1.31–1.64) < 0.001
 > 0.18 15 8 1.86 (1.67–2.07) 1.94 (1.76–2.14) < 0.001
Resection margin
 R0 20 19 1.00 (reference) 1.00 (reference)
 R1/R2 14 8 1.57 (1.44–1.70) 1.44 (1.33–1.57) < 0.001
Adjuvant chemotherapy
 Yes 21 20 1.00 (reference) 1.00 (reference)
 No 14 11 1.52 (1.41–1.65) 1.64 (1.51–1.79) < 0.001

Data after multiple imputation were used

OS overall survival, HR hazard ratio, CI confidence interval

aLymph node ratio is the number of positive lymph nodes divided by the total number of lymph nodes harvested

bHazard ratios and 95% CIs from the Cox LASSO model are presented

cp values of multivariable analyses are shown