Table 2

Univariable and multivariable COX proportional hazards regression analysis for the predictors of nodal metastasis

UnivariableMultivariable
HR (95% CI)P valueHR (95% CI)P value
Age, years0.99 (0.97 to 1.02)0.454
Sex (female vs male)1.40 (0.73 to 2.68)0.309
Laterality (left vs right)1.16 (0.62 to 2.16)0.648
Diabetes0.61 (0.15 to 2.55)0.502
History of diuretic use0.74 (0.23 to 2.39)0.613
HBV, HBsAg (+)0.05 (0 to 29.44)0.351
Diagnostic delay, months1.00 (0.99 to 1.01)0.550
Upper lid0.97 (0.52 to 1.82)0.926
Lower lid1.08 (0.58 to 2.02)0.814
Both upper lid and lower lid2.21 (0.68 to 7.18)0.186
Medial canthus3.26 (1.50 to 7.08)0.003*1.09 (0.35 to 3.44)0.878
Initial diagnosis (others vs sebaceous carcinoma)0.76 (0.38 to 1.53)0.442
Tumour presentation pattern (diffuse vs nodule)4.16 (1.73 to 10.00)0.001*4.34 (1.75 to 10.76)0.002*
Greatest basal diameter, mm1.05 (1.01 to 1.08)0.008*1.02 (0.99 to 1.06)0.220
Orbital invasion at presentation3.98 (1.83 to 8.65)<0.001*3.22 (1.42 to 7.33)0.005*
Pagetoid spread1.02 (0.52 to 2.01)0.953
Multicentric origin1.88 (0.83 to 4.25)0.130
PNI6.51 (2.30 to 18.42)<0.001*3.24 (1.11 to 9.49)0.032*
Perivascular invasion2.27 (0.31 to 16.57)0.419
Muscle infiltration1.76 (0.86 to 3.61)0.121
Ki-67 percentage1.04 (1.02 to 1.05)<0.001*1.03 (1.01 to 1.05)<0.001*
Histological differentiation
 Moderate versus well1.28 (0.45 to 3.66)0.650
 Poor versus well2.67 (0.84 to 8.53)0.097
With positive surgical margin1.01 (0.42 to 2.41)0.991
Local recurrence1.23 (0.58 to 2.58)0.590
Initial treatment
 Wide local excision versus frozen margin control2.05 (0.87 to 4.87)0.103
 Orbital exenteration versus frozen margin control10.29 (3.44 to 30.75)<0.001*
 Wide local excision versus orbital exenteration0.20 (0.08 to 0.49)<0.001*
  • *Statistically significant.

  • HBV, hepatitis B virus; PNI, perineural invasion.