Table 3

Stepwise backward elimination method of multinomial logistic regression

Parameter estimates OR (95% CI)
DysplasiaOral squamous cell carcinoma
N=Mild
(n=37)		Moderate/severe
(n=34)		Well differentiated
(n=86)		Poorly differentiated
(n=27)
Variables in the model
 hMLH10.92** (0.88 to 0.97)0.97 (0.92 to 1.03)0.92* (0.88 to 0.96)0.92** (0.85 to 1.00)
 hPMS20.94** (0.91 to 0.98)0.98 (0.94 to 1.03)0.94* (0.90 to 0.97)0.89* (0.84 to 0.95)
 hMSH20.96 (0.92 to 1.00)0.97 (0.93 to 1.01)0.93** (0.89 to 0.97)0.86* (0.81 to 0.91)
Control variables
 Age1.02 (0.98 to 1.05)1.04** (1.00 to 1.07)1.06** (1.02 to 1.09)1.07** (1.02 to 1.13)

  • The reference group is normal; *p<0.0001, **p<0.05. The general formula by which the model was designed is as follows: Embedded Image in which p=probability, Dx=Diagnosis of sample (mild dysplasia, moderate/severe dysplasia, well-differentiated OSCC and poorly differentiated OSCC), b=interface, β=coefficient for each independent variable, x=independent variable and i=score for which model is calculated. Following equations were used for data shown in the table:

    Embedded Image

    Embedded Image

    Embedded Image

    Embedded Image

  • OSCC, oral squamous cell carcinoma.