A Combination Analysis of IVIM-DWI Biomarkers and T2WI-Based Texture Features for Tumor Differentiation Grade of Cervical Squamous Cell Carcinoma

Table 4

Statistical results of regression analysis.

	Regression model	AUC (95% CI)	Sensitivity	Specificity
Poorly vs. moderately	Ŷ = 30.43 − 6.755  f − 2.772  difference average − 50.441  inverse variance − 0.007  SRHGLE	0.797 (0.718–0.875)	97.96	50.70
Moderately vs. well	Ŷ = 13.918–23.142  ADC + 22.326  D + 0.08   − 15.405  IDMN	0.954 (0.882–1.000)	93.88	94.44
Poorly vs. moderately and well	Ŷ = 1.486 + 3.903  D − 5.558  f − 0.003  range	0.795 (0.720–0.870)	88.06	64.79
Well vs. moderately and poorly	Ŷ = 14.644–30.501  ADC + 33.041  D − 16.553  IDMN	0.952 (0.888–0.952)	91.69	94.44

IDMN represented for inverse difference moment normalized. SRHGLE represented for short run high gray-level emphasis. The units of the sensitivity and specificity are %.