Table 4 Logistic regression model for differentiate MIA from precursor lesions.

From: Correlation between imaging features of pure ground-glass opacities and pathological subtypes of lung minimally invasive adenocarcinoma and precursor lesions

Features

Univariate analysis

Multivariate analysis

p

OR (95% CI)

p

OR (95% CI)

Maximum radiodensity (HU)

 < − 443.00

< 0.001

 

< 0.001

 

 − 443.00 to − 220.00

0.017

2.606 (1.188–5.718)

0.009

3.109 (1.330–7.266)

 > − 220.00

< 0.001

13.054 (3.759–45.335)

< 0.001

12.955 (3.562–47.110)

Minimum radiodensity (HU)

 < − 775.00

0.006

   

 − 775.00 to − 660.00

0.176

1.725 (0.0.783–3.803)

  

 > − 660.00

0.001

4.706 (1.824–12.141)

  

Mean radiodensity (HU)

 < − 650.00

< 0.001

   

 − 650.00 to − 564.00

0.244

1.785 (0.673–4.733)

  

 > − 564.00

< 0.001

5.355 (2.269–12.638)

  

Variance

 < 80.50

0.002

   

 80.50 to 125.00

0.009

2.840 (1.305–6.180)

  

 > 125.00

0.002

6.345 (2.026–19.877)

  

Skewness

 < − 0.07

0.011

 

0.045

 
 

0.004

0.111 (0.025–0.503)

0.013

0.136 (0.028–0.653)

 > 0.02

0.233

0.500 (0.160–1.564)

0.806

0.858 (0.251–2.926)

Entropy

 < 8.45

0.016

   

 8.45 to 8.80

0.005

4.500 (1.565–12.940)

  

 > 8.80

0.361

2.200 (0.405–11.949)

  
  1. MIA, minimally invasive adenocarcinoma.