Table 1 Demographic and clinical characteristics of patients in the training and validation datasets.

From: An ensemble deep learning model for risk stratification of invasive lung adenocarcinoma using thin-slice CT

  

Training dataset (n = 843)

Validation dataset

(n = 232)

Age

Mean (SD)

58.4(±11.5)

57.7(±11.7)

Gender

Male

198(40.6%)

54(38.8%)

Female

290(59.4%)

85(61.2%)

Pathological type

Benign

216(25.6%)

78(33.6%)

Pre-IA

288(34.2%)

78(33.6%)

IAC (Grade1)

124(14.7%)

29(12.5%)

IAC (Grade2)

116(13.8%)

26(11.2%)

IAC (Grade 3)

99(11.7%)

21(9.1%)

Location

LUL

200(23.7%)

49(21.1%)

LLL

140(16.6%)

32(13.8%)

RUL

268(31.8%)

88(38.0%)

RML

78(9.3%)

15(6.5%)

RLL

157(18.6%)

48(20.7%)

Nodule morphology

pGGN

287(34.0%)

63(27.2%)

Heterogeneous ground-glass nodule

41(4.9%)

20(8.6%)

Part-solid nodule

182(21.6%)

55(23.7%)

Pure solid nodule

333(39.5%)

94(40.5%)

Diameter (mm)

≤10 mm

429(50.9%)

122(52.6%)

10–20 mm

280(33.2%)

72(31.0%)

20–30 mm

134(15.9%)

38(16.4%)

Average diameter (mm)

Mean (SD)

Benign

Malignant

8.6( ± 7.6)

14.3( ± 8.8)

8.9( ± 7.3)

14.2( ± 8.1)

  1. n is the number of nodules, IAC invasive adenocarcinoma, RUL right upper lobe, RML right middle lobe, RLL right lower lobe, LUL left upper lobe, LLL left lower lobe, pGGN pure ground-glass nodule.