Table 1 Clinical and pathological characteristics of training set and validation set.

From: Deep-learning based discrimination of pathologic complete response using MRI in HER2-positive and triple-negative breast cancer

Characteristics

Total

(n = 852)

Training set

(n = 724)

Validation set

(n = 128)

P value

Age (years)

51 ± 10

51 ± 10

52 ± 10

0.483

Tumor size on pre-NAC MRI (cm)

4.1 ± 2.1

4.2 ± 2.1

3.7 ± 1.8

0.014

Clinical T stage

   

0.194

T1

48 (5.6)

39 (5.4)

9 (7.0)

T2

575 (67.5)

481 (66.4)

94 (73.4)

T3

141 (16.5)

124 (17.1)

17 (13.3)

T4

88 (10.3)

80 (11.0)

8 (6.3)

Clinical N stage

   

0.003

N0

124 (14.6)

94 (13.0)

30 (23.4)

N1

358 (42.0)

303 (41.9)

55 (43.0)

N2

240 (28.2)

217 (30.0)

23 (18.0)

N3

130 (15.3)

110 (15.2)

20 (15.6)

Estrogen receptor status

Negative

Positive

583 (68.4)

269 (31.6)

494 (68.2)

230 (31.8)

89 (69.5)

39 (30.5)

0.771

Progesterone receptor status

Negative

Positive

693 (81.3)

159 (18.7)

586 (80.9)

138 (19.1)

107 (83.6)

21 (16.4)

0.477

HER2 status

Negative

Positive

326 (38.3)

526 (61.7)

271 (37.4)

453 (62.6)

55 (43.0)

73 (57.0)

0.235

Ki-67 index (%)

22 ± 20

22 ± 20

22 ± 22

0.919

Histologic grade

   

0.091

Low or intermediate

381 (44.7)

315 (43.5)

66 (51.6)

High

471 (55.3)

409 (56.5)

62 (48.4)

Histologic type

   

0.475

Ductal

831 (97.5)

705 (97.4)

126 (98.4)

Others

21 (2.5)

19 (2.6)

2 (1.6)

Pathologic complete response

   

0.495

No

542 (63.6)

464 (64.1)

78 (60.9)

Yes

310 (36.4)

260 (35.9)

50 (39.1)

  1. Note. – For continuous variables, data are means with standard deviations (SD) in parentheses. For categorical variables, data are number of women with percentages in parentheses. HER2 = human epidermal growth factor receptor 2.