Table 2 Histopathologic characteristics of the NN-master set.

From: A machine learning algorithm for simulating immunohistochemistry: development of SOX10 virtual IHC and evaluation on primarily melanocytic neoplasms

Case

Location

Lesion type

Sub-images

Number of cells

Cell-fraction of data set [%]

SOX10-positive

SOX10-negative

Percent SOX10-positive [%]

1

Skin

MIS

1,987

654

317,261

0.2

8.4

2

Skin

MIS

1,290

594

187,875

0.3

5.0

3

Skin

Melanoma

2,168

19,142

336,058

5.4

9.4

4

Skin

Melanoma

2,004

19,551

264,744

6.9

7.5

5

Skin

BCC

260

1,051

58,222

1.8

1.6

6

Skin

Melanoma

975

25,493

167,175

13.2

5.1

7

Skin

Neuroma

361

556

66,314

0.8

1.8

8

Skin

Melanoma

488

31,210

99,420

23.9

3.4

9

LN

MM

2,796

18,402

661,343

2.7

17.9

10

LN

MM

2,249

61,901

623,796

9.0

18.1

11

Skin

Melanoma

768

10,964

91,754

10.7

2.7

12

LN

MM

2,776

116,648

522,706

18.2

16.9

Total

  

18,122

306,166

3,396,668

8.3

100

  1. BCC basal cell carcinoma, LN lymph node, MIS melanoma in situ, MM metastatic melanoma.