Skip to main content
. 2023 Nov 3;13:1274557. doi: 10.3389/fonc.2023.1274557

Table 2.

Patient demographics data and breast lesion characteristics.

Characteristic Retrospective
Training and Validation sets
Prospective test sets
Hospital 1 Set Hospital 2 Set
Number of patients 420 214 64
Age (years) (mean) 45 ± 12 45 ± 12 50 ± 11
Number of lesions 480 228 64
Lesion maximum diameter(mm)
2-10 167(35%) 89(39%) 13(20%)
10-20 185(38%) 90(40%) 34(53%)
20-30 128(27%) 49(21%) 17(27%)
BI-RADS category a
2 34(7%) 18(8%) 0(0%)
3 110(23%) 30(13%) 10(16%)
4a 106(23%) 70(31%) 8(13%)
4b 80(17%) 41(18%) 14(22%)
4c 74(15%) 35(15%) 22(34%)
5 76(16%) 34(15%) 10(15%)
Lesion type
 Invasive ductal carcinoma 82(17%) 26(11%) 14(22%)
 Invasive lobular carcinoma 73(15%) 43(19%) 16(25%)
 Ductal carcinoma in situ 21(4%) 10(4%) 6(9%)
Other malignant b 20(4%) 4(2%) 5(8%)
Fibroadenoma 141(30%) 64(28%) 13(20%)
Other benign c 143(30%) 81(36%) 10(16%)

The BI-RADS category a is based on the interpretation of the radiologist who originally performed the US examinations before the biopsy test, not the radiologists involved in the reader study. It should be noted that all BI-RADS categories involved in this study were determined on breast-US video images only. b Includes non-specific malignant results. c Includes adenosis, hyperplasia, benign phyllodes tumors, and papillomata.