Research Article

A Machine Learning Approach for the Association of ki-67 Scoring with Prognostic Factors

Table 1

Patient characteristics (discrete data are given as numbers, continuous as the mean ± standard deviation) (n = 223).

VariableKi-67 ≤ Ki-67 > p

Ki-67 Proliferation74 (33.2)149 (66.8)0.001

Molecular Classification
LA37 (50)1 (0.7)!0.001
LB28 (37.8)112 (75.2)
Her-26 (8.2)27 (18.1)
TN3 (4)9 (6)

Lymphovascular Invasion
-36 (48.6)32 (21.5)0.001
+38 (51.4)117 (78.5)

Age
Age <5038 (51.4)80 (53.7)0.742
Age >5036 (48.6)69 (46.3)

Number of Metastatic Lymph Nodes
0-3 LNm27 (36.5)38 25.5)0.234
4-9 LNm8 (10.8)18 (12.1)
10 LNm39 (52.7)93 (62.4)

Nuclear Grade
Grade I26 (35.1)16 (10.7)0.001
Grade II33 (44.6)82 (55)
Grade III15 (20.3)51 (34.2)

Tumor Size
T132 (43.2)41 (27.5)0.01
T238 (51.4)82 (55)
T3 and T44 (5.4)26 (17.4)

Body Mass Index
18.5-24.97 (9.5)18 (12.1)0,386
25-29.928 (37.8)43 (28.9)
30+39 (52.7)88 (59.1)

Surgical Type
Mastectomy55 (74.3)77 (51.7)0,001
Segmental Mastectomy19 (25.7)72 (48.3)

Histopathological Type
IDC61 (82.4)131 (87.9)0,403
ILC6 (8.1)6 (4)
Other7 (9.5)12 (8.1)

Perivascular Invasion
-52 (70.3)36 (24.2)0.01
+22 (29.7)113 (75.8)

Number of Lymph Nodes
2.39 ± 4.623.74 ± 6.30.07

: 6 mucinous carcinomas, 7 DCIS, 6 neuroendocrine carcinomas. !: “rfinput” imputation. Molecular classification: LA: luminal A, LB: luminal B, and TNBC: triple negative breast carcinoma; BMI: underweight <18.50, normal range 18.50-24.99, overweight ≥ 25.00, and obese ≥ 30.00.