Research Article
Imaging Biomarkers as Predictors for Breast Cancer Death
Table 2
Relationship of mammographic appearance to basal phenotype, histologic tumor distribution, and conventional tumor attributes in both cohorts combined.
| | | Stellate/Powdery/Crushed stone-like | Circular | Casting | Architectural distortion | NK | χ2 | P value |
| | Basal phenotype | | | | | | 89.71 | <0.0001 | | Non-basal | 619 (96.87) | 270 (78.03) | 59 (88.06) | 45 (83.33) | 42 | | | | Basal | 20 (3.13) | 76 (21.97) | 8 (11.94) | 9 (16.67) | 11 | | | | NK | 105 | 50 | 11 | 8 | 13 | | | | Histologic tumor distribution | | | | | | 817.74 | <0.0001 | | Unifocal | 450 (63.11) | 260 (68.78) | 7 (9.21) | 7 (11.48) | 33 | | | | Multifocal | 257 (36.04) | 116 (30.69) | 12 (15.79) | 9 (14.75) | 22 | | | | Diffuse | 6 (0.84) | 2 (0.53) | 57 (75.00) | 45 (73.77) | 4 | | | | NK | 31 | 18 | 2 | 1 | 7 | | | | Tumor size (mm) | | | | | | 53.96 | <0.0001 | | <15 | 340 (47.82) | 127 (33.78) | 31 (42.47) | 3 (5.00) | 20 | | | | ≥15 | 371 (52.18) | 249 (66.22) | 42 (57.53) | 57 (95.00) | 40 | | | | NK | 33 | 20 | 5 | 2 | 6 | | | | Histologic malignancy grade | | | | | | 128.08 | <0.0001 | | I/II | 613 (86.46) | 227 (60.21) | 35 (48.61) | 53 (86.89) | 40 | | | | III | 96 (13.54) | 150 (39.79) | 37 (51.39) | 8 (13.11) | 17 | | | | NK | 35 | 19 | 6 | 1 | 9 | | | | Lymph node involvement | | | | | | 22.52 | <0.0001 | | Negative | 479 (66.81) | 258 (68.07) | 45 (59.21) | 23 (38.33) | 39 | | | | Positive | 238 (33.19) | 121 (31.93) | 31 (40.79) | 37 (61.67) | 20 | | | | NK | 27 | 17 | 2 | 2 | 7 | | |
|
|
Those with unknown mammography feature were not included in the chi-square test. NK, not known. |
