Table 3.
Predictive capacity following the consecutive addition of biomarkers to clinical variables.
Models | DFS | OS | ||||||||
---|---|---|---|---|---|---|---|---|---|---|
AIC | LLR† | P† | LLR‡ | P‡ | AIC | LLR† | P† | LLR‡ | P‡ | |
Model 1 pCR + HR/HER2 status |
79 | 22.50 | <0.001 | 15.46 | <0.001 | 70 | 20.69 | <0.001 | 10.71 | 0.005 |
Model 2 + NLR + TILs |
68 | 7.04 | 0.030 | ref | n/a | 64 | 9.98 | 0.007 | ref | n/a |
Model 3 +AURKA + MYBL2 |
64 | ref | n/a | n/a | n/a | 58 | ref | n/a | n/a | n/a |
†Likelihood ratio test (LLRT) and p value for comparison of model 3 (reference category) and models 1 and 2 (nested models).
‡ Likelihood ratio test (LLRT) and p value for comparison of model 2 (reference category) and model 1 (nested model).
AIC, Akaike Information Criterion; AURKA, Aurora kinase A; DFS, disease-free survival; HER2, human epidermal growth factor receptor 2; HR, hormone receptor; NLR, neutrophil-to-lymphocyte ratio; MYBL2, MYB Proto-Oncogene Like 2; OS, overall survival; pCR, pathological complete response; TILs, tumor-infiltrating lymphocytes; Ref, reference; N/a, not applicable.