In a predictive outcome assessment test for predicting whether a patient undergoing a breast cancer treatment regimen will achieve pathological complete response (pCR), differential gene expression level information are generated for an input set of genes belonging to the TGF-β signaling pathway. The differential gene expression level information compares baseline gene expression level information from a baseline sample (70) of a breast tumor of a patient acquired before initiating (71) a breast cancer therapy regimen to the patient and response gene expression level information from a response sample (72) of the breast tumor acquired after initiating the breast cancer therapy regimen by administering a first dose of bevacizumab to the patient. A pCR prediction for the patient is computed based on the differential gene expression level information for the input set of genes belonging to the TGF-β signaling pathway. Related predictive outcome assessment test development methods are also disclosed.