The tool explicitly separates the survival baseline (surgery alone) from the incremental, absolute percentage gains of hormone therapy, chemotherapy, and Trastuzumab (for HER2+).

While PREDICT is mathematically brilliant for populations, it struggles with the biological heterogeneity of ER+/HER2- tumors. For an intermediate-risk ER+ patient, PREDICT might estimate a 4% chemo benefit. However, an Oncotype DX test might reveal a Recurrence Score of 10 (meaning 0% actual benefit) or a score of 40 (meaning massive benefit). Thus, PREDICT is best used *before* genomic testing to decide if the genomic test is even worth ordering.