Research Square · Nov 10, 2026
Breast cancer origin: a nine-state linear model: B3 two-axis power target combinations, SEER calibration, and state-resolved readouts
By wei q
Background Breast cancer is the most common cancer in women, and its incidence is the end readout of a latent process that begins with epithelial proliferation and passes through hyperplasia, in situ carcinoma and invasive cancer. We treat the tissue layer as the B3 two-axis module, proliferation-metabolic activation with involution-clearance failure, and ask whether a nine-state linear model reproduces SEER incidence, places the module at B3 and the six product groups at B4, and how the nine-state progression model compares with a three-state detection design. Methods We built a nine-state model (H, B1-B7, D). Targets were assigned to the two B3 axes by measured function, and PubMed records measured evidence volume and coverage. The model was calibrated to SEER 2012–2016 age-specific incidence, and AIC, BIC and leave-one-age-group cross-validation compared the one-state, three-state and nine-state structures. Results The two axes covered 47 targets and 21,555 records, and the six product groups covered 51 targets and 11,900 records. The three-state detection design had the lower AIC (10.61 versus 30.77) because the four intermediate rates were nearly identical; the nine-state model resolved B1-B7 and placed B3 before B7. State occupancy followed B1-B7, with B7 rising from 0.43% at age 40 to 9.09% at age 80. Conclusion The tissue layer of breast cancer is the B3 two-axis module. The nine-state model resolves the progression structure, and the three-state model is the detection design for the composite B3 transition. B4 to B5 is the cell-mutation axiom, and B5-B7 is the established mutation-clone chain.