Hypothesis: hematogenous metastatic cancer cells of solid tumors may disguise themselves as memory macrophages for metastasis.

German pathologist Otto Aichel suggested, a century ago, that the cancer cell acquired its metastatic property from a leukocyte via cell-cell fusion. Since then, several revised versions of this theory have been proposed. Most of the proposals attribute the generation of the metastatic cancer cell to the fusion between a primary cancer cell and a macrophage. However, these theories have not addressed several issues, such as dormancy and stem cell-like self-renewal, of the metastatic cancer cell. On the other hand, recent studies have found that, like T- and B-/plasma cells, macrophages can also be categorized into naïve, effector, and memory/trained macrophages. As a memory/trained macrophage can enter dormancy/quiescence, be awakened from the dormancy/quiescence by acquainted primers, and re-populate via stem cell-like self-renewal, we, therefore, further specify that the macrophage fusing with the cancer cell and contributing to metastasis, belongs with the memory/trained macrophage, not other subtypes of macrophages. The current theory can explain many puzzling clinical features of cancer, including the paradoxal effects (recurrence vs. regression) of microbes on tumors, "spontaneous" and Coley's toxin-induced tumor regression, anticancer activities of β-blockers and anti-inflammatory/anti-immune/antibiotic drugs, oncotaxis, surgery- and trauma-promoted metastasis, and impact of microbiota on tumors. Potential therapeutic strategies, such as Coley's toxin-like preparations, are proposed. This is the last article of our trilogy on carcinogenesis theories.