If you look at any collection of "common myths about cancer", you will probably find reassurances about the idea that having cancer surgery might cause the cancer to spread to other parts of the body. I remember coming across this one some years ago being surprised - I'd never heard that one myself, but it was apparently a pretty widespread fear. You could imagine a situation where tumor cells are mechanically detached and could spread, but that's actually less likely than it might seem, since not all of these cells can actually attach somewhere else and start growing again (although there are special situations where such disturbance, even to the level of taking a biopsy sample, is known to be risky).
What I didn't know enough about, though, is an effect seen in breast cancer treatment. After surgery, the risk of being diagnosed with metastatic breast cancer peaks about 12 to 18 months later. It's not a sure thing, fortunately, but it's definitely a real effect, and one that would certainly lead to people thinking that the surgery had (somehow) spread the cancer around. That's actually a pretty reasonable hypothesis, given the observations, so that particular "common myth" has something behind it. But how? The same effect is noted in patients who have undergone complete mastectomy, which doesn't disturb the tumor tissue itself much (or at all). Alternatively, were these actively growing tumors that had spread before the surgery and were just too small to detect? Some have argued for this explanation, and it's hard to disprove. There's yet another plausible mechanism: these tumor cells might have had already spread before the surgery, but were somehow not growing, and something happened to change that. What is it about surgery that would allow them to start growing?
This new paper [doi.org/10.1126/scitranslme...] has what may well be the answer. Studying the effects of surgery in mouse models of metastatic cancer, it appears that the post-surgical wound healing response is the culprit. There appears to be an ongoing T-cell response that is holding down these potential metastatic tumors, but it's disrupted by the inflammation response after surgery. Expressing GFP in murine mammary tumors and then injecting these into other (syngenic) mice caused a T-cell response to the GFP itself, and ultimately the tumors were rejected by the immune system. But surgically wounding these mice triggered an outgrowth of the tumors instead, probably through the actions of inflammatory monocytes (which differentiate into macrophages inside the tumor tissue).
What's even more interesting is that these effects can be abrogated simply by treating the mice with NSAIDs to damp down the inflammation response. To avoid interactions between the tumor cells and the anti-inflammatory drugs, the procedure was to do the surgical wounding (with and without NSAID treatment with meloxicam) several days before the tumor cells were even injected, with time enough to clear the system out!
Prostate cancer uses fat to spread and resist medications
Prostate cancer spread to pelvic and lymph nodes 
PSA cut in half by cancer still spreading?
Why keep prostate after cancer has spread?
Could exercise prevent prostate cancer spreading to the bones?
