This study by Jonathan W. Uhr and colleagues at the University of Texas Southwestern addresses the question of why breast cancer can recur so late after apparently successful treatment and reports a provocative observation. Such observations (and others) have led to the concept of tumor "dormancy," in which a patient lives apparently tumor-free but harbors viable tumor that is undetectable by present clinical tests. Certainly tumor dormancy has been shown conclusively in animal models. For instance, Judah Folkman has shown that aangiogenesis inhibitors can induce tumor dormancy in mice. Other researchers are working to develop strategies to induce tumor dormancy as a therapeutic strategy for cancer, in effect turning cancer into a chronic, manageable disease like diabetes or hypertension.
Here's what they did. They looked at women who had had mastectomies for breast cancer and had no clinical evidence of recurrent disease for seven or more years. Their hypothesis was that breast cancer patients could still have tumor cells circulating in the blood that long out (which is why they chose mastectomy patients, to avoid the confounding possibility that there were still tumor cells in the breast after lumpectomy) and that such patients still had "dormant" tumors that served as a source for these CTCs. They compared these women to age-matched controls. Blood samples were taken and special flow cytometry assays done for CTCs. Amazingly, they found that 13 of 36 of these patients had CTCs in their blood with the morphology of breast cancer cells. One patient positive for CTCs was 22 years out from her mastectomy. They next measured the half life of these cells by studying women undergoing mastectomy. CTC levels in the blood were measured before surgery started and then at various time points after the tumor was removed. The half-life of the CTCs was found to be between 1 and 2.5 hours, and in all cases the level fell to background levels within 24 hours. Recurrent tumor as a source of CTCs was excluded using clinical examination and imaging studies. The patients were followed for 1-2 years, and none were reported to have had a recurrence during that time.
Uhr concluded from these observations that a significant number (1/3) still had CTCs in their blood up to 22 years after definitive treatment, but had not developed developed recurrent disease. Moreover, because the half-life of these cells was short, that suggests that there must be active, replicating tumor cells within the body serving as a source of these CTCs. (One could postulate that the half life of dormant tumor cells in the blood is considerably longer than that of active tumors being surgically resected, but it is unlikely that the half life would be many years, which is what would be required to support an explanation in which no dormant tumor is present in the body but CTCs are still detectable many years later.) That supports the concept of "dormant" tumors, in which there is a balance between replication and tumor death, such that microscopic deposits of tumor do not increase in size or spread.
Unfortunately, this study produces more questions than answers. For example, we have no way yet to know whether the presence of these cells is an indicator that the patient's tumor will recur, although certainly that is certainly the next hypothesis that will need to be tested. Because recurrence occurs only at a 1% per year rate to begin with, it would take a large study with many patients followed over several years to work this out. If CTC levels are bad prognostic factors for late recurrence, then perhaps they could be used clinically as a test, much like Her-2/neu (whose presence on breast cancer cells indicates a more aggressive tumor). Next, we have no idea whether treating patients who are found to have CTCs in their blood would prevent or lower the rate of recurrence. That would take even a study with even more patients. Even more interesting would be to be able to gather enough of these cells to isolate RNA to do whole genome gene expression profiling on (the "Gene Chip"), to see how their gene expression patterns differ from that of the original tumor. Obviously something is different, if the tumor hasn't recurred. The results could suggest strategies for inducing tumor dormancy.
In any case, because my main research interest is tumor angiogenesis. Given that blockers of tumor angiogenesis can cause tumor dormancy in experimental tumor models, my interest in this is more than passing. I have to give a tip of the hat to one of my partners for pointing this paper out to me. (The issue of Clinical Cancer Research containing the paper arrived while I was on vacation over Christmas, and I never got around to looking through it.) We will be hearing more about tumor dormancy in the next few years.