M eric et al. 1 recently concluded that currently available data are not sufficient to determine conclusively whether distant metastases are caused by the development of ipsilateral breast tumor recurrence (IBTR). 1 The authors of at least four randomized, prospective trials do not agree. 2-5 For example, the National Surgical Adjuvant Breast and Bowel Project (NSABP) concluded that local recurrence is a marker for risk (and not a cause) of distant metastases. 2 Meric et al. also concluded that it would be difficult to argue biologically that malignant cells in the primary tumor could metastasize and affect survival while arguing that malignant cells in the recurrent tumor could not. I have offered this explanation. 6 It is likely that cancer cells begin to circulate when a tumor is very small. Cells shed from a small tumor may succumb to a variety of host defense factors-mechanical, biochemical, immunologic, etc. After breast-conserving surgery, cancer may recur locally. (Here, I prefer the term local persistence.) Patients who survive their first cancer without developing distant metastases may be expected to survive the recurrence of a similar volume of tumor, because the host defense factors that killed tumor cells from the primary tumor can be expected to kill tumor cells shed from the recurrence (persistence). Local recurrence after radical surgery has always been an ominous event. It may represent the return of malignant cells from a distant organ to the site of surgery. Local recurrence also may occur after breast-conserving surgery. This is one reason why patients who develop IBTR have a poorer prognosis compared with patients who do not. In NSABP trial B-06, patients who developed local recurrence were 3.41 times as likely to also develop distant metastases. 2 This is not evidence that IBTR is the cause of tumor spread. Investigators from the NSABP trial B-06 recently reported a marginally significant decrease in breast cancer deaths among women who received postoperative radiotherapy. 7 This finding is fully compatible with the above hypothesis. The excess deaths in the unirradiated group may have occurred in women whose recurrences exceeded their primary tumors in terms of volume. I asked the NSABP for data on the size of tumor recurrences and was told that such data are not available. Meric et al. discussed the possible survival advantage associated with local control. They cited important studies that will help us fine-tune our understanding of this dilemma. However, the past 15 years have witnessed a dramatic change in our understanding of local recurrence. Local persistence seldom metastasizes in patients with the soft tissue sarcomas, rectal carcinoma, melanoma, and other solid tumors.