Decreased tumor burden has sometimes been shown to increase the usefulness of systemic therapy, but I think that, all in all, case selection simply promotes breast surgery in patients with a better prognosis.
Dr. Blake Cady
To show improved survival, we must find causality and not simply coincidence. Findings from large case series or cancer databases are unable to provide the answer. Although registry interpretations can provide basic data, they often break down when you attempt to extract detailed data.
Sometimes a stage IV cancer is registered initially on the basis of a pulmonary or liver shadow on imaging, which later is proven not to have been metastatic disease – and the case is never amended in the registry. There are also registry misinterpretations of the type of surgery that’s been done.
We performed a matched pair analysis of stage IV breast cancer patients with and without resection of the primary tumor. This was based on the Partners Health Care Consortium cancer database, which included information on 19,464 invasive breast cancers treated between 1970 and 2002; 808 of these (4%) were recorded as stage IV. Of these, 622 were analyzed by case matching that considered age, diagnostic date, metastatic disease location, estrogen-receptor status, and systemic therapy (Ann. Surg. Oncol. 2008; 15:3384-95).
It’s absolutely true that there was a statistically significant difference in survival in the overall comparison of surgery vs. no surgery. With bone metastases only, it was still statistically significant, but with a substantially lower difference in survival. In visceral metastases, the difference became nonsignificant.
Looking at therapy sequence in those who had the same type of systemic therapy, but with chemotherapy before or after surgery, you see a suggestion that giving systemic therapy first with a clinical response, which is a favorable prognostic indication, leads to a selection bias in choosing patients for surgery.
I reviewed the 86 cases who survived 5 years with records available. In those records, 22 were actually not stage IV disease, leaving 64 with confirmed stage IV (M1) – an overall 10% 5-year survival rate. Patients who survived were significantly younger than those who did not survive, significantly more likely to have estrogen receptor-positive tumors, and significantly more likely to have bone rather than visceral metastases or oligometastatic disease – all of which are good prognostic factors.
Among the 64 5-year survivors, 25 had therapeutic breast surgery. Of these, 8 had an excellent chemotherapy response followed by breast surgery; 8 had delayed surgery following a failure of systemic therapy; 7 had an initial therapeutic and curative surgery, but staging revealed metastases because of unexpected positive nodes; and 2 had initial surgery with oligometastatic resection.
Thirty-nine patients had no therapeutic breast surgery, and, of these, 16 had an excellent response to systemic therapy; 15 had slowly progressive disease; and 8 had oligometastatic disease – a positive prognostic factor. Thus, favorable prognostic factors led to a selection of primary-site surgery, whereas poor prognostic factors mean that primary-site surgery was avoided, contradicting the assumption of its therapeutic benefit.
There were also surgical discrepancies among the 5-year survivors. The registry listed 11 as having had breast surgery, while a chart review showed that surgery was not therapeutic. Five were listed as having no breast surgery, while they had indeed undergone a therapeutic operation.
Cancer registry data are not reliable for defining stage IV disease, discriminating advanced local disease with bone involvement from true distant, imaging-confirmed metastasis, or assessing surgical procedures. Case selection bias accounts for most, if not all, of the apparent survival advantage.
Dr. Blake Cady is professor of surgery (emeritus) at Harvard Medical School, Boston, and at Brown University, Providence, R.I.