Abstract & Commentary
Synopsis: This study performed a retrospective analysis of breast cancer patients with an isolated local recurrence. They found that prognostic factors for death included higher tumor grade, older age at diagnosis, and disease-free interval. For premenopausal patients not previously exposed to systemic therapy, the use of chemotherapy or ovarian suppression was associated with a decreased risk of death.
Source: Le M, et al. Cancer. 2002;94:2813-2820.
As many as 12-13% of patients will develop a local recurrence (LR) after therapy for early stage breast cancer.1 Although factors contributing to a LR are well-described,1 defining prognostic factors for death after an isolated LR have been limited by small studies, inability to analyze the influence of the original therapy, varying definitions of "local" and "regional" recurrence, and inadequate follow-up. Le and colleagues studied 105 women with breast cancer and an isolated LR (breast or chest wall) who were treated at a single institution. With long follow-up after the recurrence, they were able to analyze potential prognostic factors for death including the effect of original treatment choices.
The patients were identified from a database of more than 7000 patients treated from 1954-1983. At diagnosis, patients had surgically treated tumors < 2.5 cm. Treatment consisted of either mastectomy or lumpectomy with breast irradiation (45 Gy with boost to tumor bed of 15 Gy). Regardless of the type of surgery, those patients with positive axillary nodes tended to receive axillary radiation as well. Other than 7 node-positive patients treated with ovarian suppression, adjuvant tamoxifen or chemotherapy was not given.2
The 105 patients with an isolated LR were identified after excluding those with previous or concurrent nodal or systemic relapse. By excluding those who recurred after January 1 1995, a long follow-up was achieved (median, 11.5 years). Fifty-seven of the 105 patients recurred again (8 isolated local, 33 distant, 16 both) of whom 52 patients died. This corresponded to 5- and 10-year overall survival rates (after the LR) of 76% and 56%, respectively, with a median survival of 12.9 years. This compares favorably with a median survival of 2.2 years when a distant recurrence was the first event.
Le et al chose to focus on the factors affecting the risk of death after a LR (this is not a study of the risks of having an LR). By univariate analysis, 3 factors were found to be significant: 1) tumor grade; 2) age at initial treatment; and 3) disease-free interval (DFI). Specifically, the risk of death was increased about 2-3 fold for grade 3 tumors, for patients older than age 60 at diagnosis, and for a DFI of less than 8 years. On the other hand, the risk of death was not influenced by the type of surgery at diagnosis, the use of radiation at diagnosis, or the original nodal status. The increased risk in older patients was felt to be due to competing causes of death.
Local treatment at the time of recurrence depended in part on the initial treatment. In the 55 patients originally treated with breast-conserving therapy, the LR was managed by mastectomy in most (82%) patients. Almost all of the 150 mastectomy patients received radiation therapy at relapse, even those previously radiated.
The benefits of systemic treatment at the time of recurrence depended on menopausal status. It appeared that both ovarian suppression and chemotherapy (CMF, FEC, or FAC) significantly decreased the risk of death in premenopausal patients, although the role of tamoxifen in these patients was unclear due to small numbers. In postmenopausal patients, there was no significant effect from tamoxifen or chemotherapy.
Comment by Kenneth W. Kotz, MD
Le et al from France report on a relatively large, single-institution retrospective study of breast cancer patients with a LR. Focusing on patients without nodal or systemic involvement, their goal was to define prognostic factors for death, and in particular to study the influence of the initial local treatment. They found that histologic tumor grade, age at initial diagnosis and DFI, but not the original stage or local treatment approach, were prognostic for death after a LR. This is consistent with other series which, in general, show that the DFI is the most reliable factor for predicting survival after a LR,1,3 with other factors inconsistently associated with prognosis, such as initial tumor grade and stage, tumor size at recurrence, young age, and premenopausal status.1,3 It has also been suggested that a LR after breast-conserving therapy may have a better prognosis than after mastectomy,3-5 although this was not observed by Le et al. The 2.5 cm cut off in the study by Le et al may have reduced a bias present in many other studies where patients with larger tumors tended to get a mastectomy.4
In patients previously treated with lumpectomy, the standard treatment for LR usually consists of mastectomy. Five-year relapse-free survival rates are 60-75%.3 Whether such patients could safely undergo a repeat breast-conserving approach is unknown3, although some have noted a higher LR rate with no effect on survival.4 A recent report does suggest that excision followed by repeat radiation therapy may be an acceptable alternative to mastectomy.6 Le et al treated 10 patients in this manner. Five of these patients also received a second course of radiation (details not provided).
Radiation therapy is the standard local treatment for mastectomy patients, with prior gross excision recommended, even if initial systemic therapy is required for resectability.5 The value of radiation in patients who previously received it is not well established,3 although Le et al gave repeat radiation to 90% of such patients (treatment details and side effects not reported).
Le et al found no benefit to tamoxifen or chemotherapy in the risk of death in previously untreated postmenopausal patients after a LR. On the other hand, they found a statistically significant decrease in the risk of death after a LR in previously untreated premenopausal patients associated with the use of ovarian suppression and chemotherapy (RR, 0.2 for both). For premenopausal patients, the tamoxifen data were unreliable due to the small number of patients who received it. Because the use of systemic therapy after a LR is of unproven benefit,3-5 its general use should be individualized. A randomized trial that addressed this issue showed that tamoxifen reduced the rate of local treatment failure and improved the 5-year disease-free survival rate from 36-59%.7 This difference had disappeared by 8-9 years after randomization and there was no impact on the development of distant metastases or on overall survival.3,4 Nevertheless, because of the excellent toxicity profile, hormonal therapy can be considered for patients with a LR due to the favorable toxicity profile.3,4 The use of chemotherapy is more problematic,3,4 even more so in previously treated patients at higher risk of drug resistance,5 but might be reasonable in eligible patients not previously exposed to cytotoxics.5
Dr. Kotz of is of Hanover Medical Specialists, Wilmington, NC.
1. Clemons M, et al. Cancer Treat Rev. 2001;27:67-82.
2. Le M. Personal Communication, July 2002.
3. Harris J, et al, eds. Diseases of the Breast. Philadelphia, Pa: Lippincott Williams & Wilkins, 1996.
4. Clemons M, et al. Cancer Treat Rev. 2001;27:83-92.
5. DeVita VT, et al, eds. Cancer Principles & Practice of Oncology. Philadelphia, Pa: Lippincott Williams & Wilkins, 2001.
6. Deutsch M. Int J Rad Onc Biol Phys. 2002;53:687-691.
7. Borner M, et al. J Clin Oncol. 1994;12:2071-2077.