A survey of the American Society of Plastic Surgeons in 2010 regarding the safety of autologous fat grafting (lipofilling) of the breast found that 49 percent of respondents considered the lack of evidence supporting the oncologic safety of lipofilling to be a significant obstacle to its use for breast reconstruction or cosmetic augmentation.1 A concern is that adult adipose tissue–derived stem cells transferred with the lipoaspirate may reactivate dormant tumor cells within the breast or activate primary breast cancer.
Clinical and animal studies have shown conflicting results2–5 as to whether lipofilling confers a higher risk for recurrence of breast cancer. Furthermore, it remains unknown which groups of patients might be more susceptible to any lipofilling-induced increase in the risk of recurrence. Although lipofilling of the breast is performed worldwide in thousands of patients per year, including as an alternative to implant placement for breast augmentation in some patients, there has been no study published with a control group regarding the oncologic safety of lipofilling in general or lipofilling of the breast. This lack of evidence has made many surgeons reluctant to offer lipofilling to patients, and surgeons who do perform lipofilling may have some concern that lipofilling may increase the risk for breast cancer recurrence.
The primary objective of this study was to determine whether lipofilling as an adjunct or primary breast reconstruction procedure increases the rate of locoregional recurrence of breast cancer. The secondary objective was to determine whether patients without breast cancer who underwent lipofilling after risk-reducing mastectomy had an increased risk of primary breast cancer.
PATIENTS AND METHODS
The Institutional Review Board of The University of Texas M. D. Anderson Cancer Center approved this retrospective analysis. The patients for analysis were identified in two stages. First, the database of prospectively collected data maintained by the M. D. Anderson Department of Plastic Surgery was searched for all patients who underwent segmental or total mastectomy for breast cancer or breast cancer risk reduction followed by breast reconstruction with lipofilling as an adjunct or primary procedure. For the patients identified as a result of this search, the dates of segmental or total mastectomy ranged from June of 1981 through April of 2013, and the dates of lipofilling ranged from January of 2001 through February of 2014. The patients identified from this initial search were divided into two groups: patients who underwent mastectomy for breast cancer treatment (i.e., cases) and patients who underwent mastectomy for reduction of breast cancer risk (i.e., cancer-free breasts), of which BRCA mutation information was available for 233 patients. Thirty-three patients were BRCA1/BRCA2 mutation carriers.
In the second stage of patient identification, the database of prospectively collected data maintained by the M. D. Anderson Department of Breast Medical Oncology was searched to identify patients with breast cancer who underwent mastectomy and breast reconstruction during the period from June of 1981 through April of 2013 but did not undergo lipofilling (i.e., controls).
To assess the effect of lipofilling on the risk of locoregional recurrence of breast cancer, we compared outcomes in the breasts reconstructed with (i.e., cases) and without (i.e., controls) lipofilling after mastectomy for breast cancer. Breasts reconstructed with lipofilling after mastectomy for breast cancer risk reduction or benign disease were analyzed to achieve our secondary objective of determining whether patients who underwent lipofilling after risk-reducing mastectomy had an increased risk of primary breast cancer.
Descriptive statistics, such as means and standard deviations, were used to summarize age and follow-up time. Frequencies and proportions were used to summarize the categorical characteristics. Demographic and clinical information was compared by using Fisher’s exact test and the Wilcoxon rank sum test. The recurrence-free survival time was defined as the interval from the date of mastectomy to the date of first locoregional recurrence or the date of last follow-up if no locoregional recurrence was observed. Patients without locoregional recurrence were censored in the analyses. The probability of locoregional recurrence was estimated by the Kaplan-Meier product-limit method. Because lipofilling was often performed at various time intervals after the oncologic surgery, we used time-dependent Cox proportional hazards regression models to assess the effect of lipofilling on locoregional recurrence. We also used the multivariable Cox proportional hazards regression models to adjust the potential confounding factors. All the tests were two-sided. The p value was calculated by comparing breast cancer patients who received lipofilling to breast cancer patients who did not receive lipofilling. A value of p < 0.05 was considered significant. The analyses were performed using SAS 9.3 (SAS Institute, Inc., Cary, N.C.) and R (The R Foundation for Statistical Computing, Vienna, Austria).
Our initial search of the plastic surgery database identified 1024 consecutive breasts reconstructed with lipofilling, 719 of them cancerous breasts (cases) and 305 of them cancer-free breasts (removed for breast cancer risk reduction). Of the 305 cancer-free breasts, BRCA status was available for 233 patients, of which 33 patients were carriers of the BRCA1/BRCA2 mutation. For the BRCA carriers, the mean follow-up time after mastectomy was 33.6 months and the mean time to lipofilling was 18.4 months. Through our search of the breast medical oncology database, we matched with the cases 670 cancerous breasts reconstructed without lipofilling (controls). The last patient follow-up was in November of 2014. Table 1 summarizes the patient and tumor characteristics. The mean follow-up time after mastectomy was significantly longer in cases (lipofilling) than in controls (no lipofilling) (59.6 months and 44.0 months, respectively; p < 0.001). The cases were older than the controls (mean age, 47.7 years and 46.5 years, respectively; p = 0.039). Cases and controls also differed with respect to pathologic stage, with cases more likely to have stage 0 or 1 cancer (p < 0.001). Tumor location and hormonal status (estrogen and progesterone receptor) were balanced between cases and controls. Controls were more likely to have HER2/neu-positive tumors (11.6 percent and 6.4 percent, respectively; p = 0.001) and more likely to receive chemotherapy (p < 0.001). Cases were more likely to receive hormonal therapy (p = 0.043).
Table 2 summarizes the characteristics of lipofilling (number of sessions and total volume injected) between cancer and benign cases. The two groups were similar with respect to the number of sessions. Total volume was greater in cases (p = 0.007).
Table 3 summarizes the differences in the risk of locoregional recurrence between lipofilling and no lipofilling in several subgroups. Overall, locoregional recurrence occurred in 1.3 percent of the cases (nine of 719 breasts) and 2.4 percent of the controls (16 of 670 breasts) (p = 0.455). We found that lipofilling did not affect the risk of locoregional recurrence in subgroups defined on the basis of pathologic stage, breast quadrant where the tumor was located, mastectomy, hormone receptor status, or use of chemotherapy or radiation therapy. We also used the multivariate Cox proportional hazards model in which we compared locoregional recurrence between cases and controls by adjusting for chemotherapy, radiation therapy, hormonal therapy, and clinical stage (p = 0.348). The only subgroup examined in which lipofilling was associated with an increased risk of locoregional recurrence was the subgroup treated with hormonal therapy, 1.4 percent and 0.5 percent for lipofilling and no lipofilling, respectively (p = 0.038). When a multivariate Cox proportional hazards model was performed only for hormonal therapy patients (adjusted for factors such as chemotherapy, radiation therapy, and clinical stage), lipofilling was still a significant factor with an increased risk of locoregional recurrence (p = 0.031).
Table 4 summarizes the differences in the risk of systemic recurrence between lipofilling and no lipofilling in several subgroups. Overall, systemic recurrence occurred in 2.4 percent of the cases (17 of 719 breasts) and 3.6 percent of the controls (24 of 670 breasts) (p = 0.514). There was no significant difference in the rates of systemic tumor recurrence between the breasts reconstructed with and without lipofilling overall or in any of the subgroups examined.
As shown in Figure 1, the cumulative 3-year and 5-year locoregional recurrence rates were 0.3 percent and 1.6 percent, respectively, for the cases (lipofilling) and 1.3 percent and 4.1 percent, respectively, for the controls (no lipofilling). The incidence of locoregional recurrence was 0.25 cases per 100 person-years for the cases and 0.65 cases per 100 person-years for the controls (p = 0.455). In the 305 healthy breasts reconstructed with lipofilling after risk-reducing mastectomy, no cases of breast cancer were observed during the follow-up period.
In contrast to previous studies that have evaluated the incidence of breast cancer recurrence in breast cancer patients who have undergone lipofilling for breast reconstruction after mastectomy, our study included a control group. In addition, our study included more patients than did previous studies. In our study, we found no significant differences in the rates of locoregional recurrence or systemic recurrence between breasts reconstructed with lipofilling and breasts reconstructed without lipofilling. We also compared rates of locoregional recurrence in breasts reconstructed with and without lipofilling within several subgroups defined on the basis of clinical, pathologic, and lipofilling procedure characteristics, and we found that the only subgroup in which the locoregional recurrence rate was higher for lipofilling was the subgroup treated with hormonal therapy. We also did not find any instances of primary breast cancer development in healthy breasts reconstructed with lipofilling, inclusive of 33 patients who were carriers for the BRCA1/BRCA2 mutation.
The increasing popularity of fat grafting is evidenced by the significant increase in articles being published on the subject. We could find only two articles on fat grafting for breast reconstruction published in 1993, but in 2013, more than 120 articles on this topic were published. Many of the previously reported studies focused on the technical aspects of serial fat grafting, including the number of fat grafting sessions required and the total volume usually used for breast reconstruction.
One study examined the trends in fat grafting through a national survey of members of the American Society of Plastic Surgeons (456 respondents of 2584 members sent the survey).1 In that study, 62 percent of respondents reported currently using fat grafting for reconstructive breast surgery. Twenty-eight percent of respondents reported currently using fat grafting for aesthetic breast surgery, and 59 percent of respondents had not performed fat grafting for aesthetic breast surgery and had no plans to do so in the future. When asked about potential obstacles to incorporation of fat grafting into clinical practice, 49 percent of respondents strongly agreed or agreed that the lack of evidence concerning the impact of fat grafting to the breast on breast cancer development or recurrence was an obstacle.
To overcome this obstacle so that 100 percent of surgeons can confidently incorporate fat grafting into their clinical practice, we first need to know what occurs physiologically when we inject fat grafts into patients’ breasts. Rigotti et al. in 2007 elucidated how lipoaspirate heals irradiated tissue through a process mediated by adipose-derived adult stem cells.6 Ultrastructural analysis of the lipoaspirate revealed a well-preserved stromal vascular component. However, well-preserved adipocytes were virtually absent. Cytologic characterization of the lipoaspirate by in vitro expansion showed that the mesenchymal stem cells corresponded to bone marrow–derived mesenchymal stem cells. Four to 6 months after injection of the lipoaspirate into the patient, adipocytes were normal, and the microvasculature exhibited normal ultrastructure. One year or more after treatment, the picture was substantially unchanged apart from a tendency toward shrinking extracellular spaces, with normal adipocytes and a well-formed microcirculation. Certainly, this information should guide clinicians to perform serial fat grafting procedures 4 to 6 months apart. In addition, our current study found that the median time from lipofilling to detection of locoregional recurrence (19 months) was not directly related to completion of the maturation process of the vasculature and extracellular space.
Another piece of information we need to confirm the safety of fat grafting is information regarding the interaction between the fat graft and the tumor bed. Hypothetically, the transfer of adipose tissue–derived stem cells or adipose tissue–derived mesenchymal stem cells could induce dormant tumor cells to reproduce and thereby predispose the patient to locoregional recurrence. In vitro and animal studies have produced conflicting findings regarding the impact of stem cells, with some showing positive and others showing negative associations with breast cancer cell proliferation. Petit and colleagues published a retrospective European multi-institutional study of 646 cases of lipofilling of the breast and found, to the concern of many surgeons, that the risk of breast cancer recurrence was higher in patients with in situ carcinoma than in patients with invasive breast cancer.7 In a follow-up retrospective study (59 patients), Petit et al. focused only on patients with in situ carcinoma of the breast and found that patients who underwent lipofilling had an 18 percent cumulative 5-year risk of locoregional recurrence, compared with a 3 percent cumulative 5-year risk in patients who did not undergo lipofilling.8 An important factor to consider in both studies by Petit et al. was that a large percentage of patients undergoing breast conserving therapy received only intraoperative radiation therapy.
In contrast to the findings of Petit and colleagues,7,8 we observed only 25 local regional recurrences (1.5 percent) and total locoregional recurrence incidence rates of 1.3 percent for lipofilling cancer patients and 2.4 percent for no-lipofilling cancer patients. In addition, we found no significant differences in 5-year cumulative locoregional recurrence rate between breast cancer patients with lipofilling (1.6 percent, 0.25 cases per 100 person-years) and breast cancer patients without lipofilling (4.1 percent, 0.65 cases per 100 person-years). Our findings were similar to those in a recent prospective study by Brenelli et al. of 59 patients9 that found that only three patients (4 percent incidence) had a recurrence of breast cancer, with an estimated annual rate of recurrence of 1.3 percent per year.
Theoretically, segmental mastectomy should be associated with the highest risk of locoregional recurrence after lipofilling because much of the breast tissue is not resected; however, in the previously mentioned prospective study of 59 patients,9 all three patients with a recurrence of breast cancer had invasive primary tumors and invasive recurrences. In contrast to the studies by Petit et al.7,8 and in agreement with the 59-patient prospective study,9 we found no significant differences in the risk of locoregional recurrence between breasts reconstructed with lipofilling and breasts reconstructed without lipofilling in either the invasive breast cancer or intraductal breast cancer subgroups. We found no significant differences in the risk of breast cancer recurrence between breasts treated with segmental mastectomy and breasts treated with total mastectomy along with lipofilling. We also found no significant differences in the incidence of recurrence when segmental and total mastectomy with lipofilling were individually compared to the control group.
The only variable that significantly increased the risk of breast cancer recurrence with lipofilling was receipt of hormonal therapy. Although the cases were more likely to receive hormonal therapy, the cases and controls had similar hormonal receptor status. Eight of 805 patients with hormonal therapy had locoregional recurrence. Among them, the locoregional recurrence rate in the cases was approximately three times that of the controls. A hypothetical potential role of hormonal therapy in enhancing a tumorigenic microenvironment or impacting crosstalk between adipose-derived mesenchymal stem cells and breast cancer cells is unknown based on current scientific knowledge. Although lipoaspirate is a known reservoir for adipose-derived mesenchymal stem cells, adipose-derived mesenchymal stem cells are not tumorigenic per se, as they are not able to induce neoplastic transformation of normal mammary cells. However, it is not known whether adipose-derived mesenchymal stem cells can exacerbate tumorigenic behavior in breast cancer cells, theoretically creating an inflammatory microenvironment that sustains tumor growth and angiogenesis.10 Interestingly, we also found that neither the total volume of fat injected nor the number of fat grafting sessions performed impacted locoregional recurrence. However, we found that larger grafts were required for the breasts with cancer than for the healthy breasts.
In assessment of our study, the degree to which the cases and controls were similar is an important consideration. Although there were some differences between cases and controls, including longer follow-up time, slightly older age, more stage 0 and I breast cancers, and more receipt of hormonal therapy among the cases, the cases and controls had similar hormonal receptor status, and the clinicopathologic differences tended to even the groups with respect to expected risk of locoregional recurrence. For instance, the higher stages of disease (higher risk of recurrence) and more chemotherapy (more aggressive therapy) administered in the control group were balanced against the longer follow-up time (higher risk of detecting recurrence) and more hormonal therapy (less aggressive therapy) in the case group.
The results of this study showed no increase in rates of locoregional recurrence, systemic recurrence, or second breast cancer and support the oncologic safety of fat grafting in breast reconstruction. Although receipt of hormonal therapy did significantly increase the risk of locoregional recurrence of breast cancer in patients who received lipofilling compared with patients who did not receive lipofilling, the recurrence rates were low. Before withholding lipofilling as a reconstructive option for breast cancer patients receiving hormonal therapy, a randomized trial would most be appropriate to determine the true clinical significance.
The University of Texas M. D. Anderson Cancer Center is supported in part by the National Institutes of Health through Cancer Center Support Grant CA016672. The authors wish to acknowledge Melissa Crosby, M.D., for contribution to this study.