Introduction

Based on societal image and beauty standards, the number of women willing to receive breast reconstruction after breast cancer surgery is rapidly increasing [1]. Adipose tissue is an accessible and abundant source for soft tissue reconstruction [2]. However, its oncologic safety remains controversial [3].

Autologous fat graft (AFG) augments the soft tissue volume by adding mature adipocytes and potentially numerous progenitor cells such as adipose-derived stem cells (ASCs). ASCs have a pivotal role in the survival of adipocytes after AFG by enhancing angiogenesis and tissue regeneration resulting from the secretion of various growth factors and cytokines. These processes have been demonstrated to stimulate cancer cell proliferation and progression under experimental conditions [4,5,6,7,8].

Considering the oncogenic potential and possible disturbance to breast cancer surveillance, the American Society of Plastic Surgeons (ASPS) issued a warning against the use of AFG, describing it as not safe enough and unreliable for breast reconstruction in 1987 [9]. However, starting in the early 2000s, AFG has been rediscovered and reassessed. Coleman et al. described benign postoperative findings on mammograms following AFG [10]. In 2009, the ASPS Fat Graft Task Force rescinded the prohibition of using AFG in breast reconstruction and published new recommendations on its efficacy and safety [11]. Additional clinical retrospective studies have reported consistent results supporting the safety and efficacy of AFG in breast reconstruction. These reports have reduced the initial concern over the oncogenic potential and possible disturbance of AFG in cancer surveillance [12,13,14,15,16,17]. AFG continues to gain popularity as an alternative to breast reconstruction for eligible patients that results in a breast with a natural contour and supple touch [18, 19]. AFG in low volumes can play a role as an additional procedure to implant-based, autologous tissue and partial breast reconstruction to correct minor contour irregularities, while multiple applications of AFG with higher injection volumes can be performed as an alternative to conventional whole breast reconstruction [2, 19].

Two-stage implant-based reconstruction (IBR) is the most prevalent breast reconstruction method that benefits from additional autologous lipotransplantation in various ways [19, 20]. AFG is commonly involved in implant placement after tissue expansion. AFG also may be used as an interim procedure over the tissue expander before the final implant exchange for patients who have received previous radiation treatment or with thin mastectomy flaps [21,22,23]. In the event of implant failure, implant-to-fat conversion may result as a breast of lower or nearly equal volume with a more natural contour and a more supple feeling than that of IBR for patients who cannot accept a prolonged tissue flap transposition or have experienced recurrent infections relating to prostheses [2, 19, 24]. Fat injection may also be performed secondarily or in delay to enlarger the breast volume after autologous-based breast reconstructions or simultaneously performed with pedicle flap transposition in slim patients desiring an extensive volume reconstruction but without the use of prosthetic devices [25, 26]. AFG also can be utilized to address fat necrosis, which is an unfavorable outcome resulting in abnormal palpable and/or visible densities within the reconstruction flaps [19]. Moreover, AFG has been used in the treatment and reconstruction of irradiated tissue and scars by improving fibrosis, alleviating breast cancer-related lymphedema and cellulitis, and treating postmastectomy pain syndrome [27,28,29,30,31].

Several studies have reported success with primary reconstructions only using AFG in a staged manner and multiple grafting sessions [32, 33]. Stark et al. [26] described the primary AFG after a nipple-sparing mastectomy that required at least two sessions and an average of 40% volume adsorption per session to restore the soft tissue deficiency. Zhang et al. [32] also reported 30 successful cases of total breast reconstruction utilizing high volume, condensed, and viable AFG with an average of 3.3 sessions. Herly et al. conducted a meta-analysis estimating the efficacy of using AFG alone for breast reconstruction. The estimated numbers of fat grafting sessions for a complete reconstruction were 2.84–4.66 in the mastectomy groups and 1.72 in the BCS groups, with a significantly greater number of sessions needed in the irradiated patients [1]. Relatively low and uncertain fat graft retention rates, about 20–75 %, are one limitations of using fat grafting as the primary procedure for breast reconstruction due to the need for multiple sessions [34]. The modification and enrichment of fat grafts or their use in combination with an expanded ASC population may increase fat graft retention. However, the oncological safety of these techniques, especially when patients have received oncology surgery, still needs to be determined [16, 35, 36].

Since AFG as a completely different approach from traditional reconstructive techniques, a viable alternative for controls is lacking [37]. Therefore, establishing a randomized controlled trial (RCT) for AFG is impractical, unethical, and not possible to achieve in the foreseeable future. Most of systematic reviews published up to now have primarily consisted of descriptive summations based on individual studies [3, 38, 39]. A systematic review conducted by Decker et al. [40] involved totally 2419 patients and reported that the prevalence of localized cancer recurrence was 1.69%. While there is not enough evidence to demonstrate the oncological safety of AFG, investigators have approached this goal using retrospective studies, including case series and cohort studies. In particular, in matched cohort studies, each patient undergoing AFG was matched to at least one control patient based on relevant prognostic indicators including age, tumor size, tumor histology, date of oncology surgery, modality of surgical treatment, disease stage, lymph node involvement, estrogen receptor status, Bloom and Richardson grade, human epidermal growth factor receptor 2 overexpression, and progesterone receptor status, to reduce the heterogenicity of baseline characteristics between the groups to the lowest possible level. Several recently published studies using matched cohorts have provided a viable method to estimate the oncological safety of AFG [41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56]. A meta‐analysis by Wang et al. found no increase in the risk of localized and reigonal recurrence (LRR) in AFG subgroups when the surgical modalities, tumor histology types, and whether patients received postoperative radiation therapy were considered from 11 cohort studies that included 5550 patients [57]. Krastev’s meta-analysis that estimated the unmatched cohorts revealed a significantly lower overall incidence of LRR after AFG in patients compared with controls. However, another meta-analysis of seven matched cohorts suggested no significant differences in the LRR incidence between AFG and control groups [37].

Currently, the optimal clinical evidence concerning the oncological safety of AFG in breast reconstruction has been obtained from matched cohort studies. Therefore, our study aimed to estimate the oncological safety of AFG based on matched cohort studies.

Methods

This study was performed following the guidelines published in the Preferred Reporting Items for Systematic Reviews and Meta-analysis statement [58].

Search Strategy

Li conducted an electronic literature search using PubMed on August 1, 2021. The search terms used included: (breast reconstruction OR breast conserving surgery OR mastectomy OR quadrantectomy OR lumpectomy) AND (fat graft OR fat transplantation OR fat transfer OR lipotransfer OR lipografts OR lipofilling OR lipomodeling).

Inclusion and Exclusion Criteria

The inclusion criteria were as follows: (a) the participants were patients who received breast reconstruction with AFG after breast cancer surgery, (b) the control groups were treated without AFG, (c) the oncological events were evaluated, (d) it was a matched cohort study. Studies were retrieved manually or from the reference lists of studies that met the inclusion criteria.

Data Extraction

Li and Shi screened the titles and abstracts of all the retrieved articles independently. Studies found manually or from the reference lists of included studies were selected according to the inclusion and exclusion criteria. The relevant information was extracted from each study by Li and Shi independently. Any discrepancies between the two authors were resolved by discussion among all the authors. The authors unanimously agreed on the final decisions and confirmed the extracted information from included studies.

Statistical Analysis

The pooled effect estimate was conducted using Review Manager software (RevMan, version 5.3, Copenhagen, Denmark) with a fixed-effects and Mantel–Haenszel model. Odds ratios (OR) and 95% confidence intervals (CIs) were used to express the outcomes. Subgroup analyses were performed based on the types of surgery (breast‐conserving surgery (BCS) and mastectomy) and tumor histology (in situ cancer and invasive cancer). An I2 value > 50% indicated a significant heterogeneity. Statistical significance was achieved when P was less than 0.05.

The Risk of Bias

It is not possible to establish a RCT for AFG in patients after breast cancer surgery, so no RCT was used. Considering that all the studies included in our analysis were matched cohort studies and consisted of 16 retrospective matched cohort studies and one prospective matched cohort study, we did not assess the risk of bias in our analysis.

Results

The protocol for the literature search and selection is illustrated in Fig. 1 as a flow diagram. The initial search in PubMed and other sources identified 2600 citations after duplicates were removed. According to the titles and abstracts, 124 unique studies were selected for full-text review. Seventeen matched cohort studies with a total of 7494 patients were identified. Sixteen of the studies were retrospective matched cohort studies and one was a prospective matched cohort study; the studies met all the inclusion criteria and were incorporated in the meta-analysis [41,42,43,44,45,46,47,48,49,50,51,52,53,54,55,56, 59]. The extracted information from the 17 studies included: bibliographical data (i.e., author, year), study design, sample size, basic characteristics of patients (mean age, mean body mass index (BMI)), follow-up period, tumor histology, breast cancer surgery, and incidence of cancer recurrence (Tables 1 and 2). The data used for subgroup analysis regarding types of surgery and tumor histology are listed in Table 2.

Fig. 1
figure 1

A flow diagram of the literature search and selection process

Full size image
Table 1 Baseline characteristics of the included studies
Full size table
Table2 Characteristics of subgroup
Full size table

Meta-Analysis

Localized and Regional Recurrence (LRR)

Sixteen studies reported LRR rates for the AFG and control groups. The pooled estimate of the 16 studies indicated that AFG was not associated with a significant increase in the risk of LRR compared with controls (OR 0.97; 95% CI 0.76–1.24; P = 0.82). There also was no significant heterogeneity among included studies (I2 = 0%, heterogeneity P = 0.60; Fig. 2). Thus, the outcome suggested that AFG does not increase the risk of LRR in patients after breast cancer surgery.

Fig. 2
figure 2

Comparison between AFG and control group regarding LRR

Full size image

Subgroup Analyses of LRR

Subgroup analyses assessing the effect of AFG on LRR were conducted considering different types of surgery (BCS or mastectomy) that patients received. The pooled estimate OR revealed no significant different in LRR between the AFG and control groups in patients undergoing either BCS (OR 1.02; 95% CI 0.42–2.51; P = 0.96; Fig. 3) or mastectomy (OR 0.79; 95% CI 0.51–1.22; P= 0.29; Fig. 3). Testing for the overall effect indicated that the AFG group was not associated with a statistically significant difference in LRR relative to the control group (P = 0.35), and no heterogeneity was observed among the studies (I2 = 0%, heterogeneity P = 0.91).

Fig. 3
figure 3

Subgroup analysis of AFG compared with control group in LRR. Subgroups were delimited based on patients undergoing BCS or mastectomy

Full size image

Further subgroup analyses of LRR were performed concerning the tumor histology (in situ or invasive breast cancer). The outcome showed that AFG was not associated with a significantly increased LRR in patients with either in situ (OR 0.29; 95% CI 0.08–1.11; P = 0.07) or invasive breast cancer (OR 0.76; 95% CI 0.42–1.35; P = 0.34; Fig. 4). The test for overall effects indicated that the AFG group was not associated with a statistically significant LRR relative to the control group (P = 0.10). There was no heterogeneity among the studies (I2 = 0%, heterogeneity P = 0.66). Outcomes of subgroup analyses confirmed that AFG was not associated with a higher risk of LRR in patients after breast cancer surgery.

Fig. 4
figure 4

Subgroup analysis of AFG compared with control group in LRR. Subgroups were delimited based on patients with either in situ or invasive breast cancer

Full size image

Distant Metastases (DM)

ORs for DM were available in eight studies. The pooled ORs revealed that no significant difference was present for the risk of DM between the AFG and control groups (OR 1.13; 95% CI 0.84–1.51; P = 0.42; Fig. 5). Heterogeneity also was not significant among the studies (I2 = 0%, heterogeneity P = 0.57). The outcome indicated that AFG does not increase the risk of DM in patients after breast cancer surgery.

Fig. 5
figure 5

Comparison between AFG and control group regarding DM

Full size image

Subgroup Analyses for DM

Subgroup analyses assessing the effect of AFG on DM indicated that no significant increase in DM was associated with AFG when patients underwent BCS (OR 1.22; 95% CI 0.26–5.82; P = 0.80) or mastectomy (OR 0.87; 95% CI 0.46–1.65; P = 0.69; Fig. 6). Testing for the overall effect confirmed that AFG did not present a statistically significant difference in DM relative to the control group (P = 0.75). No significant heterogeneity was observed among the subgroups (I2 = 0%, heterogeneity P = 0.64).

Fig. 6
figure 6

Subgroup analysis of AFG compared with control group in DM. Subgroups were delimited based on patients undergoing BCS or mastectomy

Full size image

Considering that the sample sizes of the subgroups relative to diverse histological types (in situ or invasive breast cancer) were limited, the results of the subgroup analyses of DM in patients with different tumor histology (in situ or invasive breast cancer) were not included. Outcomes of the subgroup analyses confirmed that AFG was not associated with a higher risk of DM in patients after BCS or mastectomy. In general, no heterogeneity was observed in the meta-analysis results for DM (I2 = 0%, Fig. 5). Therefore, we concluded that AFG did not increase the risk of DM in patients after breast cancer surgery.

Discussion

This study included increased numbers of patients in matched controlled studies to demonstrate that LRR and DM after AFG were not altered in the general cancer patient population. Similarly, the subgroup analyses revealed no significant statistical differences in LRR and DM between the AFG and control groups.

One striking aspect of the controversy concerning AFG oncologic safety focuses on the inconsistent outcomes of clinical and basic science studies [34, 60]. Several groups have repeatedly demonstrated that ASCs promote the growth and migration of breast cancer cells under experimental conditions [4,5,6,7,8]. In theory, activated ASCs should produce regenerative responses starting a few months following AFG to 1 year following the procedure. Variations in the recurrence mode often occur among tumors based on differences in their receptor status and histopathological stages. Recurrence generally occurs during the first and the fifth year of oncological follow-ups. However, the influence in the cancer recurrence rates attributed to AFG later than 5 years after the first AFG exposure has not been clarified [37]. The follow-up times reported in most studies concerning delayed AFG are approximately 36 months after treatment and 72 months after breast cancer surgery [37]. The mean total follow-up time for the studies included in our analysis was 74.9 months. The mean follow-up time after the first fat grafts was 43.1 months. A matched retrospective cohort study conducted by Stumpf et al. [43] reported that at a mean 5-year follow-up assessment, the risks of LRR in patients who underwent immediate AFG after BCS and those who received only BCS were not significantly different. In other clinic studies of immediate AFG with BCS, mean follow-ups were approximately 3 years [17, 61]. Khan et al. [17] conducted a cohort study to evaluate the immediate AFG efficiency. No difference was found in the incidence of local or systemic recurrence between the AFG and control groups after a follow-up time of 36 months. In 2018, a prospective, nonrandomized, uncontrolled study conducted by Biazus et al. involving 65 patients who underwent BCS with immediate lipofilling reported that the LRR was not different from expected, and the mean follow-up time was 40.8 months [61]. Therefore, additional studies assessing the safety of immediate or delayed AFG for at least 5 years following the initial procedure are needed. It also is necessary to clarify the divergence between clinical and basic science AFG studies in the future.

Furthermore, most clinical studies and meta-analyses described the oncologic safety of AFG after breast cancer in patients with in situ or invasive carcinomas, and the subgroup analysis in our study showed similar results [41, 42, 51, 53, 59]. However, Petit et al. reported a higher incidence of local events for intraepithelial neoplasia in patients who received AFG in one matched cohort study. This was the only included study that reported a significantly higher risk of cancer recurrence in the AFG group than the control group [55]. According to the World Health Organization (WHO), the classification of breast tumors, atypical hyperplasia, and carcinoma in situ are collectively referred to as intraepithelial neoplasia [62]. Atypical hyperplasia is considered a precancerous lesion, and its diagnosis and treatment are critical for successful treatment and tumor prevention [63]. This suggests that more specific information concerning the histology or tumor stage needs to be considered as risk factors when evaluating the oncology safety of AFG for breast reconstruction and not to focus on only invasive or in situ carcinomas.

Limitations

This study had several limitations. First, bias was an unavoidable confounder in the retrospective studies including in this study. However, the multivariate analyses and exact matching used in matched cohort studies might reduce the inherent selection biases. Second, no standardized surgical modality or uniform viable methodology was used to estimate the recurrence rate in patients who received AFG after breast cancer. Finally, the sample sizes in the subgroup analyses were relatively limited.

Additional high-quality studies with detailed information, including the type of cancer surgery, specific tumor histology, surgical modality of breast reconstruction, the timing of fat grafting, and longer follow-up times after AFG exposure, are needed to further establish the long-term safety of AFG in patients after breast cancer surgery.

Conclusion

Our meta-analysis based on matched control studies found no significant increase in the risk of LRR and DM after AFG for breast reconstruction in patients who received oncology surgery. Subgroup analysis involving tumor histology types and surgical modalities led to similar results. Thus, our study may provide evidence-based conclusions supporting the application of AFG in breast reconstruction.