Cell assisted lipotransfer in breast augmentation and reconstruction: A systematic review of safety, ef ﬁ cacy, use of patient reported outcomes and study quality

Background: Cell assisted lipotransfer serves as a novel technique for both breast reconstruction and breast augmentation. This systematic review assesses the ef ﬁ cacy, safety and use of patient reported outcome measures in studies involving cell assisted lipotransfer. We also carry out an objective assessment of study quality focussing on recruitment, follow-up and provide an up-to-date clinical trial landscaping analysis. Methods: Key electronic databases were searched according to PRISMA guidelines and pre-de ﬁ ned inclusion and exclusion criteria. Two independent reviewers examined the retrieved publications and performed data extraction. Results: 3980 publications were identi ﬁ ed. Following screening, 11 studies were included for full review, representing a total of 336 patients with a follow-up time ranging from six to 42 months. A degree of variation was noted in graft retention and reported satisfaction levels, although there were only three comparative studies with con ﬂ icting results. Complications occurred at a rate of 37%. Additionally, there was a paucity of objective outcomes assessments (e.g. 3D assessment modalities or validated patient reported outcome measures) in the selected studies. Conclusions: Cell assisted lipotransfer is a surgical technique that is currently employed sparingly within the plastic & reconstructive surgery community. Presently, further technical and outcome standardization is required, in addition to rigorous randomized controlled trials and supporting long-term follow-up data to better determine procedural safety and ef ﬁ cacy. Routine use of more objective outcome measures, particularly 3D assessments and validated patient reported outcome measures, will also help facilitate wider clinical adoption and establish procedural utility.


a b s t r a c t
Background: Cell assisted lipotransfer serves as a novel technique for both breast reconstruction and breast augmentation. This systematic review assesses the efficacy, safety and use of patient reported outcome measures in studies involving cell assisted lipotransfer. We also carry out an objective assessment of study quality focussing on recruitment, follow-up and provide an up-todate clinical trial landscaping analysis. Methods: Key electronic databases were searched according to PRISMA guidelines and pre-defined inclusion and exclusion criteria. Two independent reviewers examined the retrieved publications and performed data extraction. Results: 3980 publications were identified. Following screening, 11 studies were included for full review, representing a total of 336 patients with a follow-up time ranging from six to 42 months. A degree of variation was noted in graft retention and reported satisfaction levels, although there were only three comparative

Introduction
Breast augmentation was the most commonly performed cosmetic procedure in the US in 2014 with over 102,000 procedures taking place. 1 Conventionally, implants have been utilized. However, their use is associated with a number of complications, notably capsular contraction, malposition, and anaplastic large cell lymphoma. 2,3 In a proportion of post-mastectomies, the use of implants is not possible due to the irregularity of the soft tissue defects, particularly in post-radiotherapy patients. 4 An alternative is the employment of complex reconstructive techniques including deep inferior epigastric perforator (DIEP) and latissimus dorsi (LD) flaps, that have an inherent complication risk and longer recovery periods. 5 Although cell assisted lipotransfer (CAL) will not replace these procedures (due to the shape and projection profiles they achieve), it has potential to serve as an adjunct for small corrections or volume increases, and may serve as a less invasive option for patients hoping to achieve subtle aesthetic enhancements. It should be noted that there is limited evidence to suggest that fat grafting without the use of adipose derived stem cell supplementation can be used for complete postmastectomy reconstruction. This has, however, used a technique called BRAVA-assisted fat grafting were an external volume expansion device is applied to enhance graft survival. For example, Khouri 2015 conducted a level IV study on 488 women (616 breast) and concluded that BRAVA-assisted fat grafting is a minimally invasive, safe and economic alternative to other forms of breast reconstruction. 6 The technique has also been used by the same study author to investigate the effect of large volume fat grafting after BRAVA use or implant removal with positive results. 7 Adipose derived stem cells, fat grating and external volume expansion technology is therefore a potential area of future research, specifically with regards to whole breast and large volume fat grafting.
An option for both breast augmentation and reconstruction is autologous fat grafting. Although studies have reported a more natural breast contour, reports of fat resorption have been reported. 8 Reported graft retention using this procedure vary from 40 to 75%, and therefore there is room for improvement in the efficacy of this procedure. 9 It has been found that the key to fat graft retention is maximizing the surface area to volume ratio, and the vascularity of the recipient area. 10,11 Studies have suggested that adipose derived stem cells can survive the period of hypoxia post surgery that is thought to result in the necrosis of conventional fat. 11 This provides scientific rational to using CAL in breast surgery and is why the use of this technique can be seen as a key development in the repertoire of techniques available to surgeons. 12 CAL utilizes fat grafts that have been enriched with a patient's adipose derived stem cells (Figure 1 compares CAL to autologous fat transfer). Adipose derived stem cells are able to enhance both angiogenesis and adipogenesis. Translating this into the clinical setting, it is hoped that long-term graft retention and lower post-operative complication rates will result. 13,14 The abundance of adipose tissue Figure 1. The cell assisted lipotransfer procedure. Harvested fat is separated with approximately one half being used to harvest adipose derived stem cells and the other being processed for use as a fat graft. The two constituents are then combined and placed within the recipient site. Note that the black arrows indicate the procedure for a conventional autologous fat transfer procedure. makes the harvesting of these cells relatively easy, avoiding the need for in-vitro expansion. 15 In addition, the removal of fat from aesthetically sensitive anatomical regions offers secondary patient benefits.
Several clinical studies have assessed the use of CAL in breast surgery. However, there has been no objective assessment of the evidence in this area. Limited examples of secondary research have been published examining CAL for multiple indications, 16,17 but no studies focus exclusively on breastrelated procedures. In light of this and the prospective role of CAL in breast augmentation and reconstruction, this systematic review was conducted to assess the efficacy and safety of cell assisted lipotransfer in breast surgery. We also aimed to ascertain the quality of the completed studies with particular consideration to methods of recruitment, appropriate follow-up and assessment of patient reported outcomes. Through this, we aim to inform future clinical trial organizers of limitations of current studies.

Search strategy
A systematic review was carried out according to the PRISMA guidelines, 18 the details of which can be found in Supplementary Material Document 1.

Study selection
The literature and database searches identified 11 studies for inclusion ( Figure 2 provides a detailed breakdown of the search). The search of PubMed and Ovid yielded 3553 and 1825 studies, respectively. Searches through the bibliographies of the identified articles and relevant publications resulted in the addition of two further studies. A number of incomplete trials were identified from searches on clinical trials registers that will be analysed separately in the discussion section of this paper.

Overview of studies
The 11 studies included (Table 1) comprised of 336 participants with follow-up times that were both variable and occasionally inconsistent within the studies. They ranged from six months in Wang et al. 2015 19 and Peltoniemi et al. 2013 20 to 42 months in Yoshimura et al. 2008 12 (a single patient was followed for this length of time). The majority of trials (n ¼ 9) were observational, with eight being prospective and one retrospective with none being randomized. The further two studies comprised one case study 21 and one case series. 12 Six studies looked into CAL for breast augmentation. Five concerned the use of CAL in breast reconstruction to treat breast defects in patients undergoing cosmetic, congenital and reconstructive procedures.

Patient demographic and operative details
All procedures carried out were single-stage procedures with no in-vitro expansion of the harvested stem cells. Studies described a similar surgical technique utilizing initial stab incisions, cannula insertion and subsequent injection of the fat graft into the breast in a multilayered and circular manner to optimize distribution. Weighted mean age and BMI of participants were 42.7 (SD 7.15) years and 21.87 (SD 1.89) kg/m 2 , respectively. Weighted mean fat harvested was 691.77 (Standard Deviation ¼ SD 339.25) ml. Weighted mean fat injected into the left and right breast was 253.65 (SD ¼ 20.05) and 265.51 (SD ¼ 14.15) ml, respectively. When volume injected into each breast was not disclosed per breast this average was 174.97 (SD 63.69) ml/per breast. Fat was harvested from a number of anatomical locations with a number of different systems being used to process the fat as detailed in Table 2. Table 2 also summarizes data regarding cell count and viability within the fat grafts. A number of studies recruited women of solely Asian heritage (Table 1). 14,23,27,29,32 Quality assessment Figure 3a/b summarizes the bias assessment that was carried out. A large number of the studies exhibited selection bias as they did not provide details regarding their recruitment process. Many studies did not report the demographic of their patient cohort in detail for an assessment of whether the results could be extrapolated to the general population. The criteria used to measure outcomes were pre-specified and contained mostly objective methods of assessment but did include more subjective assessments of patient/surgeon satisfaction. For example, in Kamakura et al. 2011, 22 patients were asked to select a satisfaction rating from the options: excellent, good or fair with no focus on negative outcomes. Only three studies used comparison groups. In addition to this, one study Dos Anjos et al. 2015 23 used a "sham control" where the comparison group received fat grafts that were enriched with a low level of adipose stem cells that showed no advantage over conventional autologous fat transfer. Three studies exhibited a high level of attrition bias: Yoshimura, et al. 2010, 24 where 12 month data were collected for only six out of an original cohort of 15 patients and Peltoniemi et al. 2013, 20 where two participants were excluded as their BMI had changed too much. In addition, when it was realized that the intervention showed no advantage over conventional autologous fat transfer the study was discontinued. In Domenis et al. 2015 25 data was not collected for seven out of 30 patients. In a number of studies financial conflicts of interest were not declared with one study author declaring that he is an employee of Cytori Therapeutics (San Diego; USA): a regenerative-medicine company developing cell-based therapies from adult adipose tissue.

Graft survival
A summary of graft survival outcomes and how they were measured in each study are presented in Table 3. Jung et al. 2015, 26 27 found an average graft retention of 49.25% using CAL vs. 39% in its control group and represents the second study incorporating a control group. Dos Anjos et al. 2015 23 reported an average graft retention of 75% when using grafts that were high in their stem cell content compared to 50% in the group of patients that received low stem cell enriched grafts in what the authors describe as a 'sham control group'. Domenis et al., 25 the final controlled trial, found graft retention to be better in its cell enriched group also e however, this measurement was made though ultrasonography. In all other studies, authors report positive findings although no comparison group was present, and the quantitative data presented does not indicate any meaningful advantage over existing techniques. 27 Methods of measuring graft volume retention (summarized in Table 3) varied between studies and the case study by Calabrese et al. 2009 21 did not report any quantitative measurement of this outcome. The majority of studies used magnetic resonance imaging (MRI) due to its reliability and accuracy in measuring graft volume, 28 however a few studies opted for physical measurements (e.g. breast circumference). Yoshimura, et al. 2010 24 and Dos Anjos et al. also used 3D measurement techniques.

Complications
The complications that occurred during the course of each study are presented in Table 4. The overall complication rate across studies was 37%, with the most common side effect being calcification e comprising 83% of all complications. Fibrosis, and consequent hardening of the breast, was seen in a single case by Yoshimura et al. 2008 24 and an isolated case of Mondor's disease (a rare condition that Table 4 Complications that occurred during the course of each study. It should be noted that each complication was considered to be an independent event as information regarding multiple complications in single patients was not available. Calcification is by far the most commonly cited complication using CAL. Note that no information regarding complications was reported in Domenis  leads to thrombophlebitis of the superficial veins of the breast and anterior chest wall 29 ) was reported by Dos Anjos et al. 2015. 23 Five serious adverse effects were reported either during or immediately post-procedure by P erez-Cano et al. 2012 30 e but only two of these were described. They comprised of subcutaneous bleeding, which was thought to be due to the post-operative use of anti-coagulants, and bony metastasis; secondary to natural disease progression.

Patient reported outcomes
Validated PROMs were only used in a single study. 30 The study, by P erez-Cano et al. 2012, represents the most comprehensive assessment of PROMs and used a number of objective scales to assess patient satisfaction. This included Clough's Classification System for breast reconstruction, the Late Effects Normal Tissues (LENT)-Subjective Objective Management Analysis scoring system and a Quality of Life assessment. Overall, at 12 months, 58 out of 67 surgeons and 45 out of 67 patients were reportedly 'satisfied' with aesthetic outcomes, and 57 out of 67 surgeons and 50 out of 67 patients reported satisfaction with regards to the whole treatment process, when taking into account all of the assessment methods used. Quality of life scores remained constant throughout the follow-up period. Three studies did not report any measure of patient reported outcomes (See Table 1). Of the studies that did report patient reported outcomes, the methods for data collection were not stated. For example, in the study by Calabrese et al. 2009, 21 the patient reportedly stated their outcome as 'excellent' with no accompanying explanation of how this was justified. This also occurred in the study by Kamakura et al. 2011, 22 where 75% of patients were reportedly 'satisfied' with outcomes. Wang  Patients provided a yes/no or positive/negative evaluation in addition to a percentage value. All patients reported satisfaction with texture, softness and contour. A detailed breakdown incorporating all the outcomes assessed above was not provided.
Generally, the provision of patient reported outcomes in the studies identified was poor. Future studies, therefore, should only use validated PROMs such as the BREAST-Q that are used appropriately and time-critically. This will generate more meaningful patient insights and facilitate a more consistent degree of comparability between studies.
For additional analysis please see Supplementary Material Document 2.

Discussion
Due to the infancy of this technique in breast augmentation and reconstruction, the current level of evidence surrounding CAL makes it difficult to draw conclusions for its use in the clinical setting. The majority of studies included in this review are positive in relation to graft retention, complication rates and patient or surgeon reported outcomes. However, there is a lack of methodological rigour e the absence of control groups in eight out of the 11 studies makes it difficult to ascertain the efficacy and safety of CAL. In addition, one of the three studies 20,25 that made use of a control group found no statistical advantage in using CAL in contrast to the other two controlled studies 25,27 that did find a difference. Evidently, further work is needed to assess the full effect, if any, of CAL for this indication. Peltoniemi et al. 2013 20 highlighted that the conventional autologous fat transfer procedure takes around 90e150 min less time in addition to being $3000 cheaper than CAL. Any benefits that CAL may bring need to be weighed against such drawbacks.
One of the key limitations that this systematic review highlights is the short follow-up times that are insufficient to assess the long-term implications of using CAL. This is of particular concern as preclinical studies have shown the use of stem cell enriched autologous fat to increase rates of local cancer occurrence 31 and metastasis. 32 We also found that calcification was the second most common complication of CAL, occurring in four out of 336 patients. Although in all cases this calcification was found not to be of malignant origin, the presence of calcification may lead to false positive results on mammography causing undue stress e this can be mitigated through use of expert radiologists who can readily distinguish malignancy from fat changes in the context of fat grafting. 33 The field of plastic surgery is widely adopting the use of quality of life measures. 34,35 Although the majority of studies did acknowledge patient satisfaction e the methodology was generally not described and as such, their reliability is limited. Most notably, in Kamakura et al. 2011, 22 patients were questioned about their satisfaction with their treatment outcomes in terms of three possible responses: excellent, good, or fair, with no option to report any potential negative outcomes e demonstrating a considerable element of bias. Therefore, the use of validated PROMs, such as BREAST-Q, should be routinely utilized. These validated PROMs are superior to conventional methods of collecting patient reported outcomes and are considered reliable (in that they measure how the procedure has influenced the patient and not any other factor), valid (in that they are objective in what they intend to measure) and sensitive (in their ability to detect change in a patient). 36,37 The use of such measures is therefore twofold: firstly, this would allow a consistent measure of satisfaction and well-being that could be compared between studies to collect relevant information while minimizing bias; secondly, BREAST-Q covers a broad range of domains including psychosocial, physical and sexual well being in addition to satisfaction with breasts, care and outcome and how this relates to the patients expectations. 38 In doing so, reliable data could be pooled to take a more patient centred approach to improving CAL. It could also provide insight into patient understanding of the procedure including its limitations to foster improvement in the acquirement of informed consent.
To assess the current clinical trial landscape, active clinical trials involving CAL in breast-related procedures were identified through searches of relevant clinical trial registries (see Figure 4). In total, 15 active studies were identified e the majority in early stages of translation (9 in phase two clinical trials). This is in line with what would be expected for an intervention in the early stages of development. In contrast to the studies previously mentioned, a significant proportion of the trials already have or are planning to recruit large numbers of participants, with the largest aiming for 440. Whether this is achievable is open for debate but may offer an explanation as to why such a significant proportion (60%) are still in the recruitment stage. Six out of the 15 future trials intend to use a comparison group, for more rigorous data analysis into the effects of CAL in breast reconstruction or augmentation. Our systematic review highlighted inadequate trial follow-up times and a failure to address long-term outcomes, particularly surrounding the risk of malignancy. This is a major concern for both patients and clinicians and will likely dictate the viability of the procedure and its potential for universal application in breast-related and other procedures.

Limitations
Limitations of this study include the small sample of patients (n ¼ 336) and the high levels of selection bias found within the studies. Only three of the studies included control groups, making it difficult to compare CAL to conventional autologous fat grafting. Follow-up times varied within studies and were too short to detect long-term outcomes of interest e notably the risk of malignancy. The studies also employed heterogeneous methods to measure graft survival with a corresponding lack of validity regarding the methods of assessing patient satisfaction levels post-operatively. In addition, a number of studies failed to declare financial conflicts of interest and in one study an author declared himself an employee a company developing cell-based therapies from adult adipose tissue. The large degree of heterogeneity meant that a meta-analysis to calculate the true effect size of cell assisted lipotransfer for the outcomes of interest discussed within this systematic review was not feasible.

Conclusion
The literature indicates that CAL may be a promising surgical technique. Presently, studies demonstrate high levels of bias, lack control groups and display considerable heterogeneity, making the generalizability of study results and effect size unclear. Furthermore, lack of long-term follow-up data and associated concerns of malignant risk require mitigation. Further rigorous randomized controlled trials are needed to investigate the procedure's full clinical efficacy and safety profile that should also make use of validated PROM's.  . The majority of studies are still in the early stages of translation in an attempt to ascertain the efficacy and safety of the procedure (a) and a major proportion of the trials plan on recruiting a larger number of participants (Note sample number unknown for one study) (b). An increasing number of trials plan on employing comparison groups especially for breast reconstruction (Note that some trials are investigating CAL for both breast augmentation and reconstruction so these are shown separately in this figure) (d). On the contrary future trials still fail to show a follow-up time long enough to demonstrate long-term effects (c). companies that among other services provide cell therapy biomanufacturing, regulatory, and financial advice to pharmaceutical clients. Smith is a consultant with IP Asset Ventures Ltd. Brindley also is subject to the CFA Institute's codes, standards, and guidelines, so he must stress that this piece is provided for academic interest only and must not be construed in any way as an investment recommendation. Additionally, at time of publication, Brindley and the organizations with which he is affiliated may or may not have agreed and/or pending funding commitments from the organizations named herein.