Institutional review board approval and informed consent were obtained for the retrospective chart review of consecutive, PADM-assisted prepectoral breast implant surgery cases performed at a single academic medical center (IRB file No. SCHUH 2019-06-006-

002). After a retrospective chart review of PADM-assisted prepectoral breast implant surgery using PADM (Supporix; HansBioMed) between December 2017 and December 2019, an analysis was performed. The candidates for our study were revisionary cosmetic breast surgery (CBS) patients with severe capsular contracture or foreign body injection, immediate implant-based breast reconstruction (IIBR), and delayed implant-based breast reconstruction (DIBR). The following clinical characteristics were recorded: age, body mass index (BMI), preoperative diagnosis, operation type, implant volume, laterality, operation time, follow-up period, and complications. Complications such as animation deformity, bottoming out, palpability, visibility, infection, seroma, flap necrosis, and capsular contracture were investigated. In addition, surgical costs according to the operation time and cost of the ADM were evaluated.

Before inserting the implant, we determined the implant volume and prepared the prepectoral pocket. The implant volume was determined by the volume of the excised mastectomy specimen (for reconstructive indications) or the patient’s preference and contralateral volume (for cosmetic indications). If the pocket was much wider than the implant, the lateral part including the serratus anterior muscle was sewn to narrow the width for preventing implant migration. PADM measuring 30×20 cm (Supporix, HansBioMed) and a cohesive round silicone gel-filled implant (Bellagel, SmoothFine; HansBioMed) were used for prepectoral implant placement with the ravioli technique. Supporix is a collagen tissue matrix used to support soft tissue regeneration that is derived from freeze-dried porcine skin, which was approved for use in February 2016. Before implant insertion, the PADM was emulsified for 20 minutes. After making a stab incision on the PADM, the implant was completely wrapped with half-folded PADM. The surrounding area was sewn with interrupted Vicryl 3/0 sutures. The remnant area was removed (Fig. 1). After confirming that the pectoralis major muscle (PMM) was intact, the implant was inserted above the PMM. ADM-fixation suturing was not performed (Supplementary Video 1). One silicone suction drain was inserted between the ADM cover and the skin. Double-layer wound closure was performed with Monosyn 4/0 interrupted sutures and Nylon 5/0 interrupted sutures for skin closure. The drains were left in place in all cases until the drainage was less than 20 mL in a 24-hour period. All patients received an individualized, treated postoperative dressing to prevent complications.

A systematic review of available literature was carried out in May 2020 using PubMed. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were thoroughly followed for the literature search, selection, and reporting of the available evidence. A standardized data abstraction form was developed and utilized by the authors (Fig. 2). We included references that described the use of PADM in cosmetic and reconstructive breast surgery. A literature search was performed in PubMed using the search string “porcine” AND “acellular dermal matrix” AND “breast.” The search was limited to articles in the English language and publication years from 2010 to 2020. Initially, only case reports were excluded. The titles and abstracts were screened to meet the following inclusion criteria: human studies, biologic material, prepectoral breast reconstruction, prepectoral CBS, histological data, and complications. Brief communications, correspondences, discussions, letters, and conference abstracts were excluded. The articles deemed relevant were subsequently read and examined in full to assess their eligibility. The reference lists of the included articles were also examined to identify additional relevant papers. All publications that met the inclusion criteria were deemed potentially eligible with respect to complication rates and the effectiveness of PADM in breast surgery. A search for unpublished literature was not performed. Two authors (WJS and SGK) abstracted the data independently and resolved any differences by discussion.

All statistical analyses were performed using Rex, an Excel-based statistical analysis software version 3.5.0.1 (RexSoft Inc.). To determine the 95% confidence intervals (CIs) of PADM complications by the meta-analysis of single proportions, the Freeman-Tukey double arcsine transformation was used for the summary statistics. The study-specific prevalence estimates were then weighted by the inverse of their variance to compute the pooled prevalence and its 95% CI. Random-effects pooled estimates were calculated. To check for between-study variance, a version of the DerSimonian-Laird method was used [5]. To compare the complication rates between our study and the other studies, we performed an exact test of the probability of success of the Bernoulli experiment. To compare the effect of PADM between our study and the other studies through the meta-analysis of complications, a target group was set up as the total group of patients who underwent breast surgery using PADM. Subgroup analysis was performed for the three groups according to operation type: the CBS group, the IIBR group, and the DIBR group. All complications requiring intervention were registered and analyzed by dividing them into eight categories (overall complications, infections, skin necrosis, seroma/hematoma, implant loss/explantation, implant exposure, capsular contracture, and minor complications) according to their frequency and probability. The minor complications included partial nipple necrosis, red breast syndrome, contour deformity, delayed wound healing, and others. Publication bias was assessed using funnel plots and Rucker’s arcsine-Thompson’s (AS-Thompson’s) test. A sensitivity analysis was also conducted to assess the influence of publication bias, as estimated by fitting a Copas selection model.

A literature search was performed following the PRISMA guidelines. Our search yielded 78 original articles (33 articles in PubMed and 45 articles in the Web of Science). From those, a total of 26 articles were included in this paper for qualitative analysis. Studies (n=20) that clearly specified the numbers of breasts undergoing unilateral and bilateral procedures, the operation type, and complication rates were included in the meta-analysis (Fig. 3). In the systematic review, a total of 2,504 breasts in 1,921 female patients were analyzed for complication rates after PADM-assisted prepectoral implant placement. The pooled incidence of overall complications was 24% (95% CI, 17%–32%). Substantial heterogeneity (I²=93.97%; P<0.01) was detected. Therefore, a subgroup analysis by operation type was conducted. The overall complication rate was 12% in the CBS group, 28% in the DIBR group, and 26% in the IIBR group. A detailed analysis of complications showed that seroma and hematoma had the highest incidence in the CBS group (4%) and the IIBR group (10%). However, skin necrosis was the most common complication in the DIBR group (8%). The results of our systematic literature review are presented in Table 1 and Fig. 4 [6-25]. Publication bias analysis conducted with AS-Thompson’s test showed asymmetry in the funnel plot, with a statistic of 0.70 (P>0.05). This result means that significant asymmetry was found, and the results derived from this meta-analysis should be considered cautiously. Fourteen of the twenty-four studies fell outside the 95% CIs denoted by the diagonal dashed lines shown in the funnel plot (Fig. 5A), which suggests possible heterogeneity and publication bias. The logarithm of the PFT (Freeman-Tukey double arcsine transformation) from the meta-analysis may be sensitive to the range of (γ0, γ1) values. As the probability of publishing the study with the largest standard error decreased from 100% to 48%, the estimated PFT decreased from 0.43 to 0.52 (Fig. 5C). The model indicates statistically non-significant residual publication bias (i.e., P>0.10) for the diminishing probability of publication (Fig. 5D). Overall, adjusting for selection bias, the Copas selection model yielded an estimate of 0.25 (95% CI, 0.18–0.32), as did the random-effects model estimate of 0.25 (95% CI, 0.18–0.33).

PADM-assisted prepectoral implant placement was performed for a total of 20 breasts in 16 female patients (four bilateral: all cosmetic indications; 12 unilateral: 11 reconstructive, one cosmetic indication). The mean age of the patients was 50.2 years, and the mean BMI was 23.2 kg/m². All patients asked for an intact PMM and preferred prepectoral implant placement with complete PADM coverage (the ravioli technique). For those with cosmetic indications, two had bilateral capsular contracture, two had foreign body injection, and one had implant rupture. For those with reconstructive indications, eight patients underwent immediate reconstruction followed by unilateral nipple-sparing mastectomy and three patients underwent delayed reconstruction following a lumpectomy, skin-sparing mastectomy, and contralateral augmentation mammoplasty for breast symmetry. All patients underwent direct-to-implant (DTI) breast reconstruction. In patients undergoing delayed reconstruction, skin volume and expansion were sufficiently feasible; therefore, DTI reconstruction was used rather than a tissue expander. The median follow-up period was 8.25 months (range, 3–24 months). Arm and shoulder mobility showed no restrictions. No patients complained of postoperative pain. No animation deformity was reported. The cosmetic results and the patients’ satisfaction were excellent. The implant rims were not visible or palpable. The breasts were smooth except for one case of capsular contracture (grade I). Preoperative and postoperative images of patients undergoing immediate reconstruction, delayed reconstruction, and surgery for cosmetic indications are shown in Figs. 6-8. Five complications (two cases of seroma, one case of mild infection, one case of contour deformity, and one case of partial nipple necrosis) occurred only in three patients with immediate reconstruction. No major complications such as total skin necrosis or severe infection occurred during the mean follow-up period of 8 months. A 57-year-old patient (patient No. 6) showed contour deformity of the left breast at a 1-year postoperative follow-up. The patient was suspected of having an implant rupture and seroma formation based on the computed tomography and magnetic resonance imaging findings. There was no rupture of the implant in the reoperation findings. However, the patient requested implant removal due to there were no abnormal findings, a biopsy was not performed. A 48-year-old patient (patient No. 8) showed seroma and partial nipple necrosis, which was healed by conservative treatment. A 56-year-old patient (patient No. 11) showed seroma and mild infection with fever at 1 month postoperatively. A salvage procedure was performed without implant removal. She showed a good recovery. All seromas were managed by clinic-based aspiration. All case summaries are shown in Table 2.

In the comparative analysis, the overall rate of complications in the total group cohort was 14% in our study and 24% in the other studies, showing no significant difference. In the subgroup analysis, the overall complication rates in the CBS group and the DIBR group were lower in our study (CBS: 0% vs. 12%; DIBR: 0% vs. 28%), but significant differences were not found between our study and the other studies. Although most complications were minor and did not require hospitalization or reoperation (25% vs. 2%, P=0.01), the overall complication rate in the IIBR group was significantly higher in our study (62% vs. 26%, P=0.03). In particular, no complications were observed in the CBS group or the DIBR group in our research. In the CBS group, the incidence of infection, skin necrosis, implant loss and explantation, implant exposure, and capsular contracture was 0% in both our study and the other studies. In the IIBR group and the DIBR group, the incidence of implant exposure and capsular contracture was also 0% in both our study and the other studies. The difference in the complication rates in Tables 1 and 3 is due to the statistical difference caused by the sum of the number of study populations included in the systematic review during the subgroup analysis. All comparative analysis data are provided in Table 3.